National Design Competition: Accessible Ergometer

BME 402 Department of Biomedical Engineering

University of Wisconsin – Madison June 1, 2005

Team: Amit Mehta Jon Millin

Ryan Pope Jeff Swift

Contact: John Enderle, Ph.D.

Department of Electrical and Computer Engineering University of Connecticut

Advisor: Justin Williams, Ph.D.

Department of Biomedical Engineering University of Wisconsin – Madison

Table of Contents Executive Summary..................................................................................................................................... i Abstract........................................................................................................................................................1 Background .................................................................................................................................................1 Design Constraints .....................................................................................................................................2

Mainstream Considerations ......................................................................................................................4 95th Percentile Illusion...............................................................................................................................4 EZ Access.................................................................................................................................................5

Accessible Design Principles ....................................................................................................................6 Low Vision.................................................................................................................................................6 Blindess ....................................................................................................................................................7 Deafness...................................................................................................................................................8 Parkinson’s disease..................................................................................................................................8 Stroke........................................................................................................................................................9 Heart Failure .............................................................................................................................................9 Diabetes..................................................................................................................................................10

Current Competition .................................................................................................................................11 Chosen Device: NordicTrack® SL710.....................................................................................................11 Modifications to Existing Device.............................................................................................................13 Additions to existing device ....................................................................................................................15

Power Seats............................................................................................................................................15 Track system...........................................................................................................................................18 Seat Platform ..........................................................................................................................................19 Seat Supports .........................................................................................................................................22 Arm Motion Exercise...............................................................................................................................24 Arm Handles ...........................................................................................................................................27

Wireless Light Emitting Diode (LED) Pedals..........................................................................................29 User Interface ............................................................................................................................................36

Original User Interface (NordicTrack® SL710)........................................................................................36 New User Interface (LabVIEWTM based) ................................................................................................37 Labview Programming ............................................................................................................................44

Timer.vi...............................................................................................................................................45 Resistance.vi ......................................................................................................................................46 RPM.vi ................................................................................................................................................47 HeartRate.vi........................................................................................................................................48

Designing the Layout for a Simple User Interface..................................................................................48 Miscellaneous Additions..........................................................................................................................50 Humans Subjects Testing ........................................................................................................................50

Testing Procedure ..................................................................................................................................51 Testing Results .......................................................................................................................................52

Cost Analysis ............................................................................................................................................53 Future Work/Conclusion ..........................................................................................................................55 References.................................................................................................................................................56

-i-

Executive Summary The design team participated in the 2004-2005 National Student Design Competition. The competition is sponsored by the University of Connecticut, Marquette University, and the Rehabilitation Engineering Research Center on Accessible Medical Instrumentation (RERC on AMI). This competition is open to design teams in biomedical engineering, industrial design, and other disciplines. The competition offers multiple target designs to address, and the design team chose to address the Accessible Ergometer Design (synonymous to exercise bike). The aim of this project was to build a creative cycle ergometer that is usable by individuals with a diversity of disabilities including: Deaf, Blind, Low Vision, Parkinson’s Disease, Stroke, Heart Failure, and Diabetes. To accommodate the various disabilities as well as designing the bike to be universal, several considerations had to be taken into account. Deaf users retain their ability to see and read the user interface display, but any audio output needs to be coupled with text. Blind users would have a hard time adjusting the bike as well as using the user interface. Therefore, any adjustments should be done electronically via the user interface and should be coupled with audio/speech output. Low vision users can use the bike easier by implementing contrasting colors (white on black background), large buttons, and bright lights that help items stand out. People with Parkinson’s disease often have a hard time initiating motion, thus it is necessary for them to have a “target” to shoot for so they can use the bike effectively. Along with Parkinson’s disease, people with stroke, heart failure, and diabetes need to be able to stabilize themselves on the bike and be able to control the amount of exercise they are obtaining. As part of the design competition, the team is allowed a $2000 budget towards building a prototype where up to $500 of that budget can be used towards purchasing an existing device to be modified. Furthermore, the mainstream device must retail for no more than $1000. The team ended up purchasing the NordicTrack® SL710 for $499.99. Some of the technology from the NordicTrack® SL710 kept includes the SMRTM Silent Magnetic Resistance, the CardioGripTM Heart Rate Sensors, the seat, and some of the original steel framing. The rest was eliminated and the design team modified the rest of the bike to make it more accessible to various users. Some of the components implemented includes: zero step-over technology, power seats, lighted pedals, arm motion, and seat assist. Some users have weak lower body strength and cannot step over conventional exercise bikes; therefore, the design team dropped the original track down to ground level and built a ramp for easy access to the seat. Since the track was brought down to ground level, it was important to maintain the original seat height from the ground and distance from the pedals. Therefore, the design team built a track system and a new seat frame. In conventional exercise bikes, the only mode of adjusting the seat requires lifting a pin from the track and having the user employ their muscle to push back on the seat, thereby changing its position on the track. Users with weak lower body strength are unable to do this. The design team implemented power seats which works by way of a 500 lb linear actuator attached to the bottom of the seat system. To assist people with low vision and Parkinson’s disease, the design team built a wireless LED system in which a

-ii-

conventional doorbell circuit was modified to activate an 800 mcd LED on each pedal upon sitting down on the seat. Once situated on the bike, the user has the option to either workout their lower body by pedaling or workout their upper body by using the arm handles. The arm motion exercise is performed via four independent pistons, with two pistons on each side of the bike. Since the pistons allow for resistance in one direction, this component was designed so that when one piston contracts, the other expands. The pistons themselves have a force range of 5 – 200 lbs, while the range of actual exercise force is 3.70 – 93.08 lbs (adjustable by turning a dial). When the user has completed their workout, they can simply hold onto the raised handle bars (CardioGripTM Heart Rate Sensors attached to the ends of these handles) and allow the seat assist to help them raise from a seated position. Again, people with weak lower body strength would have a hard time getting up from a seated position; therefore, the design team attached a 100 lb piston on an angle under the seat padding. When the user first sits down on the bike, the piston contracts and the energy is stored within the piston until the user stands up again. The original user interface was insufficient because it had a small display that would make it hard for low vision users to see, and it was also hard to press a function for the user interface to work properly. Therefore, the design team programmed with LabVIEWTM and created a user interface in which the user can control the timer to count up or count down minutes, see the revolutions per minute (rpm) they are pedaling, output the heart rate from the CardioGripTM Heart Rate Sensors, adjust the resistance setting dial on a 1-10 setting, and start/stop their workout. Careful considerations were made, such as large buttons, contrasting colors, and boxes to group the timer function. Also, a proper color scheme was used to make sure that color blind users can still use the device. The team was only able to build a demo version of the user interface, and due to insufficient advanced programming knowledge, this demo version was self-operating. For full consideration for the competition, human subjects’ testing was required. The design team submitted necessary forms such as the initial application, consent form, general health survey, post-experimental survey, recruitment flyer, and a phone screening form to the University of Wisconsin – Madison Institutional Review Board (IRB) on November 8, 2004. After several rounds of revisions, the design team received approval from the IRB on March 3, 2005. Testing was conducted on April 5-10, 2005 in the University of Wisconsin – Madison Hospital. A total of four experimental and four control subjects were tested. Most users found the device readily accessible with the exception of the ability to adjust the variable resistance pistons. Since there were two pistons on each side of the prototype, participants responded it was tedious to adjust the pistons and found they were often hard to reach since one piston on each side of the bike was located close to the ground. With the completion of Human Subjects Testing, valuable information has been obtained about the design of the device. While subjects and controls in general rated the device high in quality, certain areas were found that require improvement. First, and most importantly, the user interface must be completed. To accommodate blind users, an audio output should be incorporated. This can be done using the LabVIEWTM program, and only requires additional computer programming and the downloading of

-iii-

the program onto the computer. Second, an additional handgrip should be added to the arm motion handles below the current grip to make the overall length of the grip four inches longer. This will allow shorter people a place to grab onto the handles instead of holding onto the bare metal of the arm handle. Third, there is a section of the seat track that could potentially cause injury. A section of the L-steel that was used to construct the track protrudes near the walk-through platform. When the seat is moved forward, this piece could possibly catch the front of a user’s shoe, thereby compressing and injuring the user’s foot. By cutting this small section out of the seat track, this problem can easily be addressed without destroying the track. Fourth, the weight of the device is too heavy for users with disabilities to transport. A solution to this problem is to use a lighter, yet still sufficiently strong material for the frame. With these improvements in mind, our team feels that we have built an exercise device that would be enjoyed by people with various abilities. We believe we have addressed every aspect of the competition put forth to us and that our device would be of great benefit for people who may have trouble using currently existing devices.

1

Abstract The design team is currently participating in the National Design Competition through the University of Connecticut. The goal of this competition is to build an accessible ergometer (exercise bike) for hypothetical persons of various ailments. Team budget is $2000, while the design will retail less than $1000. Team purchased an existing design for $499.99 and made necessary modifications, which includes: seat assist, power seats, arm motion, walkthrough frame for easy access, and lighted pedals. The seat assist was built to help people who can no longer get up unassisted due to insufficient upper or lower body strength. The power seat consisted of a 500 pound linear actuator and was used to allow automatic seat adjustment for users. Lastly, the arm motion includes four independent, variable pistons that provide greater variability in the user’s workout than would linked pistons. The team successfully built a working prototype by the end of the semester. For full consideration in the contest, the team is required to apply for human subject’s approval. Application was submitted to the Institutional Review Board on November 8, 2004 and received feedback on December 1, 2004 to make revisions prior to testing. After necessary revisions, human subject testing was approved on March 3, 2005 and testing was conducted on April 5-10, 2005. Future work includes improving the user interface to allow for a self-operating bike, addressing blind users, and minimizing the weight of the bike. Ultimately, we will create a more ergonomic, universal device to facilitate exercise for patients with various disabilities. Background The design team participated in the 2004-2005 National Student Design Competition. The competition is sponsored by the University of Connecticut, Marquette University, and the Rehabilitation Engineering Research Council on Accessible Medical Instrumentation (RERC on AMI) [18]. This competition is open to design teams in biomedical engineering, industrial design, and other disciplines. The competition offers multiple target designs to address, and the design team chose to address the Accessible Ergometer Design (synonymous to exercise bike). The aim of this project was to build a creative cycle ergometer that is usable by individuals with a diversity of abilities. To enter the competition, a letter of intent was sent to Dr. John Enderle, a professor at the University of Connecticut and contact person for this competition. This letter of intent detailed a proposed solution to the given problem and a timeline of events for the project. Upon acceptance into the competition, the team was given a preliminary budget of $500. Each time $500 was spent, the receipts of purchase were sent to Dr. Enderle for reimbursement. The maximum budget for each team is $2000, but the ergometer must retail for $1000 or less. In addition to the $2000 budget, the team is allowed to allocate up to $500 of that budget towards an existing device (Contest Announcement, Rules, and Letter of Intent found in Appendix A). As part of the competition, a list of six hypothetical clients was given to the team. The team then set out to design an accessible ergometer based on these six clients’ various abilities. The disabilities of the clients include the post-stroke effects of limited arm function and necessity of a cane for walking, diabetes, poor eyesight, blindness,

2

deafness, morbid obesity (BMI over 40), heart failure, generalized low strength and flexibility, and Parkinson’s disease effects. A full description and the names of the six theoretical clients can be found on the contest announcement found in Appendix A. Design Constraints The ergometer must take into account each of the client’s disabilities. To accommodate for the post-stroke symptoms of limited function in one arm, the ergometer arm motion should be independent between the left and right side to allow just one arm to exercise. Additionally, the push and pull resistances should be independently adjustable to accommodate for strength variations between the pulling and pushing muscle groups. To address the requirement of using a cane to walk, the ergometer should have an easily accessible cane holder to allow placement of the cane prior sitting. This cane holder should be strategically placed to be intuitive for a cane user and so that cane is within easy grasp when the user is ready to exit the ergometer. For users with poor eyesight and those who are blind, the user interface will have to have a large LCD screen, well-defined letters and buttons, and ideally an audio output. Obese users will require a structurally stable ergometer that is capable of supporting a 400 pound load. To accommodate heart failure users, the ergometer must be equipped with a method of determining how hard an individual is exercising and a way to warn users of potential harm they are inflicting on themselves from overexertion. To assist users of low strength and flexibility, the seat position and all resistance controllers should be readily accessible and require minimal effort to adjust. For users with Parkinson’s disease, a method of helping initiate movement to place their feet on the foot pedals and hands on the hand grips should be incorporated, as well as a method that will assist these individuals in entering information into the user interface [15].

In addition to addressing the specific disabilities of the theoretical clients, the final prototype must also be universally applicable, that is, usable by those users who have no disabilities. Further technical design constraints can be found in the Product Design Specification in Appendix B. Universal Design Universal design is the design of products and environments to be useable by all people, to the greatest extent possible, without the need for adaptation or specialized design. There are seven main principles of universal design [23]:

1. Equitable Use – The design is useful and marketable to people with diverse abilities.

• Guidelines includes providing the same means of use for all users, avoiding segregating or stigmatizing any users, provisions for privacy, security, and safety easily accessible by all users, and making the design appealing to all users.

3

2. Flexibility in Use – The design accommodates a wide range of individual preferences and abilities

• Guidelines includes providing choice in methods of use, accommodating left- or right-handed access and use, facilitating and user’s accuracy and precision, and providing adaptability to the user’s pace.

3. Simple and Intuitive – Use of the design is easy to understand, regardless of the

user’s experience, knowledge, language skills, or current concentration level. • Guidelines includes eliminating unnecessary complexity, being consistent

with the user’s intuition and expectation, accommodating a wide range of literacy and language skills, arranging information consistent with its importance, and providing effective prompting and feedback during and after task completion.

4. Perceptible Information – The design communicates necessary information

effectively to the user, regardless of ambient conditions or the user’s sensory abilities.

• Guidelines includes using different modes (pictorial, verbal, tactile) for redundant presentation of essential information, providing adequate contrast between essential information and its surroundings, maximizing “legibility” of essential information, differentiating elements in ways that can be described, providing compatibility with a variety of techniques or devices used by people with sensory limitations.

5. Tolerance for Error – The design minimizes hazards and the adverse

consequences of accidental or unintended actions • Guidelines includes arranging elements to minimize hazards and errors,

providing warnings of hazards and errors, providing fail safe features, and discouraging unconscious actions in tasks that require vigilance.

6. Low Physical Effort – The design can be used efficiently and comfortably and

with a minimum of fatigue. • Guidelines includes allowing the user to maintain a neutral body position,

using reasonable operating forces, minimizing repetitive actions, and minimizing sustained physical effort.

7. Size and Space for Approach and Use – Appropriate size and space is provided

for approach, reach, manipulation, and use regardless of user’s body size, posture, and mobility.

• Guidelines includes providing a clear line of sight to important elements for any seated or standing user, making reach to all components comfortable for any seated or standing user, accommodating variations in hand and grip size, and providing adequate space for use of assistive devices or personal assistance.

4

It is important to note that these principles of universal design address only universally usable design, while the practice of design involves more than consideration of mobility. The design team must also incorporate other considerations such as economic, engineering, cultural, gender, and environmental concerns in our design processes. Mainstream Considerations From above, considering those with functional limitations in the overall design process is good for the design process overall. Design which is more accessible to persons with disabilities typically can benefit able-bodied users as well by reducing fatigue, increasing speed and decreasing the number of errors made. One example is to be found in elevator design. Individuals in wheelchairs or on crutches had great difficulty with the large "banks" of elevators present in many buildings. Often the elevator door would open, but before the person in a wheelchair could get to the correct elevator, the door would close. An obvious solution would be for the elevators to stay open for a longer period of time. However, building codes required that a building's floors be visited by the elevators with a specified frequency. If the doors were made to stand open longer, additional elevators would need to be installed in the building to meet the level of service standards. In a building like the Sears Tower, this could result in a substantial portion of the building being consumed by elevators. Creating more accessible designs can also increase the market for many consumer products. With increasing awareness of the accessibility issues, people are beginning to look for more accessible designs. The U.S. government, for example, has recently passed legislation (Section 508 of Public Law 99-506 ) requiring that the General Services Administration develop accessibility guidelines that should apply to all future electronic office equipment acquisitions (purchase or lease). Similar measures are being examined by other countries as well as many school systems and state governments in the U.S. Accessibility features should begin to provide a market edge even in the home market. Although only one in five or six individuals in the United States has a significant functional limitation, a much higher percentage of households have individuals who have functional limitations. Products purchased for use in a household that has even one member with a disability may be more attractive if their design is more accessible. More accessible design will also increase the useful product life of many products purchased by or for individuals who are aging. 95th Percentile Illusion It should be clear that even if elderly and disabled persons are included in the mainstream design process, it is not possible to design all products and devices so that they are usable by all individuals. There will always be a "tail" of individuals who are unable to use a given product [23]. In order to include a sizeable portion of the population in the category of "those who can use a product with little or no difficulty," the 95th percentile data are often used. The problem is that there are no "95th percentile" data for specific designs. Rather, there are only data with regard to individual physical or sensory characteristics. Thus

5

there is 95th percentile data for height, a 95th percentile for vision, hearing, etc. As a result, it is not possible to determine when a product can be used by 95% of the people. It is only possible to estimate when a product can be used by 95% of the population along any one dimension. Since people in the 5% tail for any one dimension (e.g., height) are usually not the same people as the 5% tail along another dimension (e.g., vision) [10], it is possible to design a product using 95th percentile data and end up with a product that can be used by far less than 95% of the population. To illustrate this phenomenon, imagine a mini-population of ten individuals. Ten percent of them (1 of 10) have one short leg, 10% have a visual impairment, 10% have a missing arm, 10% are short and 10% cannot hear. Let's assume that we design a product that required 90th percentile ability along each of the dimensions of height, vision, leg use, arm use, and hearing. In this instance we would end up with a product which was in fact only usable by 50% of this population. This occurs because, although only 10% of this mini-population is limited in any single dimension, different individuals fall into the 10% tail for each dimension and only 50% of the population is within the 90th percentile for all five areas. In real life, the effect is not quite this dramatic, and its calculation is not so simple. First, the percentage of individual with disabilities is less than 10% along any one dimension. Secondly, there is often overlap where one individual would have more than one disability (elderly individuals, for example). On the other hand, there is a much wider range of different individual types of disability. In addition, the data from which the 95th percentiles are calculated often exclude persons with disabilities [10], making the percentage who could use the design(s) smaller than one would first calculate. EZ Access EZ Access combines simple interactive techniques in ways that work together robustly and flexibly to accommodate users. This allows more people to use the product, according to their own ability, preference, or circumstance. For example, a product that has only a touch screen may be difficult or impossible for many people to use. With the addition of just a few buttons and voice output, the product becomes usable by people who cannot see, cannot read, cannot reach the screen, or cannot make fine movements with their arms, hands, or fingers. The addition of captions further extends the product to people who cannot hear (Figure 1).

(a) (b)

Figure 1: (a) 5 button EZ Access keypad [23]. (b) 8 button EZ Access keypad [23].

6

The EZ access includes features such as Voice + 4 button navigation which give complete access to any onscreen controls and content. This feature also provides feedback and information in a logical way such that it can be used by both sighted and non-sighted users. Typical items include onscreen text, images and controls. The Touch Talk lets users touch onscreen text (and graphics), to hear them read (or described) aloud. Button Help provides a way for users to instantly identify any button on the device. At any time, a person can see and/or hear any button's name and status. They can also get more information about what that button can be used for. Layered Help provides context sensitive information about using the device. If a person needs more help, they can press the help button repeatedly, receiving more information each time. Lastly, ShowCaptions provides a visual presentation of any text or sounds created by the device that is not already visually displayed. Accessible Design Principles Two members of the design group enrolled into a one semester biomedical course dealing with the design of products for persons with physical, hearing, sensory or cognitive impairments as well as the design of standard mass market products. Through this course, we became knowledgeable about the different types of impairments and designed ways to solve these problems. We have applied these concepts into the design of the ergometer to make it more accessible to persons with different disabilities. The disabilities addressed in the design of our prototype include: low vision, blind, deaf, Parkinson’s disease, stroke, heart failure, and diabetes. For each of the disabilities listed above, a thorough description of that disease will be given followed by techniques that can be implemented to improve accessibility to those with that disability in using an ergometer. The majority of the information presented in this section was found on the University of Wisconsin Trace Research and Development Center website (http://trace.wisc.edu/) [23]. Low Vision When vision degrades significantly below normal and begins to cause difficulties seeing and carrying out activities that require sight, it is referred to as low vision. Visual impairments can take on a wide variety of forms, including blurred vision, hypersensitivity to glare, tunnel vision, peripheral vision, etc. It can also vary from mild visual impairment to total blindness. Some interventions, such as cataract surgery, can restore sight but still have visual side effects such as reduced focal range. As a result, the problems faced by people with visual impairments vary greatly from individual to individual, as do the particular strategies and assistive technologies that they would use. People with low vision can often increase their visual ability simply by increasing the light level. Increased light level not only helps individuals with visual acuity problems, but also helps individuals with night blindness or lack of light sensitivity. For individuals with sensitivity to glare, the use of diffuse light as well as adjustment of their position relative to windows or other sources of bright light can greatly facilitate their general vision in carrying out a given task. For increasing readability of printed

7

characters, having a high contrast between the printed characters and the background increased their readability, as does using typefaces that have a wider stroke. Computer access strategies for people with low vision primarily involve the enlargement of text. With the advent of graphics-based computers such as the Macintosh and Windows operating systems, the problem of screen enlargement has become easier. Because these systems use pixel-based character display, it is fairly easy to create software that will simply increase the size of the pixels, thus enlarging the image on the screen. This technique results in large letters, up to 16 times normal size, or more. Also, basic screen pan and zoom capabilities can be built into these graphics-based systems, providing at least basic screen enlargement capabilities as a standard part of the system. For the use of appliances for people with residual vision, a primary strategy is to remember the locations and functions of controls. For devices having scales or legends that must be read, applying large, high-contrast, broad-stroke lettering or labels over the top of the existing labels is often done. Blindess As a person’s vision gets so bad that it is of limited use, we begin to refer to the impairment as blindness. The most useful definition of blindness is “a condition in which a person has lost the use of vision for ordinary life purposes, although some residual vision may exist.” Legal blindness is defined as visual acuity (sharpness of vision) of 20/200 or worse in the better eye, after correction, or when the field of vision is less than 20° in the better eye, after correction. An important point worth noting is that many (but not all) people who are “blind” can see something. In other words, they may not see only blackness, but may be able to tell the difference between extreme light and dark conditions, or may have more, but still very limited, visual ability. They key thing that differentiates blindness from low vision, however, is that with blindness vision is so poor that it does not meet most of life’s needs. When designing products for people who are blind, it is important to make provision for gross visual cues (e.g., large white buttons on a dark background or the use of color) to benefit people with some minimal residual vision. For individuals who are unable to read text visually, a number of tactile strategies have been developed. By far the best known is the use of Braille. Braille is a tactile code developed by Louis Braille to represent the letters of the alphabet. There are six dots in a Braille “cell.” The different characters (letters) are formed by the presence of one or more of these six dots. Unfortunately, only about 10% of the people who are blind prefer to use Braille. This is due to a number of factors, but primarily because many individuals lose their sight later in life, when learning Braille is more difficult. Also, some individuals lose the sense of touch in their fingertips at the same time they lose their sight (e.g., from diabetes) and are unable to learn or use Braille. Another tactile mechanism for reading is the use of raised letters. Although this strategy can be used for short labels, it is not generally useful for reading text. In order for letters to be recognized tactually, they must be quite large and fairly distinct. For people who are blind, the primary computer access techniques have involved screen readers. With these software/hardware products, users can have the

8

characters, words, and blocks of text on the screen read aloud using synthetic speech. With the advent of Acrobat Reader 6.0, Adobe has embedded a screen reader of sorts into the Reader software itself. This scaled-down version of a screen reader (more accurately referred to as a text-to-speech synthesizer in this instance) can read aloud the text in nearly all PDF files, even older files that were not created with accessibility in mind [2]. Deafness In discussing hearing impairments, it is useful to distinguish between deafness and less severe hearing loss. Deafness is a profound degree of hearing loss that prevents understanding of auditory information, including speech, through the ear. Normal conversation is approximately 40 to 60 decibels in volume. A person is usually considered deaf when sound must reach at least 90 decibels to be heard at all, and even amplified speech cannot be understood. People with hearing impairments rely much more heavily on sight as well as vibrations for environmental awareness. Unfortunately for these individuals, today many warning and alerting systems are based solely on sound. For individuals who are totally deaf, however, a mechanism that does not depend upon sound is necessary for signaling an emergency situation. This mechanism could include:

1. Small visual indicator or change in the device’s display 2. Turning on an environmental light 3. A bright strobe light or flasher 4. Vibration of a device in contact with the individual 5. Vibrating the environment

In general, hearing impaired users do not have a hard time using certain devices such as computers as along as any audio output is coupled with text that appear on the monitor. Parkinson’s disease Parkinson's disease belongs to a group of conditions called motor system disorders, which are the result of the loss of dopamine-producing brain cells. The four primary symptoms of Parkinson’s disease are tremor, or trembling in hands, arms, legs, jaw, and face; rigidity, or stiffness of the limbs and trunk; bradykinesia, or slowness of movement; and postural instability, or impaired balance and coordination. As these symptoms become more pronounced, patients may have difficulty walking, talking, or completing other simple tasks. Parkinson’s disease usually affects people over the age of 50. Early symptoms of Parkinson’s disease are subtle and occur gradually. In some people the disease progresses more quickly than in others. As the disease progresses, the shaking, or tremor, which affects the majority of Parkinson’s disease patients may begin to interfere with daily activities [15]. When trying to provide effective and efficient interface techniques for people with physical impairments, the first step is to make sure they are able to fully utilize their

9

current physical abilities. Similarly, people who are weak or who have muscle activities that interfere with their control may require stabilizing in order to effectively and efficiently use different interface techniques. The first component of optimal physical control is a stable platform. A person’s trunk needs to be stable in order to have maximum control of the limbs. The most important component of an interface, therefore, is proper and stable seating or positioning. Also, people with Parkinson’s disease sometimes have a hard time initiating motion. Implementing a target for them to achieve has been shown to be beneficial for this impairment. People with Parkinson’s disease tend to have tremors and may cause multiple key activation when using a device. To eliminate spurious key activation, placing a delay on the key acceptance time would be beneficial. A key would then have to be held down for some length of time before it would be accepted (e.g., ½ to 1 second). If an individual accidentally bumped a key on the way to the desired target, it would be ignored since it would not be held down for the required time. When the desired key is pressed, it would have to be held down for the required time before being accepted. When used on a keyboard, the delay activation feature is commonly called “SlowKeys.” Some people have difficulty removing their hand from a control without activating it a second time. A technique to surmount this type of movement and allow them to operate buttons more successfully would be to introduce a period of time after a button is released during which further input on the same button would not be processed. This would eliminate dual activation if the individual bounced while either pressing a button or releasing it. When used with keyboards, this feature is often called “BounceKeys.” Stroke The three main causes of stroke are: thrombosis (blood clot in a blood vessel blocks blood flow past that point), hemorrhage (resulting in bleeding into the brain tissue; associated with high blood pressure or rupture of an aneurism), and embolism (a large clot breaks off and blocks an artery). The response of brain tissue to injury is similar whether the injury results from direct trauma or from stroke. In either case, function in the area of the brain affected either stops altogether or is impaired. In some instances, the individual is left with limited movement in the lower limbs, upper limbs, or one complete side of the body. To allow individuals with motion limited to one side of the body to exercise, the arm motion should be right and left side independent. This would allow users to workout their functional arm without the need to use their impaired arm. Other solutions to this kind of physical impairment parallel with key ideas addressed above in the Parkinson’s disease section. Heart Failure

Heart failure is a progressive disorder in which damage to the heart causes weakening of the cardiovascular system. It is clinically manifested by fluid congestion or inadequate blood flow to tissues. Heart failure progresses by inappropriate responses of the body to heart injury. Heart failure may be the sum of one or many causes. It is a progressive disorder that must be managed in regard to not only the

10

state of the heart, but the condition of the circulation, lungs, neuroendocrine system and other organs as well. Furthermore, when other conditions are present (e.g. kidney dysfunction, hypertension, vascular disease, or diabetes) it can be more of a problem. Finally, the impact it can have on a patient psychologically and socially are important as well [1, 7]. To ensure the safety of users with heart failure, CardioGripTM Heart Rate Sensors were placed on the arm handles of the seat. By allowing the user to monitor their heart rate, they will be able to determine how hard they are working their body and prevent themselves from overexertion. Other possible solutions to this kind of physical impairment also parallel with key ideas addressed above in the Parkinson’s disease section. Diabetes Diabetes is a problem with the body's fuel system. It is caused by the lack of insulin, a hormone made in the pancreas (an organ that secretes enzymes needed for digestion) that is essential for getting energy from food. There are two kinds of diabetes:

• In type 1 diabetes, which usually starts in children, the body stops making insulin

completely. • In type 2 diabetes, also called adult-onset diabetes, the body still makes some

insulin, but cannot use it properly. Most adults with diabetes have type 2; in fact, type 2 diabetes accounts for 90 percent of all diabetes cases [24]. Here’s how insulin works. Food is digested in the stomach and intestines, and carbohydrates are broken down into sugar molecules, or glucose. Glucose is then absorbed into the bloodstream, and blood glucose levels rise. This rise in blood sugar normally signals special cells in the pancreas, called beta cells, to release the right amount of insulin. Insulin allows glucose and other nutrients (such as amino acids from proteins) to enter muscle cells. There, they can be stored for later use or burned for energy. When the body has a problem making insulin or the cells do not respond to insulin in the correct way, diabetes results. One of the common symptoms found with people with diabetes is blurred vision. Techniques such as providing large lettering with high contrast between black and white is beneficial to such users. In addition, some people with diabetes lack the ability to tactically differentiate between various surfaces due to the onset of neuropathy, a complication in which the peripheral nerves degenerate and the sense of touch is diminished or lost. To overcome this, large buttons with distinct corners and contrasting colors will benefit diabetic users.

11

Current Competition Commercial devices that accomplish functions similar to the one of this project include the Schwinn® Airdyne Windjammer UBE and the RST7000 Total Body Recumbent Stepper (Figures 2 and 3, respectively). The Schwinn® Airdyne Windjammer is fully adjustable and has multi-position arms to allow for length adjustment. It consists of a dual drive train in which allows forward and backward motion. The main disadvantage for this machine is that it has a weight capacity of only 300 lbs, insufficient to support our obese patient of 400 lbs and only allows upper body workouts. In addition, the cost of this machine is $2195, far above our maximum allowance of $500.

Figure 2: Schwinn® Airdyne Figure 3: RST7000 Total Body Windjammer UBE [20]. Recumbant Stepper [19]. The RST7000 Total Body Recumbent Stepper is a more versatile device compared to the Schwinn Airdyne Windjammer described above. The Total Body Stepper allows for total body, upper body, or lower body workouts. This is beneficial since all of our clients vary in ability with use of leg/arm motion. By having different modes of exercise, it adds variability to the user’s exercise. The device has a walk through access that allows easy and safe entry for all users. The Total Body Stepper has contact heart rate sensors on the handles. This is an important component since it is vital to display the user’s heart rate during exercise to monitor their workout accordingly. This seems to be the ideal device of the team, but this machine costs $3995.00, which is also beyond our allowable $500 budget. Chosen Device: NordicTrack® SL710 The existing commercial design that was utilized for this project was the NordicTrackTM SL710 (Figure 4). It was chosen because it is a recumbent cycle that incorporated magnetic resistance, EKG/pulse sensors, a console and ergonomic pedal placement. Based on the limitation that an existing device can be purchased for no

12

more than $500, the team decided to purchase the cycle ergometer for $499.99 from Sears.

Figure 4: Nordic TrackTM SL710 [14].

The design team decided on a recumbent ergometer over an upright cycle ergometer due to two main factors: support of the user and stability. A recumbent style exercise cycle allows for the user to have their body supported when seating in a reclined seat compared to a bicycle seat that is used for upright stationary cycles. This reduces the amount of pain that is experienced by people in their lower back. Additionally, a user in a seated position is more stable than a user perched on a raised seat. The magnetic resistance used on this style of NordicTrackTM SL710 cycle is referred to as SMRTM, Silent Magnetic Resistance. The SMRTM system (Figure 5) enables changes in the pedaling resistance by having a metallic flywheel rotating through a magnetic field. As the flywheel passes through a greater portion of the magnetic field, the resistance is increased. The magnetic field is generated by permanent magnets that are mounted on a C-shaped bracket. The bracket is bolted to the frame on one end, which acts as a pivot point. The other end is attached to a cable, whose length is adjusted by an electric motor. In order to increase the resistance, the cable length is decreased. A spring is used to return the bracket to the lowest resistance setting and an adjustable bump-stop is used to prevent the bracket from making contact with the flywheel. The bump-stop also acts to control the maximum achievable resistance setting. Since the magnetic resistance design was exactly what was needed for our resistance design, we decided to keep it intact and utilize it along with the existing pedal drive train for rotating the flywheel.

13

Figure 5: Sketch of Flywheel/Magnetic Resistance System [14].

Another aspect of the NordicTrackTM SL710 that was left basically unchanged for our prototype was the CardioGripTM EKG/pulse sensors (Figure 6). These sensors work by detecting the EKG through the metallic conducting palm sensors and then relaying the signal back to the computer where the heart rate can be calculated and sent to the display. This pulse detecting system is not as accurate as other methods such as pulse oximetry or a telemetry strap, but it can be used by a wide variety of people with very low strength and dexterity and thus was appropriate for the prototype.

Figure 6: CardioGripTM EKG/pulse sensors.

One final aspect of the commercial bike was deemed sufficient for a final prototype, the seat. The original seat padding and general shape was left intact and the angle between the seat base and seat back was left unchanged. One small addition was made to the seat base, which will be addressed later. Modifications to Existing Device The first concern addressed when modifying the commercially available bike for general accessibility was the incorporation of a walkthrough access for zero step-over technology (Figure 7). Zero step-over technology means that the user would not have to lift a leg and maintain balance on one foot to get onto the bike.

14

Figure 7: Zero step-over technology access walkway.

The implementation of a zero step-over technology led to several other small

changes to the commercial bike. First, the one way clutch on the pedal linkage was reversed so that the entire pedal-magnetic resistance system could also be reversed, yet still function normally. The location of the one way clutch can be seen highlighted in Figure 8. Reversal of the pedal-magnetic resistance system netted an increase of 10 inches in terms of walk-in space and foot room.

Figure 8: Highlighted one way clutch that was reversed [14].

An increase of only 10 inches was not quite sufficient since an average foot length is approximately 12 inches long. Therefore, the original manual seat locating system was eliminated and replaced with a power seat system that was mounted lower than the original system. By lowering the seat locating track, an additional two inches of walkthrough space was gained, bringing the total walkthrough space to 12 inches, an amount of room deemed sufficient for easy access to the ergometer.

15

Additions to existing device In addition to the modifications noted above, the design team added additional components to the bike to make it more accessible to users with various disabilities. Power Seats As mentioned previously, the original manual seat locating system was eliminated in favor of a powered seat locating system. The powered seat locating system was made in such a way that the seat would travel through the same path in space with the new system that it traveled through with the original system. Thus the angle of the seat track was maintained at approximately the original 22.6° and the seat was placed with exactly the same reference to the pedals as it was originally. A new mounting platform for the seat and the arm motion was created utilizing 1 ½ inch x 2 ½ inch x 3/16 inch thick rectangular tube fashioned into a reinforced rectangle base with outside dimensions 20 inch x 13 inch. On the outside of the new seat platform, 2.17 inch rollerblade wheels were attached that would ride on the new track system. Additionally, 1 1/4 inch secondary wheels were attached to the bottom of the seat platform. The 1 1/4 inch wheels ride below the seat track to keep the seat platform locked to the track in the same way a roller coaster is locked to its track. The new track system was constructed from 1 ½ inch x 1 ½ inch x 3/16 inch thick angle iron set to the width of the seat platform and the proper length to cover the full range of travel of the linear actuator used to control the seat motion. Further detailed discussion of the seat platform and its design are addressed later in this report. To control the movement of the seat platform, a 500lb 12/24V electric linear actuator with a 1 ft stroke was used. The 500lb force is more than what is necessary for a final marketable version, but was chosen for the prototype for its cheap availability. On a final version, an actuator with half as much force would be sufficient. The actuator was mounted between the rear of the seat track and the front of the seat platform, as seen in Figure 9. This mounting position yielded the greatest forward travel for the seat by allowing the seat track to be as low to the ground as possible and also lead to the actuator being tucked under the seat platform when in the maximum retracted position. By concealing the actuator under the seat in the retracted position the overall length of the ergometer was decreased.

16

Figure 9: Linear actuator mounting location.

The limiting strength link in the linear actuator system was determined to be the mounting bolts at each end. Since the 500 lb load would be shared between both mounting bolts, only 250 lb would contribute to shear on the bolt. This load results in a shear of 1.3 ksi, much less than the critical shear value for a Grade 5 bolt of 75 ksi [17]. The Free Body Diagram used for this calculation is seen below in Figure 10 and the full calculations can be found in Appendix J.

Figure 10: FBD of linear actuator bolt in direct shear.

Once the locating system for the seat was established, improvements for the seat itself could also be addressed. Since it can be difficult for some users to stand from a fully seated position due to insufficient upper or lower body strength, a lift assist was incorporated into the seat. The bottom cushion of the seat was removed and a new elbowed mount was placed under the bottom cushion. Across the elbow and at the calculated location, a 100 lb. pressurized lift cylinder was mounted (Figure 11) [25].

V = 500 lbs V = 500 lbs

½” Diameter Grade 5 bolts

17

Figure 11: Lift seat configuration.

As the seat travels up, a portion of the 100 lbs. the lift cylinder can generate is exerted to aid the user in getting to a standing position. Additionally, a spring was added to the force cylinder so that as the seat angle approached 0°, the force generated by the spring would add to the total force exerted and the piston force would not go to 0 lbs. The lift force generated follows the graph seen in Figure 12. A limiting chain was added to allow the seat to achieve an angle no greater than 45° for ease of sitting down.

Fpiston vs. Seat Angle

0

10

20

30

40

50

60

70

80

0 5 10 15 20 25 30 35 40 45Seat Angle (degrees)

Fpis

ton

(lbs)

Figure 12: Lift force as a function of seat angle.

18

When in the extended position, the lift spring puts an even load of 100 lbs on all three of the pivot rods due to force conversion along links. Since the critical load for this Grade 5 rod would be 8283 lbs, the rods are not in a near failure mode. When the seat assist is in the compressed position, the additional force of the compressed springs is added to the shearing force on the three pivot rods. When compressed, the load on the rods is 127.8 lbs, which is, again, much less than the critical load of 8283 lbs.

The tension that is carried on the restraint chain when in the extended position was also determined. Based on the portion of the 100 lb cylinder that would be transferred to the chain and also the multiplication effect of the lever arm, the restraint chain would be subjected to 138.5 lbs of tension. Since a 150 lb load chain was used to restrain the seat, the chain was found to be close to, but not over, its load limit. The Free Body Diagrams used for these calculations are seen below in Fig. 13 and the full calculations can be found in Appendix J.

Figure 13: FBD’s of Seat Assist System in the extended and compressed positions. Track system Important considerations were first made before designing the seat frame for the seat and arm motion. The team took several measurements including:

• Distance of the foot pedal axle to the closest seat position (32.0 in.) • Distance of the foot pedal axle to the farthest seat position (43.5 in.) • Distance of the foot pedal axle from the ground (14.875 in.) • Distance of the seat from the ground (16.75 in.) • Angle of inclination (22.6 degrees)

As mentioned previously, the team decided to drop the entire track system to

near ground level compared to the original NordicTrack design which was approximately 12 inches off the ground at its lowest point. The lowering of the track

T12”

3/8” Diameter Grade 5 bolt

3/8” Diameter Grade 5 bolt

5” 5 1/2”

100 lb

Shear ~ Equal

F

F = 100 lb + Spring Force

19

system was accomplished by first cutting 8 feet of steel L-beams creating a 3 foot x 1 foot rectangle (Figure 11).

Figure 14: Dimensions of track system (welded out 8 feet of steel L-Beams).

With the track system created, it was welded to the rear support of the existing device and feet pads were placed at the front to provide support. The track system was created with attempts to emulate the original angle of elevation. The final inclination resulted in 21 degrees. Seat Platform The next task was to develop the seat platform. The platform had to be strong enough to be pulled and/or pushed by the 500 lb linear actuator and support the downward force of the seat and user. The design of the seat platform was accomplished by creating the base using 6 feet of 2 ½ inch x 1 ½ inch rectangular, hollow steel bars. The base was welded together to form a 20 inch x 13 inch rectangle (Figure 15).

36 inches

12 inches

20

Figure 15: Top view of seat platform. Dark lines indicate location of the weld.

This frame serves as the foundation on which the seat support and arm motion

mounts. The seat platform rolls on the seat track on eight 55 mm bearing-filled rollerblade wheels (four on each side). In addition to rollerblade wheels providing support on top of the track, four additional 1 ¼ inch plastic wheels (two on each side) roll along the bottom of the track to ensure the platform remains on the track system. These wheels are attached to the platform through a secondary mounting system utilizing ½ in. threaded rod. The final head-on view of the seat base is diagrammed below (Figure 16).

Figure 16: Head-on view of seat platform.

Under the normal loading condition, the 400 lb load of the user is split between

two sides of the seat track, each of which has 4 main wheels. Each wheel, or more specifically its mounting bolt, therefore carries a load of 50 lbs and a shear of 1.02 ksi. Since Grade 8 bolts where used, the critical shear is 91 ksi (Potter) and the shear on the mounting bolts is much less than the critical shear.

L-Beam

U - Bolt

Rollerblade wheels with

10”

Stop Nuts

Seat Base Seat Base

10 inches

13 inches

Plastic Wheels

20 inches

21

For the situation where the user will apply a reverse load on the seat track, Ftheory in Fig. 17, the shear on the farthest rear wheel, wheel A, and the shear on the farthest forward under wheel, wheel B, become important. The critical load for either of these wheel bolts is 4467 lbs, and from this value, Ftheory was calculated. For failure to occur at wheel B, the user would need to apply an Ftheory of 3449 lbs, which is highly unlikely if not impossible. For failure to occur at wheel A, the user would need to apply an Ftheory of 2931 lbs, which is, again, highly unlikely if not impossible. The Free Body Diagram used for all these calculations is seen below in Fig. 17 and the full calculations can be found in Appendix J.

Figure 17: FBD of seat track support wheels.

After initial welding of the seat platform and attaching the rollerblade wheels, the

platform turned out to be unbalanced. There was 1/16 inch difference between the front and back of the platform without the wheels. With the wheels, the difference increased to 3/16 inches. Since any fluctuation in the platform is not safe for a user to sit on, we decided to drill the holes again on the opposite side of the beams. The position of the new holes was determined with respect to level ground. In the end, we were able to level the seat platform and have it roll straight up and down the track. The final set-up of the seat platform on the track system and linear actuator is shown below (Figure 18). The final angle of elevation changed to 24° with respect to the ground due to the uneven seat base on the track system.

R1

R2 200 lb

15”

7.5” 0.25” Grade 8 bolts

1.75”

3”

17”

FTHEORY

22

Figure 18: Side view of track system, seat platform and linear actuator.

Seat Supports Since the team decided to drop the track system to the ground level, a seat support system is necessary to maintain the seat height at 16.75 inches above the ground. In order to do this, the team designed a support emanating from each of the three crossbars in the seat platform. This served a two fold purpose. First, the team had to ensure the user’s safety; therefore, the seat must fit securely on the seat support. Second, the center of gravity of the seat itself should be properly situated to make certain that the seat does not sway back and forth upon user movement in the seat. With these criterions, the team decided the optimal support system would consist of a single bar emanating from each of the crossbars in the seat platform. The team purchased eight feet of a 1 ½ in. square steel tubing to construct the support system. Each of the steel pieces was welded together, with the final dimensions shown below (Figure 19).

17.3348”

36”

3.5”

Seat platform

Linear Actuator

Ground Level

23

Figure 19: Seat support system with dimensions.

Seat assist situates above the 10 ½” bar.

The design of the seat frame system governs that many different areas be considered for analysis. First, the bending moment on the cantilevered portion was found, then, also, the shear across the angled bottoms of the vertical support legs was found, and finally the buckling of the vertical support legs was considered.

The portion of the 400 lb load that is applied to the cantilever was found to be 155.6 lbs. Under this load, the cantilever experiences an bending moment of 544.4 in*lbs. Since the cantilever is made of 1 in square tubing with a 1/8 in wall thickness, the maximum tension in the cantilever is 544.4 lbs, which is far below the critical tension for mild steel which is on the order of 1000’s of pounds.

The vertical pillar supports are composed of 1.5 in square tubing with a ¼ in wall thickness. From this information the shear area of each pillar was found to be 1.25 in2 for a normal cut. Since the pillar legs sit at angles, the shear area is increased. The two rear-most pillars sit at an 11o angle, meaning that their true shear area is 1.27 in2. The forward-most pillar sits at a 26o angle, meaning that its true shear area is 1.39 in2. Since the rear-most pillars have less area for shear, they are the limiting factor and the shear on each one was found to be 315 psi. Since the critical shear for mild steel is on the order of ksi, the vertical pillars are not in danger of failing.

The Free Body Diagram used for these calculations is seen below in Fig. 20 and the full calculations can be found in Appendix J.

13.25” 9.5”

5.25”

10.5”

24

Figure 20: FBD of seat frame.

Arm Motion Exercise

The last component in the entire seat system consists of the arm motion design. Although not required as part of the competition, the team felt that introducing arm motion to the design would add greater variability to the device. Along with lower body workout by the cycling motion of the foot pedals, users can also obtain an upper body workout using the arm motion. Furthermore, both components (arm motion and the foot pedals) can be performed to allow for a total body workout. The upper body and lower body workout are independent of each other, allowing the user total control of his/her exercise routine.

The initial design of the arm motion consisted of three feet of 1 ½ inch square steel tubing left over from seat support design, five feet of 1 5/16 inch circular steel hollow rod, and two independent, variable resistance pistons. The team determined the optimal placement of the pistons with respect to the seat base to allow for maximum moment endured by the user. It was determined that the insertion point of the piston should be 9.25 inches above the insertion point of the 1 ½ inch square steel in the seat base. Furthermore, a six inch piece of 1½ inch square tubing was welded outwards from the ten inch 1 ½” square tubing to bypass the width of the seat. Lastly, 2 ½ feet of hollow rod was inserted into the six inch bar upon 1300 pounds of pressure and welded together. Using independent pistons is unique in the sense that it allows patients with limited one arm motion to conduct exercise with only their functional arm and not worry about the unused handle coming back at them. The pistons were purchased from a health fitness dealer, HealthFX America [4], for $24.45 each including shipping. The pistons are capable of 5-200 lb loads with adjustable dials from 1 to 12 (increasing number on the dial corresponds to higher resistance). Additionally, brass fittings were

25

placed within the piston holes and are meant to slide over the screws to reduce friction and wear on rotating components. The team finished the arm motion by welding the ends of the handles from the NordicTrack SL710 onto the ends of the hollow bar to allow for a more ergonomic feel upon upper body exercise. The initial setup is diagrammed in Figure 21.

Figure 21: Original arm motion exercise. Single pistons used allowed resistance in only one

direction.

Upon completion of the first generation prototype, it was determined improvements were needed in the design of the arm motion. First, the two variable resistance pistons provided resistance in only one direction. Resistance was felt when pushing on the arm handles, but not when pulling on the arm handles. Second, the initial extensions that were inserted to allow the arm handles to clear the width of the seat interfered with the cycling motion of the user’s foot. If the arm motion handle was pushed forward when the foot was in the closest position of the cycle to the seat, the heel of the user’s foot would contact the arm handle extension.

To accommodate for these problems, adjustments were made to the original design. The extension that was interfering with the user’s foot was lowered from 9.25 inches above the seat base to 6.75 above the seat base and extended outwards an additional 2 inches, thereby allowing the heel of the foot to pass over the extension without making contact. To accommodate for the lack of dual resistance, two more identical pistons were purchased from HealthFX America [4] so that in the current design four independent, variable resistance pistons control the arm motion. Two pistons are attached to each handle to provide both push and pull resistance. The pivot point was moved from the bottom of the arm handle to 5 inches above the insertion point of the 1 ½ inch square steel in the seat base. This allowed for the necessary placement of the pistons to create a dual resistance system. One piston is connected

12.5”

9.25”

26

4.5 inches below the pivot point and the other piston is connected 4 inches above the pivot point. Each piston can be separately adjusted to a different resistance level, allowing the level of difficulty of the exercise to be precisely regulated by the user. The free body diagrams of the arm motion in the back, upright, and forward positions are shown in Figure 22. It was found that the minimum force a user would be required to exert on the arm handle in the back position is only 3.70 pounds, while the maximum force can be up to 93.08 pounds. In the upright position, the minimum force is 3.59 pounds and the maximum force is 80.74 pounds. Finally, in the forward position the minimum force is 3.68 pounds and the maximum force is 72.75 pounds.

Figure 22: Free body diagram of arm motion. Fuser is the force exerted by the user, and Fpiston 1

and Fpiston 2 is the force of the piston resisting the motion induced by the user. To ensure the integrity of the arm motion design, an engineering analysis was conducted on the most critical point of the system, the back bolt where each of the pistons are attached to the seat frame. The maximum allowable torque on this ½” bolt is found using the equation Tmax = K*L*d, where Tmax is the maximum allowable torque in lb*inch, K is the dynamic coefficient of friction (dimensionless), L is the clamp load in pounds (equal to the product of the minimum yield strength of the material and the stress area of the screw), and d is the nominal screw diameter (inches) [22]. For the ½” bolt, the minimum yield strength is 92,000 psi [7], K = .15, L = (92,000)(0.25)2π = 18,032 lbs, and d = 0.5 inches. This gives a maximum allowable torque of 1352 lb*inch. The actual torque on the bolt is found using the free body diagram shown in Figure 23, which is a view of the bolt from the top. It was found that the maximum torque that could be applied to the bolt is 62.75 lb*inch, well below the maximum allowable level. The calculations performed to determine the force required to move the arm motion handles and the torque applied to the bolt can be found in Appendix J.

2.92 feet

27

Figure 23: View of ½” bolt from above. Fpiston1 and Fpiston2 is the force of the piston on the

bolt and R is the reaction force at the point where the bolt is attached to the seat frame.

Figure 24: Final arm motion with dual pistons design. Two pistons were used since they only

provide resistance in one direction.

Arm Handles After designing the seat assist, it became clear that the handles attached to the seat were no longer in the correct placement. This is because when sitting on the seat if you tried to push off the handles, your arms were already at full extension. Therefore, in order to achieve better leverage and the ability to apply force, the handles were

28

raised four and a half inches. This was done by putting two bends into the handles (Figure 25), one that bends the bar vertically and one that bends it horizontally.

Figure 25: Raised arm handles with heart rate sensors

The distance from the seat to the first bend (horizontally) is 7 inches, the distance

to the second bend (vertically) is 4.5 inches and the length of the handle from that bend to the end of the handle is 16 inches. The handles were mounted on a pivoting system to allow each handle to be pivoted out of the way of the user to allow uninhibited access to the seat. The pivot system consists of two bolts that pass through the handle bars. Each bolt rides in a notch that allows a 1/8 rotation of the handle. When each of the 1/8 rotations occur together, the handle can travel through a ¼ rotation and rotate up and out of the way of the user.

The limiting link in the EKG handle rotation was determined to be the bolts on which the pivot rides. These bolts are in shear, but it is shear induced by the moment around the pivot. Assuming the maximum weight rider (400 lbs) pushes evenly with both hands and the full body weight, the load applied to the handle will be 200 lbs. The moment is then found and from this the shear found across the bolts. Since there are two bolts per handle, the shear is distributed across 4 bolt faces. This means that the shear on an individual face is 10.8 ksi, which is, again, much less than the critical shear for a Grade 5 bolt of 75 ksi (Potter). The Free Body Diagram used for this calculation is seen below in Fig. 26 and the full calculations can be found in Appendix J.

4.5”

16”

29

Figure 26: FBD of EKG handle bolts in indirect shear.

Another modification to the handles was putting the pulse rate sensors on the

handles by the seat rather than by the user console or on the arm motion handles. The placement of the handles puts them within easier reach of the user than if they were by the user console and also eliminates potential motion artifact in the EKG signal that would come if they were placed on the arm motion handles.

Wireless Light Emitting Diode (LED) Pedals

In an attempt to design the ergometer within the principles of universal design we examined possibilities to help individuals with Parkinson’s disease. Some people with Parkinson’s disease have an inability to initiate motion, which means that if you tell them to take a step they cannot do so, but if you tell them to step over a line drawn on the floor they are able to do this task. We took this same principle and applied it to the pedals. We theorize that by putting red LEDs into the pedals and couple that with audio output instructing the user to place their feet on the pedals will help user’s with Parkinson’s to overcome the problem of initiating motion.

Another advantage of incorporating wireless LED pedals is that people with low

vision or nearly blind will be able to see the pedals easier when the exercise bike is on a dark floor. This advantage was evident during human subjects testing when the user took their feet off the pedals and without the LEDs could not find the pedals once again. It is important to note that legal blindness is partly defined as having 20/200 acuity in the best eye after correction. Therefore at 20/200 it is still possible to see outlines and brightness differences of objects. Figure 27 shows the LED illuminated on the pedal.

M

200 lb

13 ¼”

M1 ¼”

30

Figure 27: Illuminated left pedal.

A major problem to be overcome with this is the fact that the pedals are continuously rotating during pedaling so it is impossible to run wires into the pedals. Some solutions that were proposed include:

1. Having the wires on a retractor so that as the pedal rotates the wires get shorter

as the pedal nears the floor. This idea causes some problems because the retractor might not be quick enough for fast pedal speeds. Also, because it is mechanical, problems and failures might occur which could injure the user of the bike if the wires became entangled on the pedals.

2. Using some sort of rotating junction that is filled with mercury to conduct the electrical signals to the pedals. This idea is highly complicated in the fact that major design changes would have to be made to the pedals. Other issues with this are safety due to the use of mercury and the cost of special mercury junctions.

3. Having all power self contained in the pedals and the ability to turn the LED on and off must be done wirelessly. Ideally all that would be contained in the pedal would be a LED, battery and the wireless receiver. The wireless transmitter would be connected to the seat so that when a user sits down it would send a signal to the receiver. The receiver would be connected to a timing circuit that is used to control the length of time the LED is illuminated.

31

2 ‘C’ Batteries

Wireless Doorbell Receiver

Monostable 555 Timer Circuit

Wireless Doorbell Transmitter

+3 VDC

+3 VDC

Ground

680 nm LED and 680 ohm resistor

assembly

555 Timer Output

Speaker negative output

Flow Block Diagram of Wireless LED pedal assembly

Ground

Ground

Figure 28: Flow Block Diagram of Wireless LED Pedal assembly

The design that was ultimately selected was the wireless design, mainly due to the ease of construction and cost. The initial design was built off a wireless doorbell system along with a monostable 555 timer circuit, an astable 555 timer circuit and an AND gate [21]. The wireless doorbell system (ACE® Wireless Chime Model 3014727 (AC-6150), Appendix E for full specifications) is modified so instead of using the speaker as the output, it is using the wires leading to the speaker as the output for the system. The negative output to the speaker can be used as a digital logic zero. Therefore if the negative output is inputted into a monostable 555 timer circuit (Figure 29), it would trigger the circuit and provide a digital logic “1” or Vcc for a set number of seconds. The period of logic “1” is determined by the values of the resistor and capacitor. In order for the period, T, to be 180 seconds, which would allow 3 minutes for the user to get their feet onto the pedals, the resistor value would need to be 1.64 MΩ and the capacitor would be 100 μF.

32

Figure 29: An input (trigger) of zero results in Vcc for T seconds, the value of T is determined

the values of R and C [12].

The astable 555 timer circuit is needed to make a square wave so that the LED blinks a determined frequency. It should be noted that the astable circuit could be removed since there are integrated circuit LED that blink at a predetermined frequency, so if an LED could be found that has the desired frequency the astable circuit could be removed as well as the AND gate. The astable circuit (Figure 30) was designed to have a frequency of approximately 700 MHz. In order to do this the circuit utilizes two resistors and a capacitor to change the frequency of the square wave. The astable circuit does not need a trigger so therefore it is always oscillating. The values chosen for the circuit were: RA = 100 kΩ, RB = 160 kΩ and C = 4.7 μF. These values result in a frequency of 729.5 MHz. As can be seen in the accompanying graph to Figure 30 the period of time the waveform is equal to Vcc is longer than when it is equal to zero. The values of TH (time when LED is illuminated) and TL (time when LED is not illuminated) are 846.846 ms and 521.136 ms respectively, as shown in Figure 30.

Figure 30: Astable 555 Timer Circuit frequency = 729.5 mHz [12].

33

The final part of the circuit before the LED circuit is the AND gate (Figure 31). The AND gate is used because it takes the output from both the monostable and astable circuits and outputs Vcc only when both the inputs for the AND gate is equal to Vcc. Accompanying the representation of an AND gate is the truth table for an AND gate in Figure 31. The LED circuit is composed of a 680Ω resistor and a red (680 nm) LED, with the resistor connected to the output of the AND gate and the other side of the LED connected to ground.

Figure 31: AND Gate and according truth table [10, 19].

The wireless doorbell currently runs on two ‘C’ batteries which is equivalent to about 3 volts. The 555 timer circuits and the AND gate are designed to operate at 5 volts, though at 3 volts both circuits are still above the minimum voltage needed, as manufacturer specified. In order for this design to work, a battery would have to be placed on the underside of the pedal. This could be problematic for the self righting pedals, except for the fact that currently there is a piece of steel rod that is used to keep the pedals with the foot surface upright. That piece of steel rod could be replaced by the battery that will provide similar results as the steel rod and make the pedals to be self righting. The final design of the wireless circuit includes the wireless doorbell receiver, the monostable 555 timer circuit described above (which lights the LED for approximately five minutes), two C batteries, 680Ω and a 680 nm LED with a brightness of 800 mcd (the specifications of the 555 timer circuit and LED can be found in Appendix H). The major change made is the removal of the astable 555 timer circuit as well as the AND logic gate. The LED chosen is not a blinking LED, though it is hoped that if this design was incorporated a suitable integrated circuit (IC) LED could be found at the desired frequency and brightness. The team was not able to find an LED that matched the desired properties and settled on a LED with good brightness and a somewhat narrow range of sight, so as not to distract non-users in the same room. The monostable circuit was changed to provide a longer time (about 5 minutes) in order to allow for users more time to get situated and work with the user interface. The finished circuit (Figure 32), before being attached to the bottom of a pedal, shows the three circuit boards and two battery holders. The monostable circuit was made so that the 555 timer IC could be easily replaced.

X Y X and Y 0 0 0 0 1 0 1 0 0 1 1 1

34

Figure 32: Wireless receiver LED circuit.

The wireless LED circuit was modified in order for it to be attached to the bottom

of a pedal (Figure 33a and 33b). The circuit boards were attached vertically and the LED circuit board was replaced with an LED holder. Also shown in Figures 33a and 33b is the protective casing used to protect the circuits from being “kicked” during use or if the pedal is spun from being stepped on.

(a) (b)

Figure 33: (a) Underside of left pedal with wireless receiver LED circuit attached. (b) Side view of left pedal, with protective plastic.

Figures 34a and 34b show the placement of the wireless trigger under the seat.

The trigger is protected from large forces by incorporating foam into the tapping device to absorb excess force. The greater the mass of the person sitting in the seat, the greater the force applied to the trigger and the greater the amount of force that is absorbed.

35

(a) (b)

Figure 34: (a) Wireless transmitter placement. (b) Close up of transmitter.

Possible improvements to the wireless LED pedals before a commercial product is produced include:

1. Examining ways to make the circuit run off the batteries for the longest period of time possible.

2. Making the receiver and monostable 555 timer circuit on one circuit board. 3. Improving the underside pedal protection. This could be done with a plastic

molding cap that could be attached with screws so that it can be removed to change the batteries.

4. Sink the LED holder into the pedal so that it does not stick up above the pedal. This would help a shoeless user not feel the LED holder sticking up from the pedal.

5. Replace the current LED with an IC LED that blinks at a frequency of about 600 mHz.

6. Replace the transmitter turn on depressible switch with something that is non mechanical, such as a pressure sensor, so that at a high enough pressure on the seat the LEDs would turn on.

7. Use the transmitter to also turn on the user interface (see below for description).

a. This can be done by having the digital logic “1” be used to turn on the user interface instead of the press-able power switch.

The wireless LED system could also be used in standard bicycles by putting the

transmitter in the seat with some sort of sensor that starts the transmitter when the individual sits on the seat. Blinking LED lights in this case would be used as a way for bike users to indicate to motorists their presence on the road.

36

User Interface Original User Interface (NordicTrack® SL710)

The original user interface that was part of the NordicTrack SL710 was initially

left alone during our first semester of work. The console (Figure 35) of the NordicTrack® SL710 has a NavigatorTM LCD console that is iFit® compatible and includes: a CoolAireTM Workout Fan, water bottle holders and a book rack. The model that we purchased had three LCD displays. The console displays time, speed, revolutions per minute (rpm), and distance pedaled. Another LCD shows the training zones and shows a graphic representation of a ¼ mile track, so that a user would know where on the track they were presently traveling. The last display shows the user heart rate, fat burned, calories burned, and the current resistance level. For all the displays, the values shown would switch between the ones shown on the display.

Figure 35: Control panel of the NordicTrack® SL710 [14].

Other aspects of the console left intact were the water bottle holders, the bookrack and the fan. The fan could be controlled by pushing on a single button to toggle through low, high and off settings. Other buttons on the console include: numbers 1 through 10 which are used to select the resistance level, program select, program start, and an iFit.com button (for a full description of the buttons, see Appendix D).

Within the system, there are eight built-in workouts that can be selected. There are six workouts that work using resistance and pace while the other two work by using the heart rate. The six workouts are: trail blazer, biker’s choice, victory hill, competitor’s challenge, winner’s pace and power drive. The first two are targeted towards weight loss, the second two are for aerobic workouts and the last two are for performance exercise routines. All the programs adjust the resistance or prompt the user to change their pace to simulate the program. Thus, for a large hill, the program will increase the resistance when going up the hill and decrease it for the descent. The other two programs work by using the user’s age and calculating a maximum heart rate by subtracting the age from 220. The two programs work to maintain your heart rate at

37

80% of maximum or 85% of maximum by adjusting the resistance level. The console also can be run by using the iFit® mode where special programs that are available on CD, video, and the internet control the resistance level of the exercise bike.

The NordicTrack® SL710 also includes CardioGripTM pulse sensors (Figure 36). These sensors work by detecting the pulse rate through the metallic, conducting palm sensors and then relaying the signal back to the console, where the heart rate is then displayed. This heart system is not as accurate as other methods such as pulse oximetry or a telemetry strap, but it can be used by a wide variety of people with very low strength and dexterity.

Figure 36: CardioGripTM pulse sensors.

The original design had some major problems with it, including:

1. Small LCD displays a. Hard to see the values displayed

2. LCD displays change between up to four different values a. Hard to see the bar designating which value it was showing b. Cognitively difficult to understand

3. Small lettering for instructions a. Hard to read small lettering for low vision users

4. Hard to use number input pad a. No raised buttons b. No physical feedback from button when it is being pushed c. No way to stop people with tremors from pressing the button twice in

sequence.

Due to these problems, as a group we discussed making our user interface more easily understood by people with low vision and cognitive problems.

New User Interface (LabVIEWTM based)

The design criteria for the user interface for the competition is that it needs to be

easy to view and manipulate. In order to make the user interface universally designed, a more detailed description of its requirements needs to be made. The list of features that the user interface should incorporate to be universally designed includes:

1. Large lettering 2. High contrast 3. Large buttons/controls

38

4. Easy to understand terminology 5. Simple controls 6. Be able to reach and adjust controls easy 7. Advanced layout and Simple layout – for power users 8. Screen reader 9. Help button 10. Audio output of controls when changed 11. Multiple frequency tones used for warnings and audio indicators

Large lettering, high contrast, large buttons and controls will help users that have low vision be able to see the labels easily and help mainstream users be able to quickly see labels without searching. Easy to understand terminology and simple controls will enable all users to quickly understand the controls of the user interface without having to read the user manual or have someone explain the controls to them. By having the controls able to be easily reached it will enable users the ability to change items during exercise without difficulty. By having an advanced layout and simple layout options it enables for users that just want to get an exercise the ability to use the ergometer without worrying about having to set any options. The screen reader and help button would enable blind users to understand the user interface, as well as help cognitively challenged users navigate the interface. The audio output is an option that should be able to be turned off so that normal users are not bothered by the interface. The multiple frequency tones enable people that are hard of hearing to hear the indicators. Some people who have trouble hearing can still hear distinct frequency ranges, so by having a multiple frequency tone it enables all users with any residual hearing function to hear the tone. Possible solutions to make the user interface more accessible that were decided not to be used include:

1. Braille a. Only 10-20% of blind people in the United States can read Braille b. Low vision and Blindness can be progressive, and thus is higher in

numbers in older people. These same people do not learn Braille as result.

c. As in (b), older people also have increased problems with diabetes. People with diabetes can lose some sensation in their fingers, a complication known as neuropathy, and as result would not be able to read Braille.

d. For the dynamic displays (such as time), a dynamic Braille display would need to be used. This type of assistive technology is expensive and would not fit within the budget of this project.

2. Sign language display a. Deaf users who use American Sign Language would be able to

understand the simple commands and numeral displays that are in English. Thus, it would not be of much benefit to have a sign language visual display.

39

3. Voice activated control a. This was not included because the technology would need to be more

advanced to respond better to normal speech before being incorporated into this device.

The main settings and readings that the user interface needs to have displayed include: rotations per minute (rpm) of the pedals, heart rate of the user, time exercised, a method for pausing time, a method to start time, a method to change time and a method to change the resistance level. Other settings that could be incorporated into an advanced window of the user interface would include: miles per hour, calories burned (which then would have to take in age and weight), watts, and different preset programs to choose from. These items are considered advanced because these items are not required for an individual to get an adequate exercise. These items would be important for the mass market aspect of selling the product. The reasons that calories burned, miles per hour, miles traveled and watts are considered advanced items is because:

1. Calories burned is only an estimate of the number of calories burned based

on an equation of age, weight and time spent exercising. The number of calories burned can be overestimated by as much as 30% on standard exercise bikes [3].

2. Miles per hour and miles traveled are inheritably linked; the problem with them is they are again not an accurate measurement when done on an exercise bike. This is because normally on a standard bike the rpm and circumference of the front wheel is used to calculate how far you have gone over a certain period of time to calculate miles per hour. However, with the exercise bike the measurement of the rpm is that of the pedals. It would be possible to use an equation to get the miles in distance. This is not accurate because it is very difficult to precisely equate resistance level into the equation. First, the resistance level could be used as an indication of gear; hence a higher resistance level means a higher gear. This idea would conclude that higher rpm means farther distance at the same resistance, and that when resistance is increased and rpm remains constant the distance increases. The other way to look at is that higher resistance level means a change in grade or incline of the road surface. Thus, a higher resistance level at the same rpm as before wouldn’t equal the same miles per hour if you equated it back to a normal road bike. Taking all of this into account, it was decided that miles per hour and miles traveled would not be used in our user interface, but that if the exercise bike was to ever be mass produced it would need to be included in order to compete in the market place.

3. The reason that watts are not included into the simple user interface is that the numbers have little to no meaning to the normal user and only confuse the user. This goes against calling the accessible ergometer an ergometer, since ergometer means to measure power and watts is the unit of power. But, because the normal user would not know what watts are or know what a

40

certain number of watts produced while pedaling translates to the vigorousness of their exercise, they were decidedly left off to simplify the user interface. Like the miles and mph, watts would need to be incorporated into a mass produced product to compete with other models that do include watts on the user interface. It is recommended that watts along with the previous three measurements be placed on an advanced user screen linked off the simple screen.

Figure 37: Flow Diagram of the electrical components of the user interface.

41

The new user interface also needs to interact with some electrical components from the NordicTrack® SL710. These include: the CardioGripTM, magnetic rpm sensor, and the tension stepper motor. The other circuit to be integrated is an ECG amplifier circuit; this is used to amplify the electrical signals that are read across the palms of the individual using the CardioGripTM. The method for incorporating and reading in the voltage values produced by the rpm sensor, tension stepper motor, and ECG amplifier circuit is to use a data acquisition program and hardware. The data acquisition (DAQ) software used to create the user interface is LabVIEWTM 7.1 by National Instruments (NI). Most programming was done in Student Edition with the executable file creation being done in professional edition. LabVIEWTM delivers a powerful graphical development environment for signal acquisition, measurement analysis, and data presentation, allowing the flexibility of a programming language without the complexity of traditional development tools. LabVIEWTM was decided upon due to its ease to program in and ability to be programmed to suit the needs of the user. The DAQ hardware that was decided upon to go along with LabVIEWTM 7.1 is the National instruments USB6008 that plugs directly into the USB port of a computer to send data. The NI USB-6008 (Figure 38) was chosen because it is small, portable, and it has two analog outputs along with numerous analog inputs. This is all important because we anticipated having two outputs (<+5V), three inputs and having to power a couple components with +5V, which the USB-6008 is able to do. Full specifications of the NI USB-6008 can be found in Appendix F. The user manual for the NI USB-6008 can be found in Appendix G

Figure 38: National Instruments USB-6008 [13].

In order to run the LabVIEWTM program to be created, we attained a used Windows system from SWAP (Surplus with a Purpose), which is an entity of the University of Wisconsin. The computer purchased is a Compaq Deskpro with an Intel® Pentium® 3 processor (800 MHz), 384 MB of ram, 2 USB ports and a hard drive of 14 GB. The computer’s operating system was selected to be Windows XP Pro, because it enabled the group to program in LabVIEWTM on a Windows system. Another option was to use the free operating system Linux and in turn program the complete LabVIEWTM program on a Linux operating system. In the future, it would be necessary to run the program on a Linux system in order to save money on licensing fees. For the user interface visual portion, we decided on using a touch screen LCD monitor. This would not be user friendly for the blind, but for all other users it is the best

42

option because touch screens are cognitively easy to use since the finger acts as the mouse. Thus the problems of using a normal mouse or tracking device are eliminated. The important features that we desired from the touch screen is that it is large enough to make it easy to view, can be viewed from an adequate angle and has a resistive touch screen (the capability of sensing any touch such as a stylus or artificial hand. The touch screen that we decided to purchase was the DellTM Touch Screen Flat Panel Monitor Model Name E153FPT (Figure 39a and 39b). For a full list of its capabilities and features see Appendix I. The monitor plugs into the normal monitor port as well as one of the USB ports in order to pass the touch screen components.

(a) (b)

Figure 39: (a) Touch screen monitor shown from an angle. (b) Dimensions of the Dell monitor.

As mentioned earlier the user interface needs to interact with several components including the CardioGripTM, magnetic rpm sensor, and the tension stepper motor. The other circuit to be integrated is an EKG amplifier circuit; this is used to amplify the electrical signals that are read across the palms of the individual using the CardioGripTM.

The tension stepper motor is used to pull in the tension cable in order to bring the C-magnet closer to the flywheel, thereby increasing the resistance of the pedals (Figure 40a). The tension stepper motor has four inputs and one output. The front of the stepper motor (blue and yellow wires on the unit, Figure 40b) has a voltage differential of +5V when turning the motor; the polarity between the two values controls the direction of rotation. There is only a differential during rotation and otherwise the voltage differential is zero volts. The side of the tension motor acts likes a potentiometer, the high input (H) is +5V while the low input (L) is ground. The wiper (W) goes between the two as the position of the winder changes; this is the sole output of the tension motor. The range of outputs and inputs is 0 to +5 volts.

43

(a) (b)

Figure 40: (a) C-magnet and tension cable (b) Tension stepper motor. [9]

The magnetic rpm sensor is a magnet on the pedals’ flywheel and a magnet sensor is attached to the frame of the exercise bike. The magnet sensor attached to the bike frame has two wires running into it. One wire is supplied with +5 volts and the other wire is the output. When the two magnets face each other the output has a peak of +5 volts. By detecting the peaks and knowing the time between them it is possible to calculate the rpm (rotations per minute) of the pedals. The last parts that the user interface needs to interact with are the CardioGripsTM and the EKG (electrocardiogram) amplifier circuit (Figure 41). The cardio grips are metal plates as described earlier. One plate on each hand is grounded in order to ground the normal electrical potential of the body. The other two plates are used as inputs into the EKG differential amplifier circuit. This circuit is a simple biopotential amplifier circuit found in John Webster’s textbook, Medical Instrumentation [27]. It is not the purpose of this document to fully describe the circuit or the testing process associated with it. The only change that is needed is the gain of the circuit. It is important not to saturate the DAQ device, thus the voltage must remain below +10 volts.

44

Figure 41: ECG differential amplifier circuit [27].

Labview Programming In the previous paragraphs, the components that the DAQ would interact with were described in detail. The following section will discuss the sub virtual instruments (sub vi’s) that were programmed to work in the simple user interface. For this project, only the simple user interface (Figure 42) was developed; the advanced user interface would be considered in future work. The simple user interface includes the following sub virtual instruments: timer.vi, resistance.vi, rpm.vi, and heart rate.vi.

Figure 42: Screenshot of simple user interface (link to advanced screen missing).

45

Timer.vi The timer.vi (Figure 43a and 43b) is a program that’s sole purpose is to take in the user input number of minutes and display the amount of time passed. A special feature incorporated was the ability to have the time either count up or down depending on the preference of the user. After looking at a number of exercise equipments, it was found that some counted up while others counted down. Thus, we found it to be beneficial to have the ability to do both. Figure 43a shows the timer.vi when the count up function is selected. Figure 43b shows the timer.vi when the count down function is selected.

Figure 43a: Timer.vi block diagram for counting up the time.

46

Figure 43b: Timer.vi block diagram for counting down the time.

Resistance.vi The resistance.vi (Figure 44) is a program that has a knob that the user can change so that the resistance against flywheel rotation is increased or decreased. The resistance.vi outputs +5 volts differential across the blue and yellow wires of the tension motor, the polarity of which defines the rotation. So if the resistance level is increased, the voltage at the blue wire is +5 volts and if decreased it is 0 volts. The opposite is true for the yellow wire.

The items missing from this virtual instrument (vi) is analog input from the sweeper. During testing, the tension motor broke resulting in a nonfunctional motor for the rest of the semester. In order to complete the virtual instrument, it is necessary to test the sweeper voltage at different levels of rotation. Once these values are known, it is possible to complete the program with the necessary components so that the voltage applied to the blue wires is zero once the desired sweeper voltage is found. The high (H) sweeper voltage is produced by the USB-6008 and so is the low sweeper voltage of ground.

47

Figure 44: Resistance.vi block diagram.

RPM.vi The purpose of the rpm.vi (Figure 45) is to take in the waveform of the magnetic rpm sensor. By detecting a peak above a set threshold and measuring the time between valid peaks it is possible to measure the rotations per minute for the pedals. This program still needs some work on it for it to work perfectly, as it was found during testing that RPM displayed was not as accurate as one would like, and thus further testing and programming is needed to complete this sub vi.

Figure 45: RPM.vi block diagram.

48

HeartRate.vi The heartrate.vi is used to measure and display the heart rate of the individual. To date the program has not been written, though the underlying principles of its design are fully understood. For this vi a similar vi as the rpm.vi will need to be created, but there will need to be a more advanced threshold detector and method for knowing the time between beats in order to determine the heart rate of the individual. Designing the Layout for a Simple User Interface When designing the layout of the simple user interface (Figure 46) it was important that the interface would be easily viewed by people with low vision as well as those who are colorblind. To be accessible for people with low vision there needs to be large lettering as well as high contrast in colors. To be accessible for people with colorblindness it was important that no information would be lost, thus making it important to convey information in a way that color was not important. The methods used to make the user interface more accessible include:

1. Grouping items together. a. The inputted time, displayed time and ability to count up or down is

grouped with a line around it. Grouping is beneficial since it is cognitively easy to understand

2. High contrast. a. The background is black while displays and controls are the highest

contrasting items (white). b. Backgrounds of number displays set as black to match background.

3. Different displays for different items. a. A gauge is used to display the rpm while the heart rate is shown as a

number. b. A knob is used to change resistance, while an up and down button is used

for changing the inputted time. 4. Large lettering and displays were used to be viewed by people with low vision. 5. Symbols used along with words for heart rate. This is, again, cognitively easy to

understand. 6. Bright LED display (lower right corner) used to show when the time is up.

49

Figure 46: Screenshot of simple user interface (link to advanced screen missing).

In order to make sure that the user interface would be accessible to people with colorblindness we used a program from vischeck.com to simulate how the interface would look like to a person with either red-green (Figure 47) or blue-yellow (Figure 48) colorblindness. The program that is free to download from http://www.vischeck.com is a plug-in to the program Vischeck ImageJ which is free software from the National Institute of Health (NIH) [26]. As can be seen for both colorblindness figures, no information is lost due to colorblindness.

Figure 47: Vischeck simulation of red-green color blindness of the simple user interface that is

shown in Figure 46.

50

Figure 48: Vischeck simulation of blue-yellow color blindness of the simple user interface that is

shown in Figure 46. Miscellaneous Additions Another improvement made on the design was adding a plastic cover to conceal the magnetic resistance components. This ensures the user will not harm himself/herself on the rotating flywheel and makes the device look more aesthetically appealing. A handle was affixed to the front of the device to facilitate transport. Using this handle, it is easier for a user to tilt the device up onto the wheels situated in the back and wheel the device to the desired location. Lastly, since some of our hypothetical users are of the elderly, some of them may be using a cane to assist them in mobility. Therefore, a cane holder was implemented on each side of the handle on the front of the device. Two cane holders were used because some users may be right- or left-handed. To avoid favoring handedness, we placed two cane holders for all users to access easily. Humans Subjects Testing For full consideration in the National Design Competition, the team was required to conduct human subjects testing on subjects containing the disabilities addressed by the hypothetical patients. The team submitted an initial review application, health surveys, and consent forms on November 8, 2004 and we received a memo requesting revisions to the application and corresponding forms on December 1, 2004. The revisions the memo requested are as follows. We must provide additional information as to the recruitment procedure. We must obtain permission from the University of Wisconsin and Meriter Hospitals to post flyers, state where the flyers will be posted, how we intend to screen our subjects, and who will approach and consent the subjects who will be invited. We need to clarify where the study will be conducted and the approximate sample size we will be using. We have to revise the consent form to

51

indicate that the consent process includes a provision that subjects who cannot see well enough to read the consent form have it read to them by an advocate (family or friend) or someone not involved in the research. The consent form should also include a statement so that the reader can act as a witness and attest to that the form signed by the subject is the same as was read to him (applicable towards low vision patients). We must clarify when subjects will complete the “Post Experimental Survey” and the general health inquiry, and instead of having the subjects sign and date the questionnaire, a study number should be recorded. We must more clearly describe the methods of each exercise session. The MR-IRB has requested that because subjects with various disabilities will be involved with this study, an MD medical advisor should be added as part of the key personnel. We need to find a physician to be present during testing and add him to our key personnel, and also state that all subjects should obtain clearance from their primary care physician for the exercise part of this study. The above changes were made and resubmitted to the IRB on February, 8, 2005 along with a cover letter indicating changes made and where they can be found in our forms. Feedback was received on February 18, 2005 which requested us to explain certain items. Certain clarifications were required, including locations of flyer posting, room locations, and the witness statement in the consent form. Additionally, some of the wordings in the consent form had to be changed to make it more readable to non-scientific participants. The changes were made and submitted to the IRB on February 21, 2005. The team received approval for human subjects testing on March 3, 2005. All final versions of the forms used in human subjects testing can be found in Appendix C including: Initial Review Application, Consent Form, General Health Survey, Post-Experimental Survey, Phone Screening Form, and the Recruitment Flier. Testing Procedure Subjects for this study will come from the Madison community between the ages of 18-70. Both individuals with normal motor control and sensory systems and those with loss of motor control and sensory system dysfunction will be used as subjects. The study will enroll four control subjects and up to ten experimental subjects. Subjects with loss of motor control and sensory system dysfunction will be recruited through hospital advertisement and by invitation. Advertisements will be placed in the following areas: School of Nursing, UW Faculty and staff newsletter and email, campus billboards, and newspaper ads. Interested individuals can contact us via email or telephone. If contacted via telephone, a phone screening form will be filled out in which we will provide the interested participant with a subject number and designate a time that works for them to conduct the study. In addition, we will make sure that, if a patient of the UW-Hospital, the patient is cleared with his/her primary care physician. We will then attempt to set up an appointment with that patient to conduct the experiment at the UW-Hospital (Room G5/170-174) between April 5-10, 2005. In addition to the participants, there will be four control subjects that will perform the experiment identical to the participants. All subjects will be asked to participate in a single experimental session. Participation in this study is entirely voluntary. No monetary compensation will be provided.

52

Testing Results After testing was completed, the team was unable to obtain all the necessary participants for testing but was able to find four experimental subjects and four control subjects. The four experimental subjects had disabilities including obesity, heart failure, post-stroke symptoms (including the necessity of walking with a cane), diabetes, and low vision. The responses obtained from each participant through the post-experimental survey were tabulated and is reported below in Figure 40.

Figure 49: Results of human subjects testing with standard deviations

It was found that most participants found the prototype readily accessible. Because of the small number of participants in both the controls and experimental group, the standard deviation was quite high. Based on the results on the ability to enter and exit the prototype, both groups found it easy to access the device and exit upon completion of exercise. The seat assist is used to help users stand from a fully seated position, thus nearly all subjects liked the idea of implementing the seat assist. Furthermore, the lighted foot pedals was another positive point brought up by our participants. Since it was activated wirelessly upon sitting, it is very easy for the subjects to relate to people with low vision to find the foot pedals. Lastly, most subjects liked the idea that we added the arm exercise to the bike. They felt that it gave them more variability in their exercise by allowing upper and lower body workouts. One of the main feedbacks that we received was that it was hard to adjust the pistons for the arm exercise. Because we had to change the actual design of the

53

pistons, it became harder for users to adjust the pistons. Especially since we have two pistons on each side of the prototype, and that one of the pistons on each side is close to the ground, users found it difficult to adjust them without restraining themselves from not falling off the bike. Our experimental group definitely pointed out that having a bike with dual independent upper and lower body exercise is the first of its kind. In addition, they appreciated added components such as the seat assist, power seats, and the simplified user interface. Our control group liked the newly designed bike we created and were able to see it being useful for people with various disabilities. Cost Analysis Description of Purchase Retailer Purchase

Price NordicTrack SL710 Sears Madison West 499.99 100 lb gas spring MSC Industrial Supply 16.82 12” Stroke Linear Actuator ebay.com 106.85 Rollerblade wheels, 55mm diameter, 8 pk

Dick’s Madison East 36.91

Momentary three way switch RadioShack East Madison 4.74 Bolts, nuts, washers, casters, 6’ angle steel

Menards East Madison 16.93

1.5” x 2.5” rect. tube, 1” x 2” rect. tube

AA Quality Welding & Mfg Madison

42.20

4 Butterfly cylinders HealthFX America 97.80 1.25” x 48” L-steel True Value Whitewater 8.43 U-bolts x 4 Home Depot FDL 4.16 Washers, nuts, thread rod Menards East Madison 3.76 Wireless doorbell chimer Stoughton Lumber (ACE) 18.98 1.5” x 1.5” x 3/16” tube, 1” tube AA Quality Welding & Mfg

Madison 73.85

Misc. nuts, bolts, and washers True Value Whitewater 30.95 Leveling feet and nuts Home Depot East

Madison 5.68

Paint, battery terminals, cable ties, wire, batteries

Fleet Farm, Beaver Dam 33.73

Electrical tape, heat shrink, terminal kit

RadioShack West Madison

12.84

Bronze bearings, chain, steel spacers

True Value Whitewater 8.59

Wireless chime doorbell, digital multimeter, misc. nuts & bolts

Wolff Kubly Madison 29.77

Compression spring McMaster-Carr 14.26 Misc. nuts & bolts Ace Hardware 3.80 Drill, misc. nuts & bolts True Value Whitewater 11.90

54

24”x24” tread plate Menards West Madison 51.67 Circuit components RadioShack Oshkosh 16.08 Computer UW Madison SWAP 131.88 Misc. nuts, bolts, washers; ½” x 10’ pipe

True Value Whitewater 28.46

Misc. nuts, bolts, washers True Value Whitewater 3.71 Circuit components RadioShack Oshkosh 6.78 LED lights, holders RadioShack West

Madison 6.29

Paint, duct tape Fleet Farm Oshkosh 8.58 Touch Screen Monitor TigerDirect.com 448.24 LabView data acquisition card National Instruments 151.03 Decals, aesthetic enhancement University Book Store 16.72 Total Expenses $1952.38 Budget Remaining $47.62

Though we spent $1952.38 developing the design, the ergometer must retail for less than $1,000. This is shown to be a feasible task after analyzing the purchases made and deciding what was bought that wasn’t necessary and what can be substituted with a less expensive alternative. First, we started this project by purchasing an already existing exercise device for $499.99, but we ended up not using most of the material purchased. The materials we did use have an estimated value of $150. By purchasing only these materials, a savings of $349.99 is assumed. Second, a smaller touch screen or non-flat panel touch screen could be used instead of the 15” flat panel touch screen that we purchased. Instead of costing $448.24, this substitution would cost only $398.95, saving $49.29. Also, the computer has more features than we require. Instead of purchasing both the computer and LabVIEWTM Data Acquisition Card, it would be more cost effective to build a separate operating device. This operating device would require a Pentium III processor, 2D video card, 384 Mb RAM, 1 Gb hard drive, 2 USB outputs, a data acquisition card, and a Linux operating system. The Linux operating system should be chosen since it is a free system, rather than paying to use the Microsoft operating system. Making these changes results in a $282.91 savings from not purchasing the computer and LabVIEWTM Data Acquisition Card, but an added cost of $150 to build our own system. This results in a net savings of $132.91. Third, we only used approximately two thirds of the material that we purchased to construct the seat frame and track. By purchasing only this amount of material, a savings of $94.02 can be tabulated. Fourth, instead of using four, one way resistance pistons each costing $24.45, two dual resistance pistons could be used, each costing $29.45, producing a savings of $38.90. Fifth, the LabVIEWTM Professional computer program that would be required to be purchased to run our user interface program costs $4095. However, since the license to use this program only needs to be purchased once, if our device were to be mass marketed this cost could be spread over the number of ergometers manufactured. With the $4095 set as a fixed cost, the average cost will decrease with the number of products manufactured due to the economy of scale. Assuming at least 5000 of these devices are produced, the cost per product for the

55

LabVIEWTM program is only $0.82. Sixth, upon construction of the LED foot pedals the first wireless doorbell chime was damaged, requiring the purchase of another $17.93 chime that would not be necessary under normal circumstances. Seventh, the original U-bolts purchased were of insufficient strength and had to be replaced, producing another $4.16 that would not be included in a normal manufacturing expense. With these adjustments and substitutions, the price to build one model is $1266. If this product were to be mass marketed, these materials could be bought in bulk and would therefore be less expensive. For the ergometer to retail for $950 a discount of 25% would be required for bulk purchase, which is not an unreasonable assumption based on preliminary inquiries. It should also be noted that this price includes all engineering research and design costs, so that any price the machine is sold at over $950 would be considered profit. Future Work/Conclusion With the completion of Human Subjects Testing, valuable information has been obtained about the design of the device. While subjects and controls in general rated the device high in quality, certain areas were found that require improvement. First, and most importantly, the user interface must be completed. To accommodate blind users, an audio output should be incorporated, a physical controller to move about the screen, a help button to explain the information on the screen (also for cognitively impaired users), and a screen reader. This can be done using the LabVIEWTM program, and only requires additional computer programming and the downloading of the program onto the computer. Second, an additional handgrip should be added to each of the arm motion handles below the current grip to make the overall length of the grip four inches longer. This will allow shorter people a place to grab onto the handles instead of holding onto the bare metal of the arm handle. Third, there is a section of the seat track that could potentially cause injury. A section of the L-steel that was used to construct the track protrudes near the walk-through platform. When the seat is moved forward, this piece could possibly catch the front of a user’s shoe, thereby squishing the user’s foot. By cutting this small section out of the seat track, this problem can easily be addressed without destroying the track. Fourth, the weight of the device is too heavy for users with disabilities to transport. A solution to this problem is to use a lighter, yet still sufficiently strong material for the frame. With these improvements in mind, our team feels that we have built an exercise device that would be enjoyed by people with various abilities. We believe we have addressed every aspect of the competition put forth to us and that our device would be of great benefit for people who may have trouble using currently existing devices.

56

References [1] Adams, C. D. & Bennett, S. “Exercise in Heart Failure: A synthesis of Current

Research.” The Online Journal of Knowledge Synthesis for Nursing. Vol. 7, #5. February 9, 2000.

[2] Adobe Acrobat Accessibility Techniques. Retrieved on April 22, 2005 from

http://www.webaim.org/techniques/acrobat/ [3] “Burning Calories is not an exact Science.” CNN. Retrieved on April 23, 2005 from http://www.cnn.com/HEALTH/heart/9909/07/cardiomachine.accuracy/. [4] ButterFly Cylinder. HealthFX America. Retrieved on November 5, 2005 from

http://healthfxamerica.com/cylinder_catalog.htm.

[5] DELL E153FPT 15-Inch Flat Panel USB Touch Monitor. Retrieved on February 8, 2005 from http://www.dell.com. [6] Diabetes. Yahoo! Health. Retrieved on April 25, 2005 from http://health.yahoo.com/centers/diabetes/ [7] Fastener Strengths/Grade Markings. Textron Fastening Systems. Retrieved on

April 28, 2005 from http://www.textronfasteningsystems.com/eng_tools_f/grades.html

[8] Heart Failure Online. Retrieved on April 24, 2005 from http://www.heartfailure.org/index.htm [9] Icon Health Customer Service. Retrieved on October 14, 2004. Refer to

Appendix J for full documentation. [10] Kroemer, K.H.E. Engineering anthropometry. In Salvendy, G. (ed.). 1987.

Handbook of Human Factors, 154-168. New York: Wiley. [11] LED Specifications: RadioShack® part number 276-309. Retrieved on April 26, 2005. Refer to Appendix H for full documentation. [12] "Microelectronic Circuits," 4th edition, A.S. Sedra and K. C. Smith, Oxford

University Press, 1998. [13] NI USB-6008 Specifications. Retrieved on February 15, 2005. Refer to

Appendix F for full documentation. [14] NordicTrack SL710. NordicTrack. Retrieved on October 1, 2004 from http://www.nordictrack.com.

57

[15] Parkinson’s Disease Information Page. National Institute of Neurological Disorders and Stroke. Retrieved on April 22, 2005 from http://www.ninds.nih.gov/disorders/parkinsons_disease/parkinsons_disease.htm [16] Parkinson’s Disease. MSN Health. Retrieved on September 6, 2004 from http://content.health.msn.com/hw/parkinsons/hw93188.asp. [17] Potter, David M. Rockcrawler.com. “Making the Grade: A Technical Discussion:

Grade 5 vs. Grade 8 Fasteners” http://www.rockcrawler.com/techreports/fasteners/index.asp

[18] Rehabilitation Engineering Research Center. Retrieved on September 3, 2004

from http://www.rerc-ami.org/index.htm. [19] RST7000 Total Body Recumbent Stepper. Pro-Med Products. Retrieved on December 6, 2004 from http://www.promedproducts.com/. [20] Schwinn Airdyne Windjammer UBE. Pro-Med Products. Retrieved on December 6, 2004 from http://www.promedproducts.com/. [21] Simple Logic Gates. Retrieved on December 3, 2004 from

http://www.brunel.ac.uk/~castjjg/hndcfund/material/logic/part5.htm. [22] Torque Limits. Engineers Edge. Retrieved on April 28, 2005 from

http://www.engineersedge.com/wwwboard/messages/289.html

[23] Trace Research and Development Center. Retrieved on April 21, 2005 from http://trace.wisc.edu/.

[24] Type 1 Diabetes. MSN Health. Retrieved on September 6, 2004 from http://content.health.msn.com/hw/diabetes_1_2/hw34305.asp.

[25] Up-Lift Seat Assist. Up-Lift Technolgies. Retrieved on October 15, 2004 from

http://www.up-lift.com/. [26] VisCheck. Retrieved on April 10, 2005 from http://www.vischeck.com/

[27] Webster, J. G. (ed.), Medical instrumentation: application and design, third

edition, John Wiley & Sons, New York, 1998