small, lightweight speed and distance sensor for skiers and snowboarders michael bekkala michael...
TRANSCRIPT
Small, Lightweight Speed and Distance Sensorfor Skiers and Snowboarders
Michael BekkalaMichael Blair
Michael Carpenter Matthew Guibord
Abhinav ParvataneniDr. Shanker Balasubramaniam
Objective Introduction Design Requirements Proposed Solution Testing Results Conclusions and Accomplishments
Introduction Living by the numbers
Measuring and interpreting your performance has never been more valuable
Critical to gain a competitive edge Most devices are based on repetitive motion but
don’t apply to skiing or snowboarding Nike+ Bicycle speedometer
Goal: Build a speed and distance sensor for skiers and snowboarders
Design RequirementsOperable in Subzero Temperatures
Useable in winter apparel
Lightweight – Less than two pounds
Cost – less than $500
Battery Life greater than 2 hours
Accurately record speed and distance
Proposed Solution Integration of two major systems
Global Positioning System (GPS) Inertial Navigation System (INS)
Inertial Navigation System Comprised of accelerometers and gyroscopes
Accelerations Angular Velocities
Requires analog to digital conversion Careful calibration is required to achieve high
accurate readings Gives great short-term
accuracy, but errors growwith time.
Global Positioning System Gives position in terms of
longitude, latitude, and altitude
Determine distance, displacement and speed
NMEA 0183 communications standard
Long term reliability, butpoor short term accuracy.
Sampled at 1Hzhttp://en.wikipedia.org/wiki/Gps
Kalman Filtering Advanced sensor integration technique
Uses statistical error measurements to provide a better estimate
Accuracy improves with time Can use the estimate to regulate INS errors
User Interface Easily accessible menu displayed on LCD
Review Performance Average and Peak Speed (mph) Distance and Displacement (miles)
Color changing buttons Color of button changes depending on their use Intuitive color scheme aids usability (ie. Red
for Stop, Green for Go)
Menu Hierarchy
PC Interface Upload run data from device
Can be saved on your PC for performance tracking and reference
Points can be plotted in Google Earth
Visual representation ofdata using graphs
PC Interface Example
Successful Failure Reached beyond design requirements with an innovative
solution Unpredicted complexity of INS and Kalman filter
Hardware implemented, software not ready Invaluable knowledge gained through:
Research Trial and error
Despite proposed solution being a successful failure GPS solution that meets and exceeds design requirements Extensive user functionality
Store, retrieve, and manage data all on device or through a PC
Higher data sample rate for greater accuracy Rechargeable battery
Kalman Filter Divergence Difficult to
troubleshoot due to 1200+ lines of code
Unsuccessful in determining cause of divergence
Poor Initialization might be to blame
Confident that with more time, root cause could be determined
Testing Results Several Trials
Moving vehicle On foot Varying speeds and distances
Peak speed had significant error One bad sample results in erroneous reading
Average speed had reasonable accuracy based on number of data points
Results verified by digital speedometer and consumer GPS.
Conclusions and Accomplishments GPS Solution
Accurate average and peak speeds Distance accurate over a large span
Inaccurate over short spans and slow speeds
Robust Display Menu Allows easy access to data and configuration
options PC Interfacing
Plot a run on Google Earth with the click of a mouse Plot speed over time to visualize a run
Future Considerations Printed Circuit Board USB Compatibility Wireless headset communication via
Bluetooth Smaller form factor INS integrated through Kalman Filtering