The Representation Challenge

• CPN’s then seem to provide the necessary structure to represent AT systems

• For this to be successful , the representation of the components of AT systems needs to be sufficiently detailed to reflect the many facets of Person, Environment and Activity which make up these systems

Representation and Tokens

• At the heart of CPN Representation are the markings which consist of sets of tokens at the various places.

• These sets of tokens define the state of the domain before and after the transitions.

• Guards are tested against the sets of tokens and outcomes describe the effects of activity on the tokens

• So tokens are at the centre of this representation scheme.

To establish what is required

• Consider Our AT System which consists of -• A Person • In Some Environment Performing • Some Activity (possibly with the help of )• Some AT• We need token definitions which represent

which represent the above

CPN tokens

• CPN Tokens have rich representation capability and are suited to representing the Person - Environment Activity and AT tokens required in modelling AT systems

• Best way of representing these elements of activity is using CPN record token structure.

• Basically a sequence of labelled types with values.

• Can use CPN Union tokens to join up subsets of various component tokens

In subsequent sections

• The structure of these AT system elements are discussed.

• These elements namely• PERSON, ACTIVITY,ENVIRONMENT and AT are

summarised in the following diagram

PERSON -ENVIRONMENT –ACTIVITY- AT

Lift Up Kettle

Carry Kettle over to sink

CONSIDER For Example

ACTS

PERSON

Environment

AT Considered Later

From the Preceding

• There is a Person• The Physical Environment is a Kitchen (Note

there are other Environmental Components)• The Activities Include:• Lifting a Kettle and Transporting it to the Sink

Representing these

• All Components of AT systems must be represented in any model

• In Other Words

Need to Represent

• The person who carries out the task• The environment in which the task is carried

out.• Objects of the Domain (Including AT

Artefacts). These are often considered Environmental Concepts

• Finally The Activities.

Representing the Person

The Agent

• The person who performs the task is known as the Agent

• Need to represent the• Agent Attributes• Agent Capabilities

Agent Attributes

• There are many attributes to consider• These include• Personal Factors Age, Gender, Height

Profession etc• Body Structures Absence or Lack of any

significant body structures such as hand leg • Anthropometric Data such as reach, height,

body dimensions

Agent Capabilities

• There are many:• Body functions such as mobility, cognitive • Activity and Participation data• Psycho Social MPT DataFor example

If a person wishes to perform this task

• 1: They need to want to do this• 2: They need to be able to locate the kettle• 3: They need to be able to move to the kettle• 4: They need to be able to grip the kettle

handle• 5: They need to be able to lift an empty kettle

In other words

• There are many attributes of the person being tested against the requirements of the task.

• These include Agent capacities e.g the ability to lift etc

• They also include assumptions about the availability of body structures such as hands and eyes

• The following are at least all of the features of a person which may be considered.

Representing the Environment

Representing the Environment

• Two Broad categories• Environmental Factors Such as

– Physical Environment– Attitudes– Policy and Legislation– Relationship and Associations

• Objects in the Environment including Assistive Objects

Representing Environmental Features

• There are two categories, Physical Objects that make up the environment and other factors that fall into environmental classification

• These latter categories based on those from the ICF include:

AttitudesServices Policy and LegislationPhysical EnvironmentRelationship and Associations

Physical Environment

• These factors include attributes of the environment such as

• Lighting• Terrain • Habitat• Whether indoors or outdoors• Objects are considered separately

Attitudes

• Attitudinal factors have been shown to impact on AT use

• These include• Attitudes of Self• Attitudes of Peers• Cultural Attitudes• Religious Attitudes

Services Policy And Legislation

• Legislative and Policy mandates have a huge bearing on successful AT use.

• Legal Aspects impacting on AT use include:• Service Provider Policies and Mandates• Local and National Legislation• AT Service Provision Policies

Relationships and Associations

• Family• Friends and Acquaintances• Strangers• People in Authority• Personal Assistants• Health and Other Professionals

Representing Objects

The Key Environmental Representation

• The objects including Assistive Technology that populate the activity domain are central to any representation scheme of AT systems

Representing Objects

• Need to represent Position• Need to represent static and dynamic

attributes• Need to represent components.• Need to represent smart knowledge- (that is how the object is interacted by an

agent)

Object Interaction

• Interacting with the objects of a task is an important ability in activity participation

• Task objects place important demands on the person carrying out the task

• These are called Capability or Capacity Demands

Representing Object Interactions

• Object interactions are described by Interaction Tokens which are associated with each Object

• Each Object may have more than one interaction token linked to it.

• This reflects that each object may be interacted with in different ways

• This knowledge is called Smart Knowledge

Predicates

• In addition to the above we need to be able to represent facts about the domain.

• These will be called Predicates and describe associations and relationships such as one object is next to another.

• Some are explicitly declared and some are calculated from object attributes such as position vectors

Examples of Dynamically Calculated Predicates

• Spatial functions – Beside , Near , On top of etc• These are example of Fluents (Facts and

Statements about the world that Change across states)

Other Issues

• Represent the action • Do all this using CPN token representation.

Representing ACTIVITIES

Representing Actions

• As stated previously• Three components of Action are identified

– Intended Action which give Parameters of the activity such as Duration Location etc

– Guards – Outcomes

Intended Actions

• Badler and others explicitly describe the action that is about to be performed

• (Uninitiated Parameterized Action Representations-uPARs)

• These uPars describe details of the Action such as Duration, Location, Source and Desination and much more

• The biggest advantage of this is that it allows us to distinguish between different instances of the same class of activity

For example

• The person making tea will have to lift a number of different objects including the kettle, teapot, tea caddy sugar bowl and so on.

• All of these instances of the activity “to Lift” have different characteristics and place different capacity demands on the person performing them.

• For example lifting a full kettle requires greater strength and a different grip than lifting a sugar bowl.

Case Structure

• A common way of capturing this kind of information is using case structure.

• For example the cases of “John Lifted the Kettle” would include– Agent (John)– Act (Lift)– Object (Kettle)– Location (Kitchen)– Time (3pm)

From this knowledge

• We can provide features to be used in the guard on the transition

• We can describe agent capacities and Performance specific to this intended ACT

• As we will see later we can use this to integrate the ICF into this representation

• Many computer Systems such as planning systems use just Pre (Guards) and Post (Outcomes) Conditions to represent Activity

Guards

• The preconditions on an activity are represented in CPN’s by the transition Guards

• These demands are the requirements that must be met by the agent and the environment and the pre activity state in general before execution of the activity

Consider for example the following

Capacity Demands(Capability Demands Clarkson)

• Looking at the above milk bottle designs:• Each bottle design demands that the user has a

capacity to perform a vertical lift by gripping the handle with a closed fist grasp.

We see that the bottles on the left will allow a greater range of hand sizes get a proper grip on the handle for lifting since it gives greater clearance dimension between handle and jug

• In other words the structure of each bottle implies the user must have particular hand dimensions in order to manipulate the bottle

• Thus each bottle places different demands on the user attributes.

• If these demands are not met then the bottle cannot be used.

• This conflict is the essence of how capacity demands define the guards of our petri nets

Guard Demands

• There are many such demands • There are Agent Capacity ( Capability ) Demands• These include Object Capability Demands - Can the

agent use an object involved in the activity.• There are environmental demands For example is

there enough space to carry out the task• There are even Activity demands- Can the agent do

that Activity for that long?

Demands of the Transition Guard

• These must be measured against the personal capacities of the agent and the attributes of the environment

• This is summarised in the following

Capacity Demands

Action and Objects

Agent Capacities,EnvironmentalFactors,State Attributes

More formally

Person Capacity Tokens

EnvironmentAttribute Tokens

Action Capacity Demands

Object Capacity Demands

Environmental Demands

Incoming Tokens representing

Person and State

Transition Guard representing Barriers

( in terms of Capacity Demands)

State

Action Tokens

Guard Construction

• Guards on a Transition then are constructed from the demands placed on the agent and environment by the activity and the objects involved in it.

• This is summarised by the following

Transition(Labelled

Description)

Guard Outcomes

Capability Demands

Person Environment

Action/Activity---Objects

Outcomes

• These are the effects or outcomes of an activity.

• More formally they are whatever changes to the state occur as a result of the activity or event

• In the next slide examples of possible state changes will be presented

Example outcomes

• Objects get moved so their position changes• This in turn will change dynamically calculated predicated

values. For example some object is no longer beside another object.

• It could be an environmental change such as an obstacle is removed from a path

• It could be a change to a person’s capacities. They get tired after performing the task.

• It might be change in the persons mood. They learn a new skill and get a job- so they are happy

Outcomes and AT

• All the AT interventions previously mentioned result in state change

• Some AT interventions restore or establish capacity for example an augmentative speech device.

• Some change the environment and the capacity demands it places - for example a ramp is easier to traverse than steps

• Some interventions change the capacity demands of the Activity

Capacity Demands And Assistive Technology

• Action and objects place capacity demands on people and environment.

• For example using a standard kettle involves a capacity demand of being able to perform a vertical lift of up to 1 kg(which is the weight of the kettle when full with water), one handed using a closed fist grip.

• Assistive Technology changes the relation between personal and environmental capacities and the capacity demands of the action being executed.

• This relationship is represented by the guard of the CPN• This is shown in the following example

Using a kettle tipper Changes Capacity Demands

In the above example

• The pouring capacity demand involved in using a standard kettle is substantial.

• Using a tipper, reduces this capacity demand.• A person may not have capacity to meet the pouring

capacity demand of a standard kettle, but will have enough using a tipper.

• The guard which represents capacity demand versus personal capacity will fail on the standard kettle transition but will be satisfied using the tipper transition

In Summary

• Outcomes can change any part of any token of the input state.

• Based on what has gone before we have at least the following tokens at each input place

Outcomes and Tokens

• Outcomes can change any part of any token of the input state.

• In other words all of the above tokens can be changed after a transition.

• This is shown in the following:

Places and Transitions

Changes that can take place

• If it’s a token field with a numeric value it can increase or decrease that value

• It can replace non-numeric fields with appropriate substitute values

• This is summarised in the following diagram

Outcomes

Transition

Outcomes

Specified Token Changes

+ Enhancements

- Inhibitors