The Need for Context |
|
why wearable applications need context and vice versa |
Wearable applications need context (non-explicit input) to simplify the question of what to do next. Classification, and higher-level situation/activity modeling required for context awareness is ideally suited to a wearable context modeling platform. | |
Wearable Applications Need Context |
simplification through perception
Context allows a wearable application to decide among sets of assumptions or axioms pertaining to making decisions. (This is the frame problem of AI fame.) For example, a communications application should know whether the wearer is busy or not before deciding to interrupt. Assuming that relevant aspects of the wearer's context can be determined, surprisingly simple rules (such as "don't interrupt me when I'm busy") can produce behavior that appears quite intelligent. The "intelligence" is in the perception of context, not the reasoning. |
Context Needs Wearables |
context modeling for personal applications is best situated on the person
Sensing is best done close to the subject of interest. Instrumenting a person is a better way to determine most aspects of her context than instrumenting her environment. Placing sensors on the body makes the wearer the natural focus and allows the application to "see" and "hear" the environment from her point of view. Body-worn sensors can provide functionality anywhere and any time, as long as they are functioning and worn. However, sensing is nothing without analysis, and the results must be accessible to the context-aware application. To ensure this accessibility (and, importantly, the privacy of the wearer) it makes sense to place the processing, analysis, and application on the body as well as the sensors, resulting in a single integrated context-aware system. |
|
|
|
Back | Up | Next |
Context Awareness and Applications |
Richard W. DeVaul |
The second annual "I Wanna Be a Cyborg" event, a borglab production. |