next up previous
Next: Conclusions Up: StartleCam: A Cybernetic Wearable Previous: Reducing Information Overload

Applications and Future Work

StartleCam reacts to the user's real-time psychophysiological reactions with a selective camera recording device. Beyond SafetyCam and selective video memory, algorithms incorporating the startle detector are being developed to measure the wearer's stress level. The control system will allow many agents to cooperate with the wearer to manage other aspects of their task load. If you choose, the computer could respond to detecting a high stress state by filtering your less urgent e-mail or playing your favorite relaxing song. The computer could also respond to a state of low arousal by reminding you of important long term goals or presenting you with exciting news stories in your areas of interest. Over time, the computer could learn your preferences based on context, for example the computer could remember that last time you were in a high stress state at 2AM and had a paper deadline approaching, you preferred to have upbeat music played rather than relaxing music but still wanted all e-mail put on hold, except messages from your co-author. By recognizing user stress level and applying user preferences, the computer can help the wearer manage information at a level that is optimal for performance where the wearer is neither bored nor over stimulated[Kah73].

If the wearer were willing to share his responses, the data from StartleCam could be used to aid the designers of computers, cars, roads and household products. The skin conductance response is correlated with arousal, attention and effort. By reviewing data from different wearers performing tasks with StartleCam, designers could see which aspects of tasks were most difficult for wearers. This information could help researchers find better ways to improve daily chores, from attempting to diaper a baby, to creating safer roadways which were not as taxing on drivers' vigilance[Hel78].

The StartleCam system could also be triggered by physiological measurements other than the skin conductivity response. With the same system shown in Figure 1, the wearer's heart rate, respiration rate and muscle activity can be monitored as well. Algorithms could be developed to trigger the camera to activate when the wearer performed an identifiable physical task, such as standing up or when the wearer interacted with other objects which triggered responses in the wearable. Although StartleCam is only triggered by patterns in one physiological signal now, it could potentially learn more complex patterns that are a function not only of the wearer's physiology, but also of context and activity. This kind of monitoring could also augment current remote medical monitoring systems.


next up previous
Next: Conclusions Up: StartleCam: A Cybernetic Wearable Previous: Reducing Information Overload
Jennifer Healey - fenn@media.mit.edu
1999-02-12