Sensory Perception & Interaction Research Group

Intelligent systems are increasingly able to offer real-time information relevant to a user's manual control of an interactive system; however, effective presentation of this information creates many challenges. We consider how to use force feedback to convey information to a user about dynamic system control space constraints that have been computed by an intelligent system. Effective display of control constraints will require careful consideration of the usability of the forces, in addition to good technical design, to assure user acceptance of the feedback. Possible dynamic systems that can benefit from this kind of interaction feedback are tasks such as driving and the control of physically-based animations. In this thesis, we studied the haptic display of control constraints in a simple driving simulation. We developed a 'look-ahead' guidance method to display usable haptic guidance suggestions to a driver based upon the predicted location of the vehicle relative to the road, and implemented this using a custom vehicle simulator based on Reynolds's Open- Steer framework. The performance and usability of our Look-Ahead Guidance method are compared to a baseline of No-Guidance, and to Potential Field Guidance, the current state-of- the-art haptic path guidance method. Our experimental results show that Look-Ahead Guidance was more usable and showed performance benefits in our task compared to both No-Guidance and to Potential Field Guidance. We identified several factors that we suspect affect the usability of haptic path guidance and suggest future work based on these observations.