Special Seminar in Computing + Mathematical Sciences

The ClearPath project's goal is to create a proof of concept robotic platform intended to guide a visually impaired person indoors.  Ideally, a sensor-rich, smart environment would assist the guide robot's planning and run-time navigation while the robot guides the blind user through the building.  The first project phase in this effort is to demonstrate the potential of indoor guidance for the blind.  It focuses on a typical scenario where a sight-challenged user needs to leave a room, traverse hallways, and eventually reaches the target destination place in the same physical building.  The Teleo-Reactive rule-based paradigm is used for travel plan representation and overall plan execution control.  It was selected for its integrated task-level, goal-directed and reactive control properties.  This solution satisfactorily matches the requirements and also allows for a highly modular design and implementation.  Though a client-server architecture is used in this first phase, software development task partitioning is the primary reason for its use.  This talk will describe how an off-the-shelf robot can demonstrate the potential for this application using the Teleo-Reactive paradigm for goal-directed task control and plan monitoring.

This effort was conducted at Carnegie Mellon University Silicon Valley.