Research will be conducted on the design of a distributed control system for a robot. The goal is to produce a high-level control network which will allow a dextrous hand (the Salisbury hand) to be used in a task-oriented manner, i.e., where actions of the hand are specified in terms of their effects on objects. Research on the dextrous hand already conducted in this laboratory provides the context in which distributed robot control will be explored. Research issues will be investigated involving both design of a logical distributed model of computation specifically for the robotic domain, and design of a dynamic mapping function to realize this model using a network of processors.
