A distributed implementation of a mode switching control program

Abstract

A distributed implementation of a mode switched control program for a robot is described. The design of the control program is given by a set of real-time automatons. One of them plans a schedule for switching between a fixed set of control functions, another dispatches the control functions according to the schedule, and a final one monitors the system for exceptions that shall lead to a halt. The implementation uses four transputers with a distribution of phases of the automatons over the individual processors. The main technical result of the paper is calculations that illustrate how to justify that the implementation meets real-time constraints.