Development and training of a learning expert system in an autonomous mobile robot via simulation

Abstract

The Center for Engineering Systems Advanced Research (CESAR) conducts basic research in the area of intelligent machines. Recently at CESAR a learning expert system was created to operate on board an autonomous robot working at a process control panel. A two- computer simulation system was used to create, evaluate and train this learning system. The simulation system has a graphics display of the current status of the process being simulated, and the same program which does the simulat ing also drives the actual control panel. Simulation results were validated on the actual robot. The speed and safety values of using a computerized simulator to train a learning computer, and future uses of the simula tion system, are discussed.