Learning to coordinate behaviors for real-time path planning of autonomous systems

We present a neural network (NN) system which learns the appropriate simultaneous activation of primitive behaviors in order to execute more complex robot behaviors. The NN implementation is part of an architecture for the execution of mobile robot tasks which are composed of several primitive behaviors in a simultaneous or concurrent fashion. We use a supervised learning technique with a human trainer generating appropriate training for the simultaneous activation of behavior in a simulated environment. The NN implementation has been tested within OPMOR, a simulation environment for mobile robots and several results are presented. The performance of the neural network is adequate. Portions of this work has been implemented in the EEC ESPRIT 2483 PANORAMA Project