作者: D. Gachet , M.A. Salichs , L. Moreno , J.R. Pimentel
关键词:
摘要: We present an implementation of a reinforcement learning algorithm through the use special neural network topology, AHC (adaptive heuristic critic). The is used as fusion supervisor primitive behaviors in order to execute more complex robot behaviors, for example go goal, surveillance or follow path. part architecture execution mobile tasks which are composed several act simultaneous concurrent fashion. allows take place at level, it incorporates experience gained executing well overall task. this autonomous approach has been tested within OPMOR, simulation environment robots and with our platform, UPM Robuter. Both, simulated actual results presented. performance adequate. Portions work implemented EEC ESPRIT 2483 PANORAMA Project. >