Clay: Integrating Motor Schemas and Reinforcement Learning

摘要

Clay is an evolutionary architecture for autonomous robots that integrates motor schema-based control and reinforcement learning. Robots utilizing Clay benefit from the real-time performance of motor schemas in continuous and dynamic environments while taking advantage of adaptive reinforcement learning. Clay coordinates assemblages (groups of motor schemas) using embedded reinforcement learning modules. The coordination modules activate specific assemblages based on the presently perceived situation. Learning occurs as the robot selects assemblages and samples a reinforcement signal over time. Experiments in a robot soccer simulation illustrate the performance and utility of the system. 1 Background and Related Work 1.1 Motor Schemas Motor schemas are an important example of behavior-based robot control. The motor schema paradigm is the central method in use at the Georgia Tech Mobile Robot Laboratory, and is the platform for this research. Motor schemas are the reactive component o...