Robot behavioral selection using discrete event language measure

摘要

This paper proposes a robot behavioral /spl mu/-selection method that maximizes a quantitative measure of languages in the discrete-event setting. This approach complements Q-learning (also called reinforcement learning) that has been widely used in behavioral robotics to learn primitive behaviors. While /spl mu/-selection assigns positive and negative weights to the marked states of a deterministic finite-state automaton (DFSA) model of robot operations, Q-learning assigns reward/penalty on each transition. While the complexity of Q-learning increases exponentially in the number of states and actions, complexity of /spl mu/-selection is polynomial in the number of DFSA states. The paper also presents results of simulation experiments for a robotic scenario to demonstrate the efficacy of the /spl mu/-selection method.