A reactive behavior design for the mobile robot based on QL

A new reactive behavior design algorithm is introduced. It is achieved through deliberative behavior learning. The deliberative behavior is a static local optimal path planning, which is learned by Q-Learning (QL) method of the reinforcement learning (RL). Control rules are formed after learning and are acted as a reflection of the implementation. A lookup-Q matrix M
̂Q
which size is 11×192 is constructed to store each state-action pair. According to the Boltzman equation, a movement is chosen at the same state. The reinforcement signal is studied carefully with a non-uniform manner. The state-action pair which has the maximum
Q̂
value in each row is selected out after QL. The optimal control rules are formed after the merger. The algorithm automatically control formation of the rules and amends them simply. At last the method performance is tested in different environments under the control of the rules.