An Algorithm And Implementation Of Behavior-Based Reactive Decision-making Control System For AUV

In recent years, Autonomous Underwater Vehicle (AUV) is widely used in the field of marine scientific research and military. To some extent, Decision-making Control System determines AUV's intelligence level, and is the key factor for AUV completing varieties of tasks successfully. This paper aims to research AUV Decision-making Control System with reactive structure based on behavior.The Behavior-Based Decision-making Control System this paper introduces possesses many Behavior-Based advantages. In order to set up the Decision-making Control System reasonably, this paper divides AUV's behavior into two classes—Rational Behavior set and Perceptual Behavior set. Rational Behavior set is superior behavior which includes complicated algorithms, and it constitutes Rational Behavior System; Perceptual Behavior set includes simple algorithms, and it has high real-time property. Rational Behavior set and Perceptual Behavior set interact closely to ensure that AUV completes tasks successfully.For the Decision-making Control System with reactive structure, behaviors are overlapped by weighting coefficient which shows off AUV's personality factors. The paper realizes controllable weighting coefficient by fuzzy logic technology. It makes use of obstacle's minimum average distance in the environment to show environment characteristics, and makes this variable as the input of fuzzy logic system, then obtains the output by fuzzy reasoning which is behavior weighting coefficient. This approach enables AUV to show different characteristics depend on different environment characteristics which greatly improves AUV's adaptability. The paper introduces a random disturbance behavior to solve the problem that reactive structure may be trapped in local minimum problem.At last, the paper selects Goal directed behavior and Obstacle avoidance behavior, and takes advantage of SIMULINK and FUZZY toolboxes in MATLAB to set up the Decision-making control system simulation platform, and collaborate with AUV motion control unit's MATLAB simulation module to carry out debug. From the simulation result we can conclude that decision-making control system with reactive structure based on behavior is valid.