Navigation Research Of AGV Based-on Multi-sensors Behaviors Fusion

Mobile robot is one of the important branches in the realm of robotics, its main aim is the research on the real-time control problem of robot system in the complicated environment, concerning the key techniques about task planning, movement planning and navigation, goal identifying and orientation, robot vision, multi-sensors information management and fusion as well as system integration. The paper mostly studies mobile robot navigation technique, introduces the theory, method and application of multi-sensors information fusion in navigation system. Also the mixed system structure is presented based-on behaviors and on the basis of it neuro-fuzzy network is applied to realize AGV navigation.The thesis adopts several ultrasonic sensors to acquire obstacle characteristic information in unknown environment, through information fusion it can eliminate single data uncertainty so as to get exact environment data, enhance AGV fault-tolerant ability. Simultaneously, we select the structure based-on behavior mix-system to solve the feedback delay problem based-on functional decomposition structure, it can resolve the lower intelligent problem based on the behavior reaction system structure, also we analyze the elementary behaviors of the AGV navigation in detail and establish the fuzzy rules of obstacle avoidance, goal tracking, wall following and emergency brake. On this basis, the navigation system with behaviors fusion is designed based-on multi-sensors. Also the neuro-fuzzy network is used to study the rules. The Simulation results demonstrate that neuro-fuzzy network and behavior fusion tactics best realize the AGV navigation behavior in unknown circumstance, in result robot can rationally evade obstacles and arrive at the target.