A FUZZY-NEURAL NETWORK APPROACH TO MULTISENSOR INTEGRATION FOR OBSTACLE AVOIDANCE OF A MOBILE ROBOT

作者：Shi Weiren^*; Wang Kai; Yang Simon X

来源：Intelligent Automation and Soft Computing, 2009, 15(2): 289-301.

DOI：10.1080/10798587.2009.10643032

• fuzzy-neural networks
• multisensor fusion
• obstacle avoidance
• mobile robot
• Takagi-Sugeno model

In this paper, a novel fuzzy-neural network approach is proposed for obstacle avoidance of mobile robots through multisensor fusion. The proposed model is constituted with a precondition network and a conclusion network. In the precondition network, each rule matches the preconditions of fuzzy rules. The conclusion network generates conclusions of fuzzy rules. The total output is the weighted addition and the weight represents the applicability of each rule. The proposed model not only has the ability to handle vague information through the furry logic but also has the ability to learn through the neural network. Multisensor integration based on the proposed fuzzy-neural networks is applied to obstacle avoidance of mobile robots, which adopt multiple ultrasonic sensors to detect the distance and direction of obstacles. The mobile robot can recognize the obstacles, the types of environment and generate collision-free motion by the fuzzy-neural network model. Simulation results show that the proposed model is capable of recognizing the environment and avoiding the obstacles and generating a collision-free path from the start point to the end point.

• 出版日期2009
• 单位重庆大学