Simulation Of Fuzzy Algorithm In Obstacle Avoidance Based On Ultrasonic Sensors

With the development of artificial intelligence and robotics technology, researchers pay more attention to autonomous mobile in more and more countries. How to implement secure autonomous mobile robot obstacle avoidance has become a hot research field. In this paper, aiming at the conventional algorithm based on ultrasonic sensor robot obstacle avoidance for its security issues and other autonomous problems, a novel Fuzzy Neural Network(FNN) control method with high autonomy and high security improvements is proposed. Firstly, obstacle avoidance controller for omnidirectional autonomous mobile robot, combined with improved fuzzy neural network, is designed on the basis of the mobile robot’s motion model and dynamics model. The improved fuzzy neural network controller for obstacle avoidance makes full use of FNN’s self-learning ablitily, and a substantial increase can be obtained in the autonomy and security of omni-directional mobile robot obstacle avoidance. Meanwhile, after the optimization on algorithm complexity and structural adjustment of the neural network, a four-layer structure of the fuzzy neural network controller is provided for obstacle avoidance. The fuzzy control rules can be self-generated by use of self-learning features of fuzzy neural network, so that the control algorithm can guarantee the accuracy of the premise and improve the algorithm real-time ability effectively for applications in the actual systems.This paper mainly focuses on the improvement of the proposed fuzzy obstacle avoidance control algorithm in Matlab simulation environment, and improved FNN control algorithm model is established for obstacle avoidance. The obstacle avoidance simulations of the mobile robot is designed and accomplished in real-time,including static and dynamic simulations on Webots platform. Furthermore, the improved FNN control algorithm is compared with traditional one on obstacle avoidance performance, and performance indicators are analysed based on simulation results. The simulation results show the effectiveness and feasibility of the proposed improved fuzzy neural network algorithm in obstacle avoidance process.