Research Of Mobile Robot Path Planning Based On Fuzzy Neural Network

Mobile robot as an important branch of robotics has had the rapid development, widely applications in industrial automation, mining, excluding danger, military, service, etc in recent years. Mobile robot path planning is point at the environment with obstacles, according to a certain evaluation standard, a path without collision from starting state to target state is looked. And the path planning problem is one of the important directions in the research of intelligent mobile robotics.The university version of intelligent robot MT-UROBOT installs five infrared sensors, the obstacles of the former, left before, right before, left, right direction can be recognized, the mobile robot discovery ability to the environment will be improved. So this paper takes the MT-UROBOT as the research platform, a research is made according to the path planning problem. In this paper, the method of mobile robot pose representation and control is analyzed and deduced, mathematical model of surrounding environment is established by the grid method, the strategy of fuzzy neural network is used to evade obstacles, the function of evading obstacles is realized by using fuzzy neural network searching the next feasible node. According to problem of parameters optimized about fuzzy neural network, a mixed method combined conjugate gradient method and least square method is used to optimize the parameters in this paper, to avoid improper selection for learning rate causing the system instability, and simulation results prove that the method is effective.Simulations about path planning of having obstacles and no obstacle are made, and the results show that mobile robot path planning based on fuzzy neural network realizes the performance indicator that the sum of cost about the journey and expense about obstacles are minimum.