Obstacle Avoidance Method For AUV Based On Fuzzy Neural Network

With the development of ocean engineering,autonomous underwater vehicle(AUV)has attracted more and more attention.There are AUVs in the fields of submarine topographic mapping,resource exploration and marine defense.The reason is that AUV has the characteristics of flexibility and wide range of motion.But the marine environment is very complex,there are many uncertain obstacles such as trenches,steep slopes and islands,so AUV also has a huge risk in the course of navigation tasks.Aiming at the autonomous obstacle avoidance of AUV in the unknown underwater environment,a fuzzy neural network obstacle avoidance method is proposed in this paper.This method is applied to the autonomous obstacle avoidance control of AUV,which makes AUV have the ability of logical reasoning and self-learning.The obstacle information of the AUV obstacle avoidance sensor is input into the fuzzy neural network,and the AUV's course angle,pitch angle and speed regulation are output.In order to realize autonomous obstacle avoidance of AUV in underwater environment according to existing experience.The main purpose of this paper is to reduce the risk of AUV in underwater navigation,improve the autonomy of AUV in underwater navigation,and make AUV tend to intelligent development by establishing the fuzzy neural network obstacle avoidance algorithm.The main contents of this paper are as follows:(1)The obstacle avoidance model of AUV in underwater space is established.According to the mature models of torpedo and submarine,the motion of AUV in underwater space is analyzed.The kinematics equation and dynamics equation of AUV in underwater three-dimensional space are established,and the hydrodynamics of AUV is calculated.Finally,the key control factors of AUV in obstacle avoidance control are obtained,that is,the factors that affect the horizontal rotation and vertical latent motion of AUV.(2)According to the factors that affect the obstacle avoidance motion of AUV,the obstacle avoidance behavior of AUV is planned and a fuzzy neural network obstacle avoidance control algorithm is established.By analyzing the forms of different obstacles,the behavior of AUV in obstacle avoidance is determined.According to the analysis of the dynamic obstacle avoidance process of AUV,the control strategy of AUV in obstacle avoidance motion is obtained.Finally,according to the above analysis,the fuzzy rule base is established,the structure of fuzzy neural network is designed,and the neural network is trained.(3)Finally,the fuzzy neural network obstacle avoidance control algorithm is simulated and verified.Two types of obstacles,simple and complex,are set in the horizontal plane to verify the algorithm.In the vertical plane,the algorithm is verified by simulating the typical obstacle environment encountered in the AUV fixed altitude navigation.Finally,the algorithm is verified in the three-dimensional complex environment.Several simulation results show the feasibility of the algorithm.