Research On Biologically Inspired Cooperative Tracking Method Of Multiple AUVs

In the face of large-scale,high-efficiency ocean exploration and development tasks,it is difficult to guarantee the operational reliability of a single AUV.Because of its obvious advantages in distribution and redundancy,the collaborative technology of multiple AUVs has been widely concerned by research institutions in various countries.Formation control,as a typical problem of multiple AUVs coordination technology,is the basis for studying other coordination problems.As one of the most widely used formation control methods,the masterslave tracking formation,that is,the leader-follower formation method,relies on the target tracking of the leader and the coordinated formation strategy between followers.Aiming at the above problems,this thesis proposes a biologically inspired cooperative tracking formation strategy for multiple AUVs and a vision-based target tracking method for AUV.This paper reviews the research status of multiple AUVs swarm systems in recent years,summarizes the existing AUV formation methods,and analyzes their advantages and disadvantages.The target tracking method of AUV is summarized and the method of visually perceiving the target movement state is proposed to improve the adaptability of the leaderfollower formation method to the underwater environment.Aimed at the problems of low color contrast,low resolution and uneven illumination,the underwater image was preprocessed and the reflection was removed by morphological method.An underwater image acquisition system was built and the parameters of the underwater camera were corrected by camera calibration.At the same time,based on the YOLOv3 deep learning algorithm and DSST related filtering algorithm,real-time visual tracking of underwater targets is used to perceive the target's distance,direction and other motion state information.The tracking controller of AUV based on fuzzy PD is designed.Based on the kinematics and dynamics modeling of AUV,a biologically inspired Cooperative Tracking formation strategy based on fish behavior of multiple AUVs is proposed.The neighborhood topology is analyzed and three behavior rule areas are defined,which are collision avoidance behavior area,orientating behavior area,and formation maintenance behavior area.The kinematic output of the AUV is obtained according to a single behavior rule or the coupling of multiple behavior rules.The numerical simulation results under the MATLAB platform show that the designed cooperative tracking formation method based on fish behavior can realize the behaviors like formation forming,formation maintenance and obstacle avoidance of multiple AUVs under limited kinematics and dynamics constraints.In order to verify the effectiveness and feasibility of the target tracking algorithm designed in this thesis,the Dolphin series AUVs were selected to complete the vision-based AUV target tracking experiment in a pool environment,and satisfactory results were obtained in terms of tracking reliability and response time.