Study On Distributed Formation Control Of Multiple Autonomous Underwater Vehicles

In recent decades,the autonomous underwater vehicle(AUV)technology has made great progress and has played an increasingly important role in various fields such as ocean hydrological data collection,sea bottom exqploration,and target search.With the continuous improvement of task requirements,a single AUV is unable to meet the increasingly complex operational needs.The multiple AUV system can be distributed in different spatial locations in the detection area,and the task can be executed coordinatly with each other,and the system composition and the sensor configuration can be redundant.With these advantages,multiple AUV system has tremendous advantages over a single AUV in large-scale,rapidly-changed marine exploration activities.In the past two decades,the theory and application research of multiple AUV system has increasingly attracted the attention ofrelevant researchers.This dissertation focuses on a key and basic problem in multiple AUV system,which is the coordinated formation control of multiple AUVs.The research work of the dissertation mainly includes the following aspects:1.The research work on multi-AUV formation control algorithm architecture was carried out.Based on the results of the sea trials and analysis of the virtual leader-follower formation control algorithm,the virtual leader and formation location vector are abstracted as the formation organization layer,forming the algorithm framework of this paper.This dissertation adopts the formation control algorithm of the flocking formation control,the artificial potential field and the virtual structure combined formation control algorithm to organize the formation formation.2.A multi-AUV flocking control algorithm based on target tracking was developed.The multi-agent flocking algorithm was used to organize the formation,and a target tracking algorithm for underactuated AUV was designed.The multi-agents form a quasi-lattice formation around the virtual leader.The AUV tracks the corresponding agents and gradually generates the same formation The robustness and scalability of the AUV formation control algorithm was enhanced.3.The study of multi-AUV flocking control algorithm based on feedback linearization was carried out.The feedback linearization method was used to reduce the AUV motion model into a particle-like model,and was further decomposed into a translational motion subsystem and a sway motion subsystem.The flocking algorithm was used to control the translational motion subsystem,and a phase consensus algorithm was used to control the sway motion subsystem.4.The decentralization of the virtual leader's states calculation was carried out.The virtual leader's calculations are migrated from the information center to each AUV in the system.A distributed information consensus algorithm was designed to make the virtual leaders' states of different AUV calculations consistent,which makes the formation control algorithm truly distributed.5.A formation control algorithm combining artificial potential field and virtual structure is designed.In this algorithm,the artificial potential field was used for collision avoidance and obstacle avoidance,and the virtual structure was used to organize furmatioan.The combination of the two algorithms served as the formation organization layer.The AUVs trace the corresponding intermediate media points,and generate a specified formation or make formation transformation without collision gradually.6.The research on obstacle avoidance of multi-AUV formation was carried out.The obstacles which could be encountered in sea trials were abstracted into three types,and corresponding obstacle avoidance algorithms were designed.Four obstacle avoidance strategies were designed for the formation,which can be selected based on different formation avoidance scenarios.