Research On AUV Autonomous Decision-Making System In Large Scale And Complex Marine Environment

Autonomous underwater vehicle(AUV)has unique advantages and broad application prospects in exploring marine resources and performing marine tasks.More and more scholars are devoted to the research of AUV's autonomous decision-making systems and related algorithms.AUV has the advantages of high flexibility,low energy consumption and small size in the marine environment,and because it is equipped with a certain degree of autonomous decision-making system,AUV has the ability to autonomously complete a variety of marine tasks with very few people involved.However,the current degree of AUV autonomy is generally not high,especially in large-scale and complex marine environments,the ability to complete tasks autonomously is not strong.Therefore,this paper takes the AUV decision-making system as the research object,and focuses on the autonomous decision-making system architecture,task planning and path planning algorithms that can adapt to large-scale and complex marine environment,and develops a set of AUV autonomous decision-making system software,which has achieved the following achievements:(1)Research and design an AUV autonomous decision-making system architecture that can adapt to large-scale and complex marine environments.Based on the classic hybrid three-layer decision-making and control architecture,the AUV decision-making system architecture including decision-making layer,planning layer and control layer is designed according to the project requirements.The system is divided into decision-making unit,navigation unit,control unit,communication unit and monitoring unit.The function division between each unit and the definition of data transmission interface between them are clear.The system effectively combines high-level deliberation decision-making and low-level real-time planning,which enhances the ability of AUV to perform missions safely and autonomously in large-scale and complex marine environments.(2)The research on the establishment of mission planning model and algorithm verification in large-scale marine environment is carried out.The task planning problem in AUV large-scale environment is modeled as the optimal route search problem in the task network graph that satisfies the time,task quality and quantity constraints.The model is solved by several optimization algorithms.The results show the correctness and effectiveness of the model,and the ant colony optimization algorithm has better performance when solving the model.(3)The research on the establishment of path planning model and algorithm verification in complex marine environment is carried out.The B-spline curve is used to represent the smooth local path of AUV in complex ocean environment.Considering the influence of ocean current on AUV navigation energy consumption,the problem is modeled as an optimization problem with multi-constraint condition with minimum time,no collision,smoothness and less energy consumption.The model is solved using the swarm hyper-heuristic algorithm and compared with the traditional firefly algorithm.The results show that the swarm hyperheuristic algorithm has better performance in terms of time stability and solution results.(4)Software integration of each unit of AUV's autonomous decision-making system and simulation test verification.The system software is implemented based on ROS,including a decision node and two communication nodes.The decision node implements the functions of all units except the communication unit,and internally uses the multi-thread technology to realize the functions of each unit.The communication node implements communication between the system software and the command center.The software was tested and functionally verified,indicating that the software correctly completed the intended function.