Study On Path Planning And Dynamic Obstacle Avoidance Of Wave Glider

With the prosperity of the unmanned maritime navigation system in the major countries of Europe and America,as well as the necessity of the strategic adjustment of China’s maritime power and navy,the development of domestic unmanned maritime autonomous vehicles is imminent.As one of the unmanned autonomous vehicles in Marine environment observation,wave glider has a wide range of applications.It can carry a variety of sensors as required to carry out Marine environment data acquisition,target surveillance,sea surface,water and air communication relay and other work.In the future,wave gliders will play an important role in Marine environment monitoring,Marine resources survey,sea rescue and other fields.Firstly,this thesis designs a wave glider control system with industrial computer as the core controller.The system is mainly composed of software and hardware.The hardware includes shore-based monitoring system and wave glider terminal control system,and the software includes Beidou wireless communication,positioning analysis program,driver program and main control program.There is no need to write complex communication protocols between each layer.The system test also proves that the system has strong feasibility,and also lays a platform foundation for the path planning research in this thesis.The path planning of wave gliders is more dependent on the Marine environment than traditional robot path planning.Firstly,quadtree and simulated force field are used to establish mathematical environment models for global path planning and local path planning respectively,which makes the formulation of path planning strategies more efficient and reasonable.Secondly,for Biologically Inspired Neural Network(BINN)algorithm,this thesis analyzes the global path planning problem of wave glider,and fully considers the difference between the actual sailing path of wave glider and the algorithm planning path.In view of the problems that the path change Angle is too large and the activity value is the same in the algorithm,this thesis uses the heading control analysis and distance influence mechanism to improve and study.Aiming at the influence of sea current on the sailing path of wave glider,this thesis uses the heading weighting scheme to carry out in-depth exploration,and uses the Bezier curve to optimize the broken path.In this thesis,Artificial Potential Field(APF)method is used to study the dynamic obstacle avoidance problem of wave glider.Aiming at the inaccessible problem of target point in the algorithm,distance factor coefficient is optimized to solve the problem.Considering the motion characteristics of wave glider and the motion constraints of wave glider,and considering the influence of wave on path planning,this thesis solves the problem by adding wave virtual force into the repulsive potential field function.Finally,this thesis conducts simulation and sea experiments for the improvement of the algorithm,and the experiments confirm that the path planning of wave glider by using the improved Biologically Inspired Neural Network algorithm and artificial potential field method will not have a large change of Angle in the planned path,but also improve the situation of falling into the dead zone.Dynamic obstacle avoidance can avoid the threat of static and dynamic obstacles under the influence of waves.