Study On Multiple Destinations Path Planning Based On The Spherical Amphibious Robot

With the increasing demand for high maneuverability,long endurance,energy saving and even stealth of autonomous underwater vehicles,bionic amphibious robot based on biological system idea shows better characteristics in adapting to underwater,land and air environment.And amphibious robots are widely used in water patrol,military reconnaissance,environmental monitoring,resource development and so on.Path planning is an indispensable technology to achieve these tasks.The core problem of bionic robot includes path planning,which is also an important factor of artificial intelligence.Path planning is to find a way from the starting point to avoid obstacles and finally reach the end point in an environment full of all kinds of obstacles.Firstly,in order to make the spherical amphibious robot respond correctly to the environment and complete path planning,this paper presents a multi-objective path planning method based on spherical amphibious robot.Implanting embedded system in spherical amphibious robot,laser sensor is selected as the main sensor to collect environmental information and can detect the minimum safe distance,the overall system hardware structure and software module are designed.It makes the spherical amphibious robot easier to detect obstacles.Secondly,in the path planning problem of spherical amphibious robot,multiple target points are set in complex environment,and the distance between each target is obtained by A~* algorithm.Then,the generated path fusion Floyd algorithm is optimized for bidirectional smoothness,so as to reduce the turning points and path distance,Finally,the sequence of reaching each target points is solved by integrating ant colony algorithm,the spherical amphibious robot passes through multiple targets from the starting point,which avoid static obstacles along the shortest path and finally return to the starting point.Then,the sub-target points are dynamically set on the basis of the global static multiple targets path,and the local dynamic obstacle avoidance is realized by integrating the dynamic window method in combination with the path situation.Finally,the robot can not only walk along the planned global static path,but also avoid obstacles in real time during the path,then return to the global static path and go ahead,the spherical amphibious robot completes the global static and dynamic multiple targets path planning.Finally,based on the simulation,the algorithm is transplanted to the spherical amphibious robot,and the path planning experiment in the real environment is completed by using the spherical amphibious robot.In the experiments of land and underwater path planning,the spherical amphibious robot can safely avoid obstacles,reach each target in turn,and finally return to the starting point.In this paper,NDI optical tracking system is used to collect data in the experiment of spherical amphibious robot,and the characteristics of land experiment and underwater experiment are evaluated.The experimental results further verify the feasibility and effectiveness of the algorithm.