Research On Path Planning Algorithm Of Mobile Robot In Complex Environment

With the fusion of autonomous navigation,intelligent control and sensor technology,mobile robots are now becoming more and more "smart".At present,mobile robots can be seen in many application fields.However,the path planning process is a prerequisite for realizing various activities of mobile robots,and it is also an important manifestation of robot intelligence.Under normal circumstances,based on the differences in the environmental status information of the robot planning area,path planning problems are subdivided into two categories: global planning problems and local area planning problems.In fact,the actual application scenarios of robots often contain both global static environmental information and local dynamic obstacles.Therefore,how to achieve path planning tasks in complex environments is a very challenging scientific research topic.In order to successfully apply path planning to mobile robots in complicated scenes,a new path planning algorithm is designed,which combines the genetic algorithm and the artificial potential field algorithm to form a compound double-layer path planning algorithm.The main work of this paper is as follows:(1)Apply the improved genetic algorithm to the path planning task of mobile robots in the global environment.First,to solve the problem of infeasible paths or feasible paths with redundant nodes in the population initialization process of the traditional genetic algorithm,an effective population initialization improvement scheme is proposed.Secondly,in order to overcome the phenomenon that the movement path planned by the mobile robot in a terrain environment with many obstacles is prone to tortuous back and forth,a path smoothness function is added to the standard form of genetic algorithm.Besides,the traditional fitness function is transformed to improve the performance of the designed algorithm.Finally,through the simulation of planning in four different environments,the experimental results verify that the improved genetic algorithm is effective.(2)For locally path planning for mobile robots,the artificial potential filed method is adopted.Aiming at the problem of unreachable targets and local minimums in the algorithm,a new method is proposed.By adding the coordination factor and relative motion speed information to the standard form of repulsive force field function,the problem that the mobile robot cannot reach the target position can be effectively solved.Secondly,this article puts forward the selection plan of the virtual target point,and gives the derivation formula of the reasonable target point in detail.Introducing a virtual target place can help the robot not go into a local minimum point.Finally,the availability of the improved method is verified through the simulation experiment of multi-robot formation and multi-dynamic obstacle avoidance.(3)To address the trajectory planning of the robot in complicated environment,which is tougher,a new hybrid double-layer control method is designed.Utilizing the difference between the global and the local planning method,the two-level planning idea is adopted in this paper to better ensure the overall performance of the hybrid algorithm.Firstly,an improved genetic algorithm is used to pre-plan the global environment,and the child nodes of the global optimal path are solved and saved.Secondly,with the help of the information from the sub node,the dynamic obstacle avoidance task in local environment can be achieved.At the end of the thesis,the related path planning experiments have verified that the hybrid two-level path planning algorithm is feasible.