Research On Fast Path Planning Algorithm For Dynamic Target In Uncertain Environment

With the rapid development of computer technology,mobile robot has been playing an increasingly important role in many areas,especially in the military,industrial,underwater exploration,disaster relief,and other fields.Due to path planning for mobile robot autonomous navigation in the process of the research is the core technology,it has been widely concerned by many scholars.This paper first introduces the current research status of mobile robot and path planning,and then explains the system structure and coordinate system of four-wheel omnidirectional mobile robot.The kinematics model and dynamics model of the mobile robot is establishes through analyzing its geometric relationship.As it is applied to fire rescue and other harsh conditions,the external load sensor of mobile robot selects the laser sensor with high precision and strong anti-interference.The sensing quantity of surrounding environment information obtained by the laser sensor is modeled.SLAM and navigation are involved in the mobile robot navigation process.Aiming at the problem of SLAM,firstly the principle of SLAM system is described and the particle filter algorithm applied in the process of mapping and localization is introduced.On the basis of the particle filter to create a map,RBPF algorithm was improved in the importance of probability density is more close to the target distribution,and selective resampling of particles can solve the problems of large particle number and serious particle degradation in RBPF algorithm.At the same time the adaptive monte carlo algorithm is used to locate accurately in the map.Both the effectiveness of the algorithm is verified by experiment.Aiming at the path planning problem in the navigation process of mobile robot,a kalman filter and D* Lite algorithm for dynamic target point path planning is proposed.The improved A* smooth path algorithm and D* Lite algorithm were used for the path planning of single and multiple target points in static and dynamic environments.The two algorithms above are used for global path planning on the constructed global raster cost map.The dynamic window algorithm is adopted to make the mobile robot safely avoid the sudden obstacles in the process of movement so as to realize the local path planning.Both can be applied in the collision-free path planning from the starting point to the target point in the map environment.Finally,the navigation system of the simulated mobile robot is built on the Ubuntu Kinetic ROS platform,and the single-target and multi-target path planning experiments are achieved on the basis of map construction and robot positioning in the experimental environment.Experimental results show that the path planning algorithm proposed in this paper is effective on the simulation robot navigation platform.