Study On Path Planning And Location Of Mobile Robot Based On Intelligent Optimization Algorithm

Mobile robots can work in complex environments,with the ability of self planning,self-organizing and self-adaptation,which are widely used in important military and civilian fields,especially in severe environments.Autonomous navigation is a core technology to improve the perception and action ability of mobile robot.Path planning and location methods are critical parts of mobile navigation system to achieve autonomous and intelligent mobile function effectively.With the increasing complexity and high dimension of the obstacle environment,higher and more urgent requirements have been proposed for the environment adaptation and planning efficiency of the mobile robot navigation system than before.Therefore,in order to improve the performance of mobile robot navigation system,the path planning and locating methods in the intelligent navigation system based on intelligent optimization algorithms are studied in this paper.(1)For the two dimensional(2D)global path planning problem in the known model environment,a smooth path planning method based on dynamic feedback A*ant colony algorithm is proposed on the basis of the grid environment model and the basic ant colony algorithm.The improved ant colony algorithm is mainly considered from the three aspects:the optimization of the initial pheromone,the increase of the evolutionary strategy and the dynamic closed-loop adjustment of parameters based on the ant colony search results.After obtaining the planning path,B-spline curve is used for smooth processing.The simulation results show that the improved ant colony algorithm can effectively improve the search efficiency of the path and avoid the local minimum.Meanwhile,the improved algorithm can get a shorter planning path compared with standard ant colony algorithm.(2)Because the problem of three-dimensional global planning is more complicated than two-dimensional planning,kinematics and dynamics constraints are more complex.For the three dimensional(3D)global path planning problem in the known model environment,taking the fixed wing unmanned aerial vehicle(UAV)as an example,the mathematical model of the 3D global path planning is set up.Meanwhile,the evaluation functions of the whole path and the single segment path are given,and the basic particle swarm optimization algorithm is analyzed.Because the 3D global path planning of the fixed wing UAV belongs to the multi-objective optimization problem with constraints,the 3D global path planning of UAV with standard particle swarm optimization(PSO)is slow and easy to fall into local minimum.Then an improved PSO method based on global optimal path competition is proposed.For the improved PSO algorithm,a candidate optimal path is found according to a single waypoint path evaluation function as the evaluation criterion based on a single path point.The optimal path is selected by comparing with the other candidate path by considering all the waypoints as a whole.The simulation results show that the improved PSO algorithm has a fast convergence speed in the 3D global path planning,and it can find a collision-free optimization path in different environments with good robustness.(3)For the local path planning problem with unknown or partly unknown environments,a path planning method based on improved dynamic window method is proposed.The basic dynamic window method is analyzed from the three aspects:robot motion model,speed search space and evaluation function.Because the evaluation function of the standard dynamic window method uses the fixed weight parameter combination in the face of different environment obstacles.In order to improve the planning effect in different environments and reduce the influence of weight parameters on the performance of the algorithm,the laser radar data is introduced to measure the distance between the obstacles,comparing with the size of the mobile robot.The weight parameters are then selected by fuzzy control with the comparison result.The simulation results show that the standard dynamic window algorithm can not reach the target position through the obstacles,whereas the improved method can avoid the obstacles to reach the target,improving the environmental adaptability adaptability of the standard algorithm.(4)For the problem of simultaneous localization and mapping(SLAM),the motion model and observation model of mobile robot are established,and the FastSLAM algorithm is analyzed.In order to reduce the computational complexity,the FastSLAM algorithm is used to estimate the pose of mobile robot and the map characteristics by Rao-Blackwellized decomposition using particle filters and Kalman filters,respectively.Due to the problem of particle degradation and consumption in the particle filtering algorithm,the improved particle filter algorithm is proposed based on improved cuckoo search(CS)algorithm.In order to increase the diversity of the cuckoo population and improve the search efficiency,a multi strategy difference evolution algorithm is introduced to improve the preference random walk search strategy for the CS algorithm.The simulation results show that the FastSLAM method based on improved CS algorithm can effectively improve the accuracy of the position estimation and the environment feature estimation of mobile robot.For the autonomous navigation problem of mobile robot,several path planning and location methods are studied and improved,including ant colony algorithm,particle swarm optimization algorithm,dynamic window algorithm,particle filter and FastSLAM algorithm.The improved algorithms in this paper can improve the navigation performance of the mobile robot system,which can provide reference for autonomous control of mobile robot.