Research On Path Planning For Mobile Robot Based On ROS

With the development of science and technology,the path planning technology of mobile robots has received extensive attention from researchers.At present,the calculation of global path planning is too complicated.The planned path is sometimes not conducive to robot running.In addition,Local path planning has low real-time performance.In order to solve these problems,TurtleBot platform based on the ROS of mobile robot as the research object,the path planning technology of the mobile robot has been theoretically analyzed and simulated,and the navigation hardware,software system construction and path planning experimental verification have been completed.We use A* algorithm as a global path planning of search algorithm based on map construction and location.Because there are many break points in the traditional global planning.the path is not smooth and there are too many problems in searching for grid points.So we use GUI to design algorithm simulation interface in MATLAB and there are three improvements to the global planning algorithm.In order to reduce the number of search grid points and reduce the planning time,a threshold value N is added to the openlist.If the number of search times is greater than N and the first node inserted is not expanded,the node is set to the highest priority.We use the path planning based on the combination of Floyd algorithm and A* algorithm to remove redundant nodes and reduce the sharpness of the breakpoints.Finally,the geometric optimization process results in a smooth path,and the improved simulation path is reduced by more than 5 unit lengths.On the basis of successful using global path planning,the DWA algorithm is used to solve obstacle avoidance problems for robots facing unknown obstacles.Local path planning is verified on MATLAB simulation maps.The three parameters of the trajectory evaluation function evaluation index ?,the speed evaluation index ?,and the obstacle distance evaluation index ? are optimized.The simulation results show that when(28),02.0?,0?(28)0.4,?and(28)the real-time performance is the most excellent.The hardware experimental platforms such as Kinect and Kobuki chassis were first integrated,and the overall framework of the navigation system based on the ROS system and the scheme design of each node were set up for completing the navigation system experiment.Then in the global_planner function package we use C++ code to design the node corresponding to the A* algorithm,adding AstarExpansion::Floyd(),AstarExpansion::smoothing()to improve the algorithm function.At the same time in thedwa_local_planner function package the DWA algorithm node is designed and python to write multi-target navigation script is used.Finally,under the well-established ROS robot platform,the global path planning experiment of A* algorithm based on Gmapping map,the obstacle avoidance experiment of local path planning of DWA algorithm,and multi-target navigation experiment were implemented.The experimental results were analyzed.Simulations verify that the improved A* algorithm reduces the path length and the path is smoother;the simulation results of the optimized evaluation function parameters for the DWA algorithm show that the real-time performance of the local planning can be improved.The navigation system test based on the Gmapping algorithm map construction verifies the feasibility of the path planning.The simulation and experiment based on ROS of robot platform are basically the same.It proves that the path planning technology of this paper can be well applied to mobile robot navigation.