Research On The Path Planning Algorithm Of Wheeled Robot Based On ROS Platform

The popularity of robots has changed the mode of human production and life.Robots are more widely used in the service field.When robots move in the crowd,the safety and comfort are more important.In order to meet the requirements of mobile robots performing tasks safely and efficiently in complex environments,this paper makes full use of the advantages of micro-pedestrian flow model to study the path planning algorithm of wheeled robots based on the ROS(Robot Operating System).Firstly,the hardware and software platforms of wheeled robot are introduced and the environment perception algorithm based on lidar is studied.The model and simulation environment are built based on ROS platform.Aiming at the kinematic distortion of indoor robot lidar,odometry information fusion is used for correction.The algorithm is optimized for the detection of walls and pedestrians,as well as the construction and positioning effects.The simulation and physical experiments verify the effectiveness of the algorithm and make preliminary preparations for the path planning of robots.Secondly,global path planning algorithms are analyzed and compared.In order to make the path more reasonable and fully consider the radius of the robot,the global A* algorithm is improved from three aspects.The calculation method of global path is modified by double-layer expansion method.The narrow channel detection algorithm is added and the path is replanned.The generated path is smoothed and optimized at last.The effectiveness of the improved algorithm is verified by simulation.Thirdly,based on the principle of Social Force Model,an improved local path planning algorithm is proposed.Taking the improved global path as the input,the selection of local target points and the calculation method of robot control are analyzed.In order to improve the safety performance,the method of calculating the expected rate is improved.In order to avoid pedestrians more naturally,the adaptive virtual force is added to realize the dynamic adjustment in advance prediction.The corresponding simulation proves that the improved Social Force Model local path planning algorithm has a good obstacle avoidance effect.Lastly,comprehensive experiments are carried out in three scenarios of passing through a static single pedestrian and the narrow corridor,beyond a dynamic single pedestrian and meeting with a dynamic single pedestrian on the Gazebo simulation software and physical platform respectively.Experimental results show that the improved algorithms are effective.