Research On Path Planning Technology Of Mobile Robots Using ROS In A Dynamic Environment

Path planning technology is an important research field of mobile robots. Based on actual needs, this paper is committed to obstacle avoidance and path planning research in order to make the robot arrive at the specified target safely and smoothly in a complicated dynamic environment. Dynamic obstacle avoidance is a difficult problem in the field of robotics which has not been really resolved for years. At the same time some problems such as oscillation and target-unreachable still exist in the robot path planning.In this paper, some improvements of algorithm have been made to solve problems existing in path planning and they were finally proved by the robot built by myself. The main research contents in the paper are listed as follows:1.The robot based on ROS( Robot Operating System) was built in order to test and verify algorithm rapidly. First of all, analyze the kinematic model of the robot to compute its odom. Then design its motion control system to make the robot be able to subscribe velocity command and publish odom information. At last, build the robot model in the format of URDF(Unified Robot Description Format) and the node of laser. With all the work done, the robot with basic function of navigation is built.2.Use ant colony algorithm as the robot global path planning algorithm. To make the algorithm converge faster, improve the algorithm in terms of both initial pheromone and the pheromone-update model. A method is also proposed to overcome the circumstance of trapping into local optimum. At last optimize the final path. The simulation results show the feasibility and advantages of the improved algorithm.3.To meet the demand of high real-time in local path planning, select the artificial potential field as local path planning method. The approach of providing partial target point on the global path is proposed to allow the robot to walk along the global path. Furthermore, an adaptive adjustment mechanism makes the algorithm not easy to fall into local minimum and smooth the path of robot.In response to the dynamic obstacle, repulsion potential function is improved by adding relative velocity information between the robot and an obstacle, which makes the robot be able to avoid dynamic obstacles quickly.The simulation results show that the improved algorithm not only make the robot walk along the planning global path, but also has better dynamic obstacle avoidance performance.4.With the ROS-built robot, some experiments were done respectively in a single obstacle, multiple obstacles, dynamic obstacles and high dynamic obstacles environment. The experiments proved the feasibility of the improved algorithm.