| Nowadays,with the rapid development of robot industry,people’s quality of life is getting higher and higher,and indoor service robots begin to enter people’s daily life.Among them,robot path planning and map construction have been the core of indoor mobile robot technology.This paper mainly studies the path planning method of VSLAM robot based on visual sensor,which can be used as a service-oriented autonomous navigation robot in indoor places such as restaurants.In order to improve the navigation efficiency of mobile robot path planning method and achieve good dynamic obstacle avoidance effect,the path planning method is optimized and improved.Firstly,the environment model is established in ROS.The depth map collected by the depth camera is transformed into dense point cloud map and octree map,and then the octree map is projected to the plane to obtain the occupancy grid map for robot path planning.RRT(rapid exploring random tree)has strong exploration ability in unknown environment.In order to facilitate independent exploration and mapping,a frontier unknown point detector based on RRT method is designed to select the optimal frontier point to participate in independent environment detection.In the improvement of the global path planning algorithm,the goal paranoia strategy with variable probability is introduced.For the growth of fast random tree,the direction of the goal is added to make the search direction more targeted.In order to solve the problems of many inflection points,complex redundant points and slow convergence speed of traditional RRT algorithm,the optimization method of removing redundant points is used to improve the convergence speed of RRT algorithm.The cubic B-spline method is used to smooth and optimize the target curve.In order to improve the obstacle avoidance ability of the robot in the dynamic environment,the optimal speed sampling simulation trajectory in DWA algorithm is selected,the evaluation function mechanism is established,and the normalization processing is used to improve the standard evaluation function,and finally the optimal evaluation parameters are obtained.The experiment is carried out on the MATLAB simulation platform.Compared with the optimized RRT method,the planning efficiency is improved by 44.89%,and the path smoothness is significantly improved.It meets the performance requirements of optimizing the length,time and search efficiency of RRT path.Combining the improved RRT with the optimized DWA,the efficiency of robot path planning is significantly improved.In order to verify the feasibility of the optimized algorithm in the actual environment,the robot hardware platform is built,mainly including the robot chassis,on-board computer terminal,host computer and depth camera.Software programming is developed in ROS(robot operating system).It realizes the combination of multiple computers,that is,the robot terminal realizes the positioning of the robot,and sends the relevant data such as the position of the robot to the upper computer in the form of topic through the ROS platform,so as to realize obstacle avoidance mapping and path planning on the upper computer.The improved RRT method has significantly improved the search efficiency and path quality compared with the traditional RRT.In the indoor dynamic environment,the improved RRT method combined with DWA path planning can achieve dynamic obstacle avoidance,and the running time is reduced by 39.53%,which proves the effectiveness of the improved method in this paper. |
