Research On Indoor Positioning And Obstacle Avoidance Algorithm Of Mobile Robot Based On ROS Platfor

With the progress of science and technology,mobile robots have gradually become an indispensable part of production and life.Mobile robots replace human beings to engage in some tasks that are repetitive and not suitable for human,and it has better flexibility,higher work efficiency.For mobile robots,localization and obstacle avoidance technology are key indicators to evaluate their performance,and also the basis for mobile robots to successfully complete various tasks.Therefore,this dissertation conducts in-depth research on indoor localization and obstacle avoidance algorithm of mobile robots based on Robot Operating System(ROS).The main work is as follows:First of all,ROS is analyzed and the ROS platform is built.Then Turtle Bot3(Waffle Pi)is introduced,including its hardware structure,kinematics model and TF coordinate transformation,etc.Then the construction of mobile robot map environment is described.An improved algorithm is proposed to solve the problem that the odometer kinematic model used in Adaptive Monte Carlo Localization(AMCL)algorithm would have accumulated errors which could not be eliminated voluntarily.Based on the idea of multi-source information fusion,cubature kalman filter is used to fuse odometer and IMU data,and then a new kinematic model is constructed as the input of AMCL algorithm.Then the lidar based measurement model is used to estimate the pose of the mobile robot in the environment.Finally,an indoor localization accuracy test experiment based on Turtle Bot3 robot is designed to verify the effectiveness of the improved AMCL algorithm.When the Timed Elastic Band(TEB)algorithm is applied to traverse dense obstacle environment,the motion trajectory of mobile robot is unreasonable and the safety of obstacle avoidance cannot be guaranteed.Then,an improved TEB algorithm based on dynamic constraint of obstacle penalty weight is proposed.The algorithm adds dynamic constraints to the original optimization objective function based on the obstacle information collected by robot lidar.In this way,the robot can output relatively safe trajectory and improve the overall passing efficiency.After that,a fuzzy controller acting on the expansion radius of the obstacle is designed to further optimize the obstacle avoidance performance.Finally,a simulation experiment and a physical experiment are designed to test the safety and obstacle avoidance efficiency of the improved algorithm.In the multi-robot environment,the TEB algorithm is studied for its poor safety and unstable speed output in dynamic obstacle avoidance.By integrating TEB algorithm and Velocity Obstacle(VO),TEB-VO trajectory planning algorithm is proposed to restrict the Obstacle avoidance velocity of the robot near dynamic obstacles.Furthermore,an adaptive parameter adjustment module is designed to dynamically adjust the discrete interval of the planned trajectory and the maximum allowable linear velocity of the robot.This makes the robot’s trajectory flatten and the velocity output more reasonable.Finally,simulation and experiment based on ROS and Turtlebot3 robot verify that the improved algorithm has good trajectory planning and dynamic obstacle avoidance ability.