Research On Path Planning Of Four-wheeled Omnidirectional Mobile Robot

Along with the rapid development of robotics,mobile robots are becoming more and more frequent in everyday life and factory work.The thesis addresses the study of omnidirectional mobile robots proposed for life and industry that can move autonomously to perform specific tasks,such as monitoring the environment and carrying small objects.It analyses the kinematic equations of the Mc Namee wheel and the steering relations of its omnidirectional moving wheels,conducts research on simultaneous localisation and map creation(SLAM)and path navigation planning for mobile robots,designs an autonomous navigation system using ROS as the software basis,and also builds a four-wheeled omnidirectional mobile robot prototype to verify the autonomous map construction and path planning functions.The main research work includes:First,it was determined that the Mc Namee wheel was selected as the moving mechanism.The overall structure of the omnidirectional mobile robot is designed.Based on the mathematical model of the Mc Namee wheel,the relationship between the omnidirectional motion and the steering of the Mc Namee wheel is analysed.The design uses the PID method to achieve omnidirectional motion control of the Mc Namee wheel.Afterwards,the research and design of autonomous navigation of the omnidirectional mobile robot is based on the ROS system adopting a modular concept,planning and designing three parts: sensor sensing part,SLAM part and path planning.This is followed by a study and analysis of sensors including airplane projection and odometry models,the selection of IMUs and suitable Li DAR.Finally,the coordinate transformation of the ROS visualisation is highlighted to explain how the data can be fused and applied in different coordinate systems,laying the foundation for subsequent autonomous navigation research.Then,simultaneous localisation and map construction techniques for mobile robots are investigated.The thesis begins with an introduction to three environmental map representations,identifying the choice of raster map representation.Two different SLAM algorithms are investigated,based on the improved RBPF algorithm,which firstly improves the proposal distribution and reduces the number of particles,and secondly improves resampling;based on the optimised Hector?SLAM algorithm,which uses extended Kalman filtering for the fusion of odometry,Li DAR and IMU sensor data;the two algorithms are compared and analysed through real environment construction experiments to verify the The practicality of the optimised Hector?SLAM-based algorithm was verified by comparing the two algorithms to create a clear and neatly contoured map.This is followed by a study of the path planning part of the mobile robot.For the commonly used global path planning algorithms Dijkstra's algorithm and A* algorithm are analysed and explained in principle,and the two algorithms are verified through simulation experiments,and the more functional A* algorithm is chosen as the global planner for the thesis after comparison.For local path planning,the Dynamic Window DWA algorithm is used,and the working principle of the Dynamic Window algorithm is verified using the MATLAB platform.Finally,a prototype four-wheeled omnidirectional mobile robot was built to experimentally validate the previously studied SLAM technique and path planning navigation.The optimised Hector?SLAM algorithm can effectively construct a laboratory 2D raster map with a complete and accurate map representation.In the simulation experiments,the A* global path algorithm and the DWA dynamic window algorithm have autonomous navigation functions in the laboratory-based simulation environment,and can plan and move to the specified target location on their own,with functional integrity and good practicality.