Research On ROS-based Indoor Omnidirectional Autonomous Navigation Robot

With the continuous development of robot technology,various types of intelligent robots continue to emerge.Among them,indoor wheeled autonomous navigation robots have been widely used in warehousing logistics,smart home,resource exploration and other fields.However,the indoor navigation scene is very complicated,and the application effect of the non-omnidirectional wheel robot is not ideal due to various constraints of its kinematic model.At the same time,traditional ROS navigation systems still suffer from poor accuracy and applicability.Therefore,according to the development status of the navigation robot in the indoor environment,this thesis studies the omnidirectional navigation robot based on the ROS operating system.The main work of this thesis is as follows:1.The current development status of indoor navigation robots and the current status of ROS research are studied,ane the indoor omnidirectional navigation robot is systematically modeled.The basic framework of the omnidirectional navigation robot is designed,and the related motion model,sensor model,coordinate system conversion,map model are studied.2.The commonly used indoor SLAM algorithms and positioning algorithms are analyzed.Aiming at the problem that the commonly used SLAM algorithms are too dependent on the accuracy of the odometer or the radar frame rate,Cartographer was finally selected as the indoor mapping algorithm.It can effectively reduce the cumulative error through closed-loop detection,and it can also achieve good results by using lower cost radar.In indoor positioning,a probabilistic positioning algorithm based on particle filtering is studied,and simulation analysis is performed.In order to provide accurate odometer information for mapping and positioning,this thesis builds an odometer with encoder fusion IMU based on EKF.3.The commonly used robot path planning algorithms and trajectory optimization algorithms are analyzed.Aiming at the problem that the traditional global path planning algorithm has too many sharp turns or corner turns caused by searching in the grid and the shape of the obstacle,which affects the local planning and control accuracy of the robot,based on the application scenario of the omnidirectional navigation robot,this thesis uses IPOPT non-Linear optimization method to smooth the global path.Next,by constructing the trajectory smooth objective function and constraint function,a trajectory smoothing module is constructed.In local path planning,the DWA algorithm based on omnidirectional motion model is studied,and the DWA algorithm is improved in ROS.4.Mecanum wheel omnidirectional navigation platform is built to conduct navigation test in indoor actual environment.The experimental results verify that the motion control system designed in this thesis has high control accuracy;the positioning accuracy and pose tracking effect of the indoor positioning method integrated with the odometer meet the navigation needs;the trajectory smoothing module based on nonlinear optimization can quickly solve smoother trajectories,improves the planning and control effect;the improved local path planner enables the omnidirectional robot to complete the autonomous navigation task in the indoor environment and has high adaptability.The paper has 74 pictures,18 tables,and 63 references.