Research On Localization/Obstacle Avoidance And System Implementation Of Outdoor Mobile Robot

With the rapid development of technologies such as sensors,computers,and artificial intelligence,mobile robot technology has also been rapidly developed and widely used.It has also become a research hotspot in the academic and industrial circles.However,it is not easy to play the "moving" characteristics of mobile robots in an outdoor environment.The acquisition of the pose of the robot and the avoidance of obstacles are important links for achieving autonomous movement.The main work of this paper is to research the localization technology and obstacle avoidance technology of mobile robots in outdoor environment,and develop a robot experiment platform with autonomous mobility(called outdoor mobile robot platform)based on the above research.Firstly,the system requirements of the outdoor mobile robot platform are analyzed.According to requirements,the mechanical structure of the platform and the embedded system hardware structure is designed,and the navigation system framework is determined.Based on the above framework,the coordinate system of the mobile robot platform navigation and the kinematics of the chassis are established.The mathematical model of the model and the corresponding sensor are also established.At the same time,a simulation platform was built based on the Gazebo simulator.Secondly,for the outdoor localization problem of mobile robot platform,the extended Kalman filter fusion localization algorithm is firstly adopted.By establishing the motion model and the measurement model,the satellite positioning data,inertial measurement data and odometer data are combined.The simulation analysis by Gazebo results show that the method in the case of good satellite signals,positioning accuracy reaches decimeter level.For the problem that the satellite signal is blocked or electromagnetic interference causes the fusion positioning method to be invalid,this paper uses the particle filter-based probabilistic map positioning method to obtain the robot pose.At the same time,the mode switching method of fusion location and probability map location is proposed,which makes the outdoor location method of mobile robot platform more robust.Then,for the local obstacle avoidance problem of mobile robot platform,this paper adopts the obstacle information acquisition method that combines stereo camera data and single-line lidar data.So the system has the ability to detect obstacles of different heights and 360 degrees.In the aspect of obstacle avoidance planning algorithm,this paper proposes a local planning algorithm based on multi-constraint optimization.The algorithm comprehensively considers the obstacle constraint,the path constraint,the velocity constraint,the geometric constraint and the minimum constraint time of the trajectory execution time,and establishes multiple objective functions which are deformed by the above constraints,and uses the graph optimization method to solve the optimal trajectory.The simulation results show that the proposed algorithm has a higher success rate than the dynamic window approach.Finally,the autonomous navigation system of the outdoor mobile robot platform integrated with the above positioning and obstacle avoidance methods is developed,and the autonomous navigation system scheme is implemented in the simulator.At the same time,the outdoor positioning function and obstacle avoidance function of the autonomous navigation system are verified by the prototype of the mobile robot platform.