Research On Autonomous Positioning And Navigation Technology Of Autonomous Vehicle In Urban Road Environment

Autonomous localization and navigation technology is the technic core of unmanned vehicles,which is also a hot topic in the field of unmanned vehicles in recent years.In a urban road condition,because of the complexity and unpreditity of its traffic conditions,the unmaned vehicles are expected to improve its security,stability and economy.However,there are still some restrictions with current autonomous localization and navigation methods.This paper focuses on the autonomous localization and navigation thchnology on urban roads,and the main research contents are as below:(1)The sensor models of camera,IMU and solid-state LiDAR are studied respectively,and a joint calibration of camera/IMU and lidar is carried out.Then,a SLAM framework that combines visual,IMU and lidar is proposed and an incremental smoothing factor graph optimization method is adopted.The factor graph is saved in the Bayesian tree.When a new factor node is added,the affected variable nodes can be identified and optimized so as to maintain the fully optimized sparsity characteristics and achieve the goal of multi-sensor fusion localization and mapping.A global factor graph is constructed,and the starting factor,registration factor,IMU factor are inserted into the factor graph by each model.When a new factor node is inserted,it is optimized and updated through the factor graph optimization algorithm to achieve the coupling of more sensors and to provide reliable position,gestures and map information for the perception of autonomous vehicles.The algorithm is tested in the KITTI dataset,real urban road conditions and extreme tunnel conditions,and it’s also compared with algorithms such as A-LOAM,LOAM-Livox and VINS-Fusion.The experimental results show that the designed autonomous localization and map construction system that conbines fusion vision,IMU and laser radar can achieve high accuracy and high precision in complicated conditions,and it still has high real-time and robustness under the premise of ensuring accuracy.(2)In order to realize the goal of the autonomous navigation of the vehicle,a dynamic model of the Ackerman vehicle is analysed at first,then the constraints that it needs to meet is calculated,and the map of obstacle in static state can be obtained through processing the constructed high-precision map.By detecting dynamic obstacles and estimating continuous time state,the information of localization,speed,and category of the obstacles are calculated,then the dynamic obstacle constraints are added to the TEB trajectory planner to achieve real-time and highly safe trajectory planning.After the targeted trajectory of autonomous vehicles is generated by the trajectory planner,the trajectory is tracked by the trajectory tracker based on the kinematics model,and the smooth time-varying feedback control is implemented on the vehicle to realize the point-to-point autonomous navigation of the vehicle.Finally,the autonomous navigation method is verified through an unmanned vehicle experimental platform.The experimental results show that the designed autonomous localization and navigation system is correct and effective.