Research On Data Fusion Positioning Method Based On GPS And 2D Lidar

Vehicle positioning is a key issue in the research of autonomous vehicles.Only a precise positioning system can ensure the vehicle to follow the preset trajectory.At present,GPS is widely used in vehicle positioning,but the positioning error of civil level GPS is about 3-30 m,which cannot meet the positioning requirements of autonomous vehicles.However,autonomous vehicles are equipped with more abundant sensing devices,such as lidar,ultrasound,etc.,which are complementary to each other when applied to positioning.Therefore,it is of great significance for the development of autonomous vehicles to study the positioning method of multi-sensor fusion and improve the positioning accuracy.Aiming at the above problems,in order to improve the accuracy of vehicle positioning system,this paper proposes a positioning method based on data fusion of GPS and 2D lidar by using the advantages of GPS in global positioning and 2D lidar in local environment for target recognition and ranging.This method firstly USES GPS for vehicle rough positioning,and then uses 2D lidar as the means of environmental perception to conduct environmental modeling,matching and positioning in local areas.Finally,a multi-sensor data fusion method integrating GPS,lidar,vehicle speed and heading is presented.The design and implementation of data fusion positioning system based on GPS and 2D lidar are completed.The main research contents include:(1)Road environment feature modeling based on 2D lidar data.Considering the static error of the lidar and the dynamic error caused by the speed of the vehicle,an error correction method is adopted.Then,an improved DBSCAN clustering algorithm is proposed according to the data characteristics of lidar.Finally,considering the requirements and environment characteristics of the subject,the Split-Merge algorithm is optimized and supplemented.The experimental results show that the method proposed in this paper can accurately establish the characteristic model of road environment.(2)Combined positioning method based on GPS and lidar.Aiming at the problem of similar road environment in the external environment,a key grid region determination method based on GPS is proposed.Considering the linear feature of the road environment feature model,a matching method between the feature model and the map information is proposed,and a method to calculate the vehicle location through the position information of key points is proposed.The experimental results show that the proposed method can effectively match the environmental feature model with the map information and obtain a relatively accurate vehicle positioning result.(3)Data fusion localization method based on extended kalman filter(EKF).Firstly,the vehicle combination navigation and positioning model is established,and then the multi-sensor data fusion is carried out in segments in the driving process considering the success or failure of key point matching when the vehicle is driving.Comparative experiments show that the proposed method is more accurate than the traditional GPS positioning method.Finally,the integrated design and implementation of the data fusion positioning system based on GPS and 2D lidar are carried out based on the above research results.The application results show that the positioning method designed in this paper is simple and easy to operate,and the required sensors are all carried by the original autonomous vehicle without external auxiliary facilities,which can effectively improve the positioning accuracy.