Design Of Multi-sensor Vehicle Positioning System Based On Combined Inertial Navigation

Intelligentization is becoming the development trend and trend of the automobile industry.Autonomous driving,as one of the important connotations of intelligence,has received the attention of various automobile powers and has launched its own development strategies.Precise positioning is one of the foundations and cores of autonomous driving.It provides technical support for vehicle control and path planning.Therefore,the research on high-precision positioning has received great attention from academia and industry.Traditional single positioning solutions such as satellite positioning,inertial navigation positioning,lidar matching positioning,and visual positioning have their own advantages and disadvantages,and cannot meet the needs of autonomous vehicles for high-precision positioning.To this end,this paper proposes a 2D lidar scan matching algorithm to provide pose estimation.Based on inertial navigation elements,satellite positioning systems and lidar,a multi-sensor combined positioning system is designed.The main research contents of this article include the following:(1)Designed a 2D lidar scan matching algorithm.Lidar positioning requires lidar to provide carrier pose estimation.Due to the ICP(Iterative Closest point)algorithm's problem of low computational efficiency when searching for corresponding points,this paper uses K-D tree algorithm to search for corresponding points and reduce calculation time.On the other hand,because the ICP algorithm uses the point-point criterion measurement method to converge slowly,and there is a certain error between the point-to-point distance and the point-to-object surface distance,this paper uses the point-to-line distance as a measurement error to improve The registration accuracy is reduced and the number of iterations is reduced.The proposed algorithm comprehensively considers factors such as calculation time,number of iterations,matching accuracy and other factors,improves the ICP algorithm method,and verifies the feasibility and effectiveness of the algorithm through multiple sets of experiments,laying the foundation for INS / Li DAR combined positioning.(2)Designed a multi-sensor combined positioning system integrating INS(Inertial Navigation System),GNSS(Global Navigation Satellite System)and Li DAR(Light Detection and Ranging).First,analyze the error of the inertial navigation system;then,for the redundant and complementary characteristics of GPS(Global Positioning System)and Li DAR,GPS / INS and Li DAR / INS combined positioning are designed according to the output signals of GPS and Li DAR,respectively System;Finally,different sensor information is fused through an extended Kalman filter to compensate the INS in different scenarios.(3)Tested and verified the performance of the designed multi-sensor combination positioning system.A multi-dimensional perception wheeled mobile robot equipped with a global positioning system,2D lidar and inertial navigation elements was tested on campus to test and compare the positioning performance of different combined modes.The results show that the integrated positioning system based on INS / GPS can compensate the accumulated error of INS,and in the GPS rejection area,the integrated positioning system based on INS / Li DAR can also compensate the accumulated error of INS.The multi-sensor fusion method not only makes up for the shortcomings of a single positioning and navigation method,but also takes advantage of each sensor,so that the positioning system can run in different scenarios,and the positioning error meets the needs of autonomous driving throughout the experiment.