Development And Application Of High-definition Map Based On Autonomous Minibus In The Park

With the rapid development of the economy,people’s demand for automobiles has become increasingly diversified.After large-scale operation,autonomous vehicles can greatly alleviate road traffic congestion,improve vehicle driving safety,and can partially or even completely replace human driving.It has friendly maneuverability due to the driver’s driving action,and gradually entered the public’s field of vision and became the trend of social development.Adhering to the automotive industry development concept of "safety,comfort,efficiency and energy saving",autonomous driving technology is one of the most promising and technological innovation directions for the automotive industry in the future.As an important implementation carrier of intelligent vehicles,self-driving minibuses in the park are widely used in daily scenes such as factories and scenic spots,and have great prospects.As an important module of the autonomous driving system,high-precision map technology plays a pivotal role in the implementation of autonomous driving in the park for autonomous minibuses.This paper studies the development and application of high-precision maps for autonomous driving minibuses in park scenarios,from the experimental platform sensor scheme design,multi-sensor fusion calibration,laser inertial navigation odometer mapping algorithm construction,point cloud map semantic annotation and The four aspects of simulation application are discussed.First of all,in the design of the sensor scheme of the experimental platform,according to the functional requirements of the self-driving minibus and the characteristics of the usage scenarios,the appropriate equipment model was selected,and the sensor scheme and electronic circuit diagram of the experimental platform for the self-driving minibus were designed,installation and debugging.A self-driving minibus that meets the experimental requirements.Then,in terms of multi-sensor fusion calibration,the basic principle and significance of multi-sensor calibration time-space synchronization are firstly introduced,and then a hand-eye calibration algorithm model is established to conduct joint external parameters for Lidar and Inertial Measurement Unit(IMU).Calibration,and get the calibration parameters of the autonomous driving minibus experimental platform in this paper through the experimental data.Then,in the construction of the mapping algorithm of the laser inertial navigation odometry,the lidar odometry factor,the IMU pre-integration factor,the GPS factor(Global Positioning System)and the closed-loop factor are introduced as the state quantities participating in the factor map optimization.The lidar factor uses a key frame registration method to reduce the number of frames involved in map optimization and improve the computational efficiency of the algorithm,while still ensuring the point cloud density.The IMU pre-integration factor can reduce the IMU measurement state variables involved in the calculation in the factor graph,and constrain the continuous key frame motion state of the lidar factor in the factor graph.And by introducing GPS factor and closed-loop factor,the influence of accumulated error on mapping accuracy is greatly reduced.Then,real vehicle tests in different scenarios are designed and evaluation indicators are set to verify the accuracy and effectiveness of the laser inertial navigation odometry-based mapping algorithm constructed in this paper.Finally,in terms of semantic labeling and simulation application of point cloud map,the high-precision map is obtained by semantic labeling of the point cloud map obtained from the real vehicle experiment,and a simulation application verification scheme is designed.The MAP simulation application verification study was carried out.The test results show that the point cloud map generated by the mapping algorithm constructed in this paper has high accuracy and effectiveness,and the high-precision map developed through semantic annotation can be used for the built automatic driving simulation platform,which is feasible.