Research On Trajectory Tracking For Autonomous Driving Based On GPS/INS Integrated In-car Navigation Technology

With the advancement of the Internet,big data,communication technology,and artificial intelligence,modern society is accelerating towards intelligence,and autonomous driving technology is also taking advantage of this opportunity to rapidly develop and will play an important role in solving the current traffic congestion and safety problems Trajectory tracking control,as one of the key technologies for autonomous vehicles,is the basis for ensuring that vehicles can drive safely and effectively in complex and changing environments.In order to meet the requirements of higher-level autonomous driving,the accuracy and smoothness of trajectory tracking control need to be further improved.This requires not only research on more effective control algorithms,but also the support of high-precision positioning systems.Under this background,this subject proposes the research on trajectory tracking for autonomous driving based on GPS/INS integrated in-car navigation technology according to the existing experimental conditions,and finally realizes the accurate and stable following of the test vehicle to a given desired trajectory The main research contents of this article include(1)This paper improves the traditional Model Predictive Control(MPC)-based lateral motion control algorithm.Based on the established dynamic model in terms of error with respect to road,this paper proposes a prediction model that takes into account the lateral displacement preview.This makes the MPC lateral motion controller approximately achieve the equivalent preview time using only half of the original prediction horizon,which can effectively reduce the computational burden.In addition,the idea of optimal preview control was introduced to utilize the future path information,and the steady-state steering angle of the vehicle was used as the reference control input to gain the effect of feed-forward control.Compared with the traditional MPC lateral motion controller,the path tracking accuracy is higher,and the algorithm has better real-time performance.(2)In order to achieve the corresponding expected speed of each way point in trajectory tracking,this paper designs a hierarchical longitudinal motion control algorithm,including a low-level acceleration controller based on the vehicle longitudinal inverse model and a speed controller based on receeding horizon linear quadratic tracking control.This type of vehicle speed tracking controller solves the control law through reverse iteration based on dynamic programming,which makes it have advantages in saving computing resources compared to the equivalent unconstrained MPC output tracking controller.In addition,this article also designs a longitudinal and lateral coordination strategy based on fuzzy logic to incorporate the driver’s experience and knowledge,so that the preview time of the MPC lateral motion controller can be adjusted properly according to changes in vehicle speed and road curvature.The adjustment effectively guarantees the effect of trajectory tracking under complicated working conditions.(3)In order to provide the necessary vehicle position and attitude information for the trajectory tracking control algorithm stably and reliably,this paper designs an integrated navigation and positioning algorithm that considers the characteristics of the vehicle.It uses the error state Kalman filter algorithm to loosely couple the data of the global positioning system and the inertial navigation system,and by designing observation equations,it makes full use of vehicle speed information obtained from the CAN bus and the non-holonomic constraints of the ground vehicle to suppress positioning errors.In addition,in the extreme scenario where the satellite signal is interrupted for a long time,the usability of integrated navigation and positioning system is effectively improved by establishing a dead reckoning model based on vehicle dynamics and using extended Kalman filter algorithm for multi-sensor data fusion.(4)Based on a domestic SUV model,a real vehicle test platform is built.The chasis control algorithm is developed for it and corresponding algorithm improvements are made for problems encountered in real vehicle experiments,including compensation for steering system hysteresis and the optimization of the reference way point search algorithm Finally,the effectiveness of the trajectory tracking algorithm for autonomous driving based on GPS/INS integrated in-car navigation technology is verified by real vehicle tests.