Research On MEC Multi-sensor Data Fusion Target Tracking Algorithm Based On Adaptive Square Root UKF

In recent years,unmanned autonomous driving technology has been advancing with each passing day and has achieved rapid development.Sensors are an indispensable part of autonomous driving.How to extract effective information from the complex sensor data and fuse the information of multiple different sensors to effectively track the target has become an urgent and challenging task.In view of this,this paper studies the multi-sensor data fusion and target tracking algorithm.Aiming at the shortcomings of the traditional filtering algorithm,a method of using the adaptive square root Unscented Kalman filter(UKF)and Hungary matching is proposed to achieve Multi-sensor data fusion,and using the fused data to track the target.The main research work of this article consists of the following parts:First,this article explains the current development of Multi-Access Edge Computing(MEC)in the field of autonomous driving,as well as the background and significance of multi-sensor fusion algorithms in autonomous driving,and introduces current research of domestic and foreign multi-sensor data fusion.Then the Kalman filter algorithm(Kalman Filter,KF),the extended Kalman filter algorithm(Extended Kalman Filter,EKF),the unscented Kalman filter algorithm(Unscented Kalman Filter,UKF)is introduced in detail,and based on this,the advantages and disadvantages of the above algorithms are analyzed in detail.Then in view of the current development of multi-sensor data fusion and the pros and cons of traditional filtering algorithms,the adaptive square root UKF algorithm is proposed and combined with the Hungarian Matching algorithm to realize the fusion of multi-sensor data.For the motion model of the moving target,this paper adopts a more complicated Constant Turn Rate and Velocity(CTRV)model,which can well describe the motion of the vehicle during smooth driving.The simulation results show that the algorithm proposed in this paper can avoid the non-positive definite noise matrix in the iterative process of common UKF algorithm,and it has good convergence for nonlinear systems.Finally,on the basis of the above-mentioned series of research,this paper uses the Pre Scan simulation platform,through the method of C++ and Matlab mixed programming,to build a simulation test scene for the algorithm of this paper,and realizes the test of the MEC multi-sensor data fusion target tracking based on adaptive UKF Algorithm.