Self-driving Marshalling Bus Pose Sensing And Lateral Control

Based the background that the driverless technology for smart cars has not been completely implemented,and it is very urgent to solve the problem of traffic congestion in cities,the vehicle grouping method emerges as the times require as a driving method derived from the platooning of vehicles.The bus grouping adopts semi-autonomous driving technology.The first bus is in manual driving mode,and the following vehicles use automatic driving technology to track the trajectory of the preceding vehicle by detecting the position and attitude information of the preceding vehicle,thereby realizing the grouping of buses.In this paper,with the help of the autonomous driving marshalling bus platform,in the aspect of pose sensing,an FPGA-based pose estimation platform is designed to obtain the relative pose information of the front and rear vehicles;In terms of lateral tracking algorithm,a lateral controller based on fuzzy adaptive algorithm is designed.The experiments show that the following vehicle can automatically track the leading vehicle in the scene of straight road,lane change and turning.Therefore,the research content of this thesis mainly focuses on the research on the pose sensing and lateral control of the autonomous driving marshalling vehicle,and the main content includes the following three aspects:First,an autonomous driving marshalling bus platform is built based on buses.Aiming at the characteristics of the bus-only roads and closed stations of the BRT system,this paper reconstructs two buses and builds the autonomous driving marshalling bus platform of this paper,which mainly includes the modification of the vehicle body and the wire-controlled chassis.The device selection of the underlying control system and the pose estimation system,the circuit design of the core control board and the peripheral interface.Second,for the pose sensing of marshalling vehicles,an FPGA-based pose estimation platform is designed.The pose estimation platform uses monocular vision and FPGA processor to complete video image acquisition,SDRAM buffering,and image preprocessing,and completes the pose extraction of image markers in static and dynamic scenarios through the positioning marks of image markers.The frame rate of image marker detection can reach30 fps,and the positioning and marking of different scenes can be realized by adjusting the threshold.The experimental results prove the feasibility of the pose estimation platform.Third,for the lateral tracking control algorithm,a lateral controller based on fuzzy adaptive PID control algorithm is designed.Aiming at the limitations of traditional PID control and fuzzy control,this paper combines the two control algorithms to realize the lateral tracking of the following vehicle,and uses the MATLAB/SIMULINK platform to build a simulation model and compare the traditional PID control to verify the effectiveness of the algorithm.And with the help of the marshalling vehicle platform,the lateral control algorithm designed in this paper is tested on the actual road.Simulation and real vehicle experiments show that the lateral tracking control algorithm can perform parameter self-tuning under different road conditions and vehicle speeds,so that it has better dynamic response characteristics and robustness in the tracking process.Compared with classical PID control,it shows strong superiority.