Vehicle path tracking prediction controller and FPGA heterogeneous acceleration implementation method

Abstract

The invention discloses a vehicle path tracking prediction controller and an FPGA heterogeneous acceleration implementation method, and belongs to the technical field of control. Vehicle system characteristics are learned by using a support vector machine, then a vehicle support vector machine model is established, the control for vehicle path tracking is realized based on model prediction control, and finally an FPGA heterogeneous acceleration method is carried out on the controller. Support vector machine model training is carried out on three selected vehicle state quantities respectively,each model training sample contains six inputs and one output, and the vehicle path tracking prediction controller is designed. A controlled design oriented vehicle model which is simple in structureand contains all information of a vehicle system can be obtained effectively, so that the controller meets all working condition requirements of the system; moreover, the computation speed of the vehicle path tracking prediction controller is improved, and the real-time requirement and low-cost computation and storage requirements of the system are met; and moreover, a control algorithm is convenient to transplant between software and hardware, so that the difficulty of system development is reduced greatly.

Description

technical field [0001] The invention belongs to the technical field of control. Background technique [0002] The autonomous driving system is mainly composed of three modules: the perception system, the planning system and the control system. The path tracking control system is one of the key technologies of the autonomous driving system, and its research is of great significance. The main goal of path tracking control is to control the vehicle to drive on the desired road and complete the driving process under the premise of ensuring safety by controlling the steering system and the driving and braking system of the vehicle according to the planned expected driving path. Therefore, how to design the controller so that the vehicle system can quickly track the desired path, while ensuring the stability of the vehicle during driving, and meeting the real-time, miniaturized, and low-cost requirements of the automotive electronic control system has always been a key issue for t...

AI Technical Summary

Problems solved by technology

[0003] 1. The structure of the vehicle system is complex. In the traditional method of designing controllers, the use of low-degree-of-freedom vehicle models cannot contain all the dynamic characteristics of the vehicle, resulting in the design of the controller cannot meet the requirements of all working conditions of the vehicle; while using high-degree-of-freedom The vehicle model is used to design the controller, which greatly increases the difficulty of controller design and brings a lot of computational burden to the controller;

[0004] 2. For the hardware acceleration implementation of the controller, the traditional FPGA implementation method based on the hardware description language relies on manually generated different register transfer level (RTL) designs. Developers need to have a deep understanding of the underlying architecture of the FPGA device, including timing convergence, state machine Control and accurate clock cycle implementation, and, in order to meet the time requirements of the application program, it is necessary to modify the hardware description language code through multiple debugging during the development process, resulting in difficult development and long development cycle; and the vehicle control system has a low impact on the controller. The requirements for cost calculation, storage and design costs are gradually increasing, and the requirements for system flexibility and communication capabilities are gradually increasing, and a single FPGA chip implementation solution is gradually unable to meet the requirements

Images