基于改进预瞄控制模型的车道保持系统设计
Design of Lane Keeping Assist Systems Based on Improved Preview Control Model
梁永彬 1付广 1林智桂 1何智成 2张家洛 1陈涛1
作者信息
- 1. 上汽通用五菱汽车股份有限公司,柳州,545007
- 2. 湖南大学汽车车身先进设计制造国家重点实验室,长沙,410082
- 折叠
摘要
为解决扭矩控制的车道保持(LKA)系统鲁棒性不高、受车辆制造一致性以及路面激烈干扰影响大的问题,采用神经网络技术、Autofix 算法及预瞄反馈控制理论,利用预期轨迹决策和跟随 PID控制算法设计了一套基于角度控制的车道保持系统.基于 Carsim/Veristand/MATLAB 搭建了硬件在环仿真试验平台,通过虚拟仿真验证了该车道保持系统设计的有效性与准确性.基于 GB/T 39323-2020、CN-CAP—2021、Euro-NCAP—2022 并结合实车调试及用户关注场景对车道保持系统的测试要求,通过仿真及实际场景对比验证得出以下结论:相对于扭矩控制的 LKA 系统,在相同使用场景下,使用角度控制的车道保持系统具有较好的车道保持能力、稳定性、适应性和鲁棒性.
Abstract
To address the issues of low robustness,consistency in vehicle manufacturing,and se-vere road surface interference in LKA system of torque control,a LKA system was designed based on angle control by using neural network technology,Autofix algorithm,and preview feedback control theory,through expected trajectory decision-making and following PID control algorithm.A hard-ware-in-the-loop simulation test platform was built to verify the effectiveness and accuracy of the de-sign of the angle based lane protection systems through virtual simulation based on Carsim/Veristand/MATLAB.Based on GB/T 39323-2020,CN-CAP—2021,Euro-NCAP—2022 and the testing re-quirements of the car retention systems based on real-road commissioning and user concern scenarios,the simulation and comparison with real-road scenarios show that the angle-controlled LKA system has better lane keeping capability,stability,adaptability and robustness than that of the torque-con-trolled LKA system in the same usage scenarios.
关键词
车道保持系统/角度控制/预瞄控制/Autofix算法/比例-积分-微分控制Key words
lane keeping assist(LKA)system/angle control/preview control/Autofix algo-rithm/proportional-integral-derivative(PID)control引用本文复制引用
基金项目
国家自然科学基金(U20A20285)
湖南省自然科学基金(2021JJ10016)
广西壮族自治区科技重大专项(AA22068108)
出版年
2024
国家科技期刊平台
NSTL
万方数据