视觉失效条件下车道保持辅助控制策略研究

Research on Control Strategy for Lane Keeping Assistance under Visual Failure

为保证车道保持辅助系统在视觉失效情况下能够顺利移交车辆控制权,利用目标车方位信息间接估算道路曲率,在纯追踪(Pure Pursuit)算法的基础上,参考车速和道路曲率,利用模糊控制器进行转向修正,并提出了车道保持指数用以评价车道保持性能。构建了硬件在环(HIL)仿真测试平台,利用快速控制原型和某视觉感知系统分别对3种控制策略进行了4种测试场景下的硬件在环测试。测试结果表明,使用改进型Pure Pursuit策略的自车能够有效避免驶出车道,4种测试场景下平均车道保持指数为65.1%,较Pure Pursuit策略和转向盘保持策略分别提高73.9%和135.2%。

In order to ensure that the lane keeping assistance system can transfer the control of the vehicle smoothly in case of visual failure,road curvature information is obtained indirectly with the azimuth and distance information of the target vehicle.Based on the Pure Pursuit algorithm,steering correction is realized via fuzzy controller with reference to vehicle speed and road curvature and the lane keeping index is proposed to evaluate the lane keeping performance.A Hardware-In-the-Loop(HIL)simulation test platform is constructed,and the HIL tests of the 3 control strategies are carried out through the rapid control prototype and a visual perception system in 4 test scenarios.The test results show that the subject vehicle with the improved Pure Pursuit strategy can avoid the rushing out of lane effectively,and the average lane keeping index in four test scenarios is 65.1%,enhanced by 73.9%and 135.2%respectively compared with Pure Pursuit strategy and steering wheel holding strategy respectively.