摘要
针对车辆自主驾驶中的车道线保持问题,基于车辆横向动力学简化线性模型,考虑前向轮自动转动系统的惯性延迟特性,首先将模型按照实际情况分为相对确定与不确定两个子系统,提出一类反演控制与滑模控制相融合的横向位置误差复合控制方法。该方法能保留滑模控制对不确定横向动力学模型的强鲁棒性优点;同时,针对确定的惯性延迟系统,又能继承反演控制层层倒推、环环相扣、精确巧妙设计的思想。由于柔化函数的引入,减弱了传统滑模控制的颤振问题。通过构造Lyapunov函数,保证了系统所有闭环信号均有界。最后,针对轮胎刚度系数不确定的情况,进行了100次随机数字仿真,结果验证了所提方法的良好鲁棒性与快速性。值得说明的是,把复杂系统按照物理实际情况分解为相对确定与不确定两部分的思想,对自动驾驶中车道线保持问题非常有物理意义,能够为未来的车辆试验提供理论参考。
The lane line keeping problem of automatic driving for vehicles is studied based on the simplified linear lateral dynamic model with consideration of the inertial delay characteristic of automatic forward tire steering system. First, the model is divided into a relative certain subsystem and an uncertain subsystem according to the actual situation, a kind of hybrid control method combining inversion control and sliding mode control is proposed to eliminate the lateral position error of vehicle. It makes use of the advantage of sliding mode control that has a strong robustness for uncertain lateral dynamic models of vehicles. Then, aiming at the certain inertial delay system, it inherits the idea of backward induction, interlocking and accurate tactical design of inversion control. A kind of soft function is used to reduce the chattering problem of traditional sliding mode control, and a Lyapunov function is constructed to guarantee all signals of close loop system are bounded. At last, 100 times of random numerical simulation are conducted for the situation of uncertain tire coefficients, and simulation result testified the rapidity and robustness of the proposed method. What is worthy of pointing out is that it is physical meaningful to divide the complex system into a uncertain part and a certain part according to the actual physical situation, which can provide a theoretical reference for the future experiment of automatic driving.
出处
《公路交通科技》
CAS
CSCD
北大核心
2015年第10期153-158,共6页
Journal of Highway and Transportation Research and Development
基金
国家自然科学基金项目(61174031)
山东省自然基金项目(ZR2012FQ010)
关键词
汽车工程
车道保持
滑模控制
反演控制
鲁棒性
automobile engineering
lane keeping
sliding mode control
inversion control
robustness
