摘要
将主动转向系统模型与2自由度整车模型相结合,组成被控对象模型。为提高具有非线性特性和不确定性的主动转向系统的自适应性能和鲁棒性能,采用基于线性RBF神经网络的模型逼近自适应控制方法,以横摆率和以横摆率与侧偏角变化率相组合为被控对象参数分别设计C1和C2两种自适应控制器,并对它们进行鲁棒性和稳定性分析。采用1阶参考模型和指令滤波器,对两种自适应控制器和考虑不确定性的H∞鲁棒控制器进行了多种非线性对比仿真。结果表明,自适应控制器比H∞鲁棒控制器有更好的动态和静态跟随性能;而两种自适应控制器中,C2控制器又略胜一筹。
An active steering system model and a 2-DOF vehicle model are combined to form a plant model.For improving the adaptive and robust performances of active steering system with nonlinearities and uncertainty,an adaptive control scheme with model approximation based on linear RBF neural network is employed.With yaw rate and the combination of yaw rate and the changing rate of sideslip angle as control parameters,two adaptive controllers C1 and C2 are designed respectively,on which robustness and stability analyses are performed.1 st order reference model and command filter are adopted to conduct several nonlinear comparative simulations on two adaptive controllers and a H∞ robust controller with uncertainty.The results show that two adaptive controllers have better static and dynamic tracking performances than H∞ robust controller,and in two adaptive controllers,C2 is slightly superior to C1.
出处
《汽车工程》
EI
CSCD
北大核心
2014年第1期107-113,共7页
Automotive Engineering
关键词
主动转向
自适应控制
神经网络
模型逼近
鲁棒控制
active steering
adaptive control
neural network
model approximation
robust control
