基于有限时间稳定理论的汽车主动悬架混合减振与容错控制

The requirements of driving safety and ride comfort are increasing due to the rapid development of automobile industry. Therefore, the problem of vehicle active suspension control has been paid much attention in both engineering and academia fields. It is difficult to guarantee the dynamic performance of the system if only the asymptotic stability (AS) is required. In light of this disadvantage, finite-time stability (FTS) theory is introduced into active suspension systems for the design of vibration control and fault-tolerant control scheme which is different from the traditional AS –based one. Four parts are mainly included in this project. Considering the existence of constraints in active suspension, the conservativeness of FTS criteria and robustness of the finite-time controller are firstly investigated. On this basis, according to the respective characteristics of finite-time control and traditional control methods, a hybrid vibration control strategy is proposed for active suspension systems. The finite-time controller and traditional controller are respectively designed for different types of road disturbances. Moreover, the switching control method is employed to make full use of these two controllers’ advantages. The proposed hybrid strategy can not only guarantee the normal operation, but also improve the overall performance of the closed-loop system. Thirdly, taking the sensor/actuator faults into account, the rapid and accurate fault estimation method based on FTS is studied, and then an active fault-tolerant control law is designed. Finally, the simulation and experiment platforms will be constructed. The performance of vibration and fault-tolerant control will be verified on these platforms. This project aims to provide an efficient, reliable and practical control strategy for vehicle active suspension systems, which will promote the further development of the FTS theory and active suspension control technology.

随着汽车工业的快速发展，人们对行驶安全性和乘坐舒适性的要求不断提高，因此汽车主动悬架控制问题引起了工程界和学界的广泛关注。针对传统稳定性理论在处理系统高动态行为方面的不足，将有限时间稳定理论引入主动悬架系统，设计不同于传统渐近稳定意义上的减振与容错控制方案，主要包括四方面工作：首先考虑主动悬架受限特性，研究有限时间稳定条件保守性与控制器鲁棒性；在此基础上，根据有限时间控制与传统控制方法的各自特点，提出一种混合减振控制策略，该策略针对不同路面扰动分别设计控制器，然后采取切换控制方法，确保闭环系统正常运行的前提下充分改善其总体性能；其次考虑传感器/执行器故障，研究快速、准确的有限时间故障估计方法，基于故障估计结果设计主动容错控制律；最后构建仿真与实验平台，验证系统的减振和容错性能。本项目旨在为汽车主动悬架系统提供高效、可靠、实用的控制策略，促进有限时间稳定理论和主动悬架控制技术的进一步发展。

随着汽车工业的快速发展，人们对汽车行驶安全性和乘坐舒适性的要求不断提高，而这两种性能很大程度上由悬架系统决定。在此背景下，研究主动悬架的先进振动控制策略以及对故障具有容忍性的容错控制具有重要理论与实际意义。本项目围绕基于有限时间稳定理论的主动悬架振动控制、基于观测器的主动悬架故障估计与容错控制以及主动悬架智能混合振动控制等方面展开研究。首先，针对冲击型路面扰动持续时间短、强度高的特点，引入能更好刻画系统短时间内动态性能的有限时间稳定性（FTS），设计主动悬架混合FTS/H∞控制策略，抑制车身加速度方均根值及峰值的同时，能保证轮胎相对动载荷、悬架行程限制、执行器最大输出力限制等约束条件；其次，将快速故障估计方法扩展到主动悬架系统，并考虑三种典型执行器故障，应用遗传算法（GA）优化可调参数以改善故障估计与容错性能。在此基础上，考虑主动悬架簧载质量变化，引入径向基函数神经网络（RBFNN）观测器，设计参数依赖型故障估计方法，提高故障估计的精确性、快速性以及对于簧载质量变化的适应性；第三，在传统模糊控制基础上，利用GA同时优化模糊规则和隶属度函数参数，并设计基于模糊小波神经网络（FWNN）的主动悬架振动控制策略，仿真实验证明了该策略在改善车辆乘坐舒适性方面的优越性。另外，为了悬架系统的实际应用，我们针对典型的半主动悬架执行器——磁流变阻尼器，初步开展其基于智能优化技术的数据驱动建模，提高其正逆模型的建模精度。