汽车阻尼可调半主动悬架混杂模型预测控制

摘要

To coordinate the contradiction between ride comfort and handling stability of damping continuously variable semi-active suspension,and considering the satisfied nonlinear constraints for damping force,a hybrid model predictive control algorithm for semi-active suspension is proposed.A single wheel model of semi-active suspension is established,and the optimal objective function that comprehensively determines comfort and stability and the satisfied nonlinear constraints for semi-active suspension are put forward.The semi-active suspension hybrid system is described with mixed logic dynamic modeling method.A finite-horizon optimal control problem of semi-active suspension is set up following model predictive control theory.For the convenience of solving,the nonlinear constrained optimization problem can be transformed into a mixed integer quadratic programming one including real and binary vectors,and solved by means of branch and bound algorithm.The simulations on random profile and single sine wave bump pavement show that the proposed hybrid model predictive control algorithm improves the comprehensive performance of damping continuously variable semi-active suspension system effectively compared with passive suspension and the traditional linear quadratic regulator algorithm,and better control performance is realized under different road input excitations and at different vehicle speeds.%为协调汽车阻尼连续可调半主动悬架舒适性和操稳性之间的矛盾,并考虑减振器阻尼力须满足的非线性约束条件,研究了一种适用于半主动悬架的混杂模型预测控制方法,包括建立半主动悬架单轮车辆模型,提出综合描述舒适性和操稳性的悬架优化控制目标函数及半主动悬架须满足的非线性约束条件.利用混合逻辑动态建模方法描述半主动悬架混杂系统模型,基于模型预测控制理论研究半主动悬架混杂系统的有限时域优化控制问题.为便于问题求解,将非线性约束优化问题转化为包含实值变量和二值变量的混合整数二次规划问题,并借助分支定界算法进行求解.在随机路面和正弦波突起路面进行仿真验证后表明,所提混杂模型预测方法明显优于被动悬架和传统线性二次型调节器算法,能有效提高阻尼可调半主动悬架的综合性能,并在不同的路面输入激励和车速下均能取得较好的控制效果.