基于VSL-MPC的半主动悬架预瞄控制研究

Research on Semi-active Suspension Preview Control Based on VSL-MPC

车载传感器为智能汽车提供了丰富的环境感知信息,然而,在电控悬架控制算法中,车辆所感知的路面信息尚未能被充分利用,造成车辆动力学控制效果不佳。本文以半主动悬架高性能预瞄控制问题为研究主题,提出了一种变步长模型预测控制(VSL-MPC)算法。该算法根据实时车速和双目相机采集的路面信息来确定预瞄控制步长,使得纳入控制算法中的路面感知信息能够更准确地反映路面特征,有助于半主动悬架在更恰当的时刻对悬架阻尼特性进行调节,能够实现更理想的悬架决策控制。利用双目相机对真实道路开展路面信息采集,引入半主动悬架系统最优性能界限作为性能评价基准,建立4种基于模型预测控制的半主动悬架仿真模型,仿真对比结果表明,驶过连续减速带和井盖冲击等典型城市路面特征时,所提出的VSL-MPC算法控制下的簧载质量垂向加速度与最优性能界限的差距仅为0.72和2.33 dB,相比传统预瞄MPC算法的4.31和4.46 dB、传统无预瞄MPC算法的4.04和4.74 dB具有显著提升,新算法能有效提升半主动悬架的动力学性能。

On board sensors provide rich environment information for intelligent vehicles.However,in the electronically controlled suspension control algorithm,the road information perceived by vehicles has not been fully utilized,resulting in poor vehicle dynamics control effect.In this paper,a variable step length model predictive control(VSL-MPC)algorithm is proposed based on the high-performance preview control of semi-active suspension.The VSL-MPC algorithm determines the step length of preview control by real-time vehicle velocity and the road information collected by the binocular camera,so that the road perception information included in the control algorithm can reflect the road features more accurately,which is helpful for the semi-active suspension to adjust the damping characteristics of the suspension at a more appropriate time to realize a more ideal suspension decisionmaking control.The road profile information is collected by the binocular camera first.Then the optimal performance limit of semi-active suspension system is introduced as the performance evaluation benchmark,and four different semi-active suspension simulation models based on model predictive control are established.The results of simulation show that under the typical urban road conditions such as continuous deceleration belts and manhole cover impact,the performance gap between the VSL-MPC algorithm and the benchmark is only 0.72 and 2.33 dB,which are much smaller than 4.31 and 4.46 dB of traditional preview MPC algorithm,and 4.04 and 4.74 dB of non-preview MPC algorithm,when taking the vertical acceleration of sprung mass as the indicator.The VSL-MPC algorithm can enhance the dynamic performance of semi-active suspension effectively.