Design of Nonlinear Controllers for Active Vehicle Suspensions Using Parameter-Varying Control Synthesis

Nonlinear suspension controllers have the potential to achieve superior performance compared to their linear counterparts. A nonlinear controller can focus on maximizing passenger comfort when the suspension deflection is small compared to its structural limit. As the deflection limit is approached, the controller can shift focus to prevent the suspension deflection from exceeding this limit. This results in superior ride quality over the range of road surfaces, as well as reduced wear of suspension components. This paper presents a novel approach to the design of such nonlinear controllers, based on linear parameter-varying control techniques. Parameter-dependent weighting functions are used to design active suspensions that stiffen as the suspension limits are reached. The controllers use only suspension deflection as a feedback signal. The proposed framework easily extends to the more general case where all the three main performance metrics, i.e., passenger comfort, suspension travel and road holding are considered, and to the design of road adaptive suspensions.     

Design of Nonlinear Controllers for Active Vehicle Suspensions Using Parameter-Varying Control Synthesis

Nonlinear suspension controllers have the potential to achieve superior performance compared to their linear counterparts. A nonlinear controller can focus on maximizing passenger comfort when the suspension deflection is small compared to its structural limit. As the deflection limit is approached, the controller can shift focus to prevent the suspension deflection from exceeding this limit. This results in superior ride quality over the range of road surfaces, as well as reduced wear of suspension components. This paper presents a novel approach to the design of such nonlinear controllers, based on linear parameter-varying control techniques. Parameter-dependent weighting functions are used to design active suspensions that stiffen as the suspension limits are reached. The controllers use only suspension deflection as a feedback signal. The proposed framework easily extends to the more general case where all the three main performance metrics, i.e., passenger comfort, suspension travel and road holding are considered, and to the design of road adaptive suspensions.     

汽车非线性半主动悬架的模糊神经网络控制

考虑磁流变减振器阻尼力和悬架弹性元件非线性特性，建立车辆6自由度的半主动悬架非线性动力学模型。提出了一种基于模糊神经网络系统结构的模型参考自适应控制方法来研究汽车半主动悬架的非线性控制问题，并考虑半车模型前后悬架的输入时滞，对其进行了仿真研究。研究结果表明：运用模糊神经网络非线性控制方法能够使人体和车身垂直加速度、俯仰角加速度都得到很大的衰减，证实这种模糊神经网络控制方法可大大减少路面对车身的振动冲击，提高汽车行驶平顺性。     

A Pitch-Plane Model of a Vehicle with Controlled Suspension

A vehicle model incorporating front and rear wheel suspensions and seat suspension is presented. The suspension control includes algorithms to provide both dynamic and steady state (levelling) control. Vehicle response to (a) vertical inputs due to ground disturbances at the wheels and (b) longitudinal inputs due to the inertial forces during braking and accelerating, are investigated. It is shown that the static (self-levelling) control causes a slight deterioration in dynamic performance. The active ride control produces improvements of ride comfort under dynamic conditions compared to an equivalent passively suspended vehicle. In steady state the proposed control eliminates the error heave of the body caused by tilting of the vehicle with active suspension.     

6自由度半车悬架解耦及其分层振动控制的研究

通过对6自由度半车悬架簧载质量的受力分析,推导出其前后1/4悬架间的定量耦合关系,并以其为基础构建分层振动控制算法.中央控制层以悬架质心处的垂向加速度和俯仰角加速度为控制目标,前后两个1/4悬架构成的两个底层分别采用H_∞和LQR控制策略,并接受中央控制层的协调指令.利用MATLAB的仿真表明,与传统控制相比,分层控制由于前后两个1/4悬架的控制量可以并行解算,计算时间大幅缩短,因而可针对路面激励实施详尽的控制,达到了改善车辆行驶平顺性的目的.     

汽车磁流变非线性悬架模糊控制

建立了整车悬架系统的三维模型,根据试验数据得出了前后悬架弹簧的非线性特性曲线。前后悬架减振器均采用磁流变减振器,采用Bouc-Wen参数化模型为其阻尼力模型。采用模糊控制算法为整车半主动控制算法,采用ADAMS和Matlab联合对整车平顺性进行仿真。结果表明,采用模糊控制算法控制磁流变非线性悬架可提高整车的平顺性。     

基于正交试验的虚拟样车平顺性分析及参数选择

利用多体力学理论,在机械系统动力学分析软件ADAMS中建立了某车的整车多体力学模型。将虚拟样车在三维空间道路上进行平顺性仿真试验,通过正交试验方法研究了前、后悬架弹簧刚度及相关主要衬套等性能参数对汽车行驶平顺性的影响,并根据正交试验的结果优选了相应的设计参数。     

基于ADAMS的空气悬架客车平顺性仿真与试验

以多体系统动力学理论为基础,应用机械系统仿真分析软件ADAMS,创建空气悬架客车前悬架、后悬架的多体系统动力学模型,包括转向系、发动机、车身、前后轮胎等在内的整车虚拟样机模型。并通过编制路面谱文件对虚拟模型进行平顺性仿真和悬挂系统固有频率仿真试验,结果显示该车的平顺性能比较理想。将仿真结果与样车道路试验结果进行对比,发现二者比较吻合,从而验证了所创建的虚拟样机模型的可靠性。研究结果表明虚拟试验可以有效地分析汽车的平顺性。     

Frequency-domain method for evaluating the ride comfort of a motorcycle

In many European towns, the demand for fast and efficient mobility is frequently satisfied by means of two-wheeled vehicles. The improvement of comfort of two-wheeled vehicles used by tired and busy workers can increase safety in ground transport. Nowadays, multibody codes make it possible to predict the ride comfort of two-wheeled vehicles by means of time-domain or frequency-domain simulations. Comfort indices can be developed by post-processing the results of numerical simulations. This task is difficult, because the indices should depend on vehicle characteristics and should be independent of road quality and vehicle speed. Poor quality roads may generate nonlinear effects. Speed influences the trim of the vehicle and the wheelbase filtering, which takes place because the same road unevenness excites the front and rear wheel with a time delay which depends on the vehicle’s speed.

In this paper, the comfort of two-wheeled vehicles is studied by means of a frequency-domain approach. The wheelbase filtering is averaged considering typical missions of the vehicle. The missions are journeys with a forward speed that assumes different values according to a probability density function. Indices of comfort are calculated taking into account the human sensitivity. The examples show that the proposed comfort indices depend on suspensions’ characteristics and, hence, are useful design tools. Finally, some time-domain calculations are carried out to give emphasis to nonlinear effects and to show the limits of the frequency-domain analysis.     

Optimal Preview Control of Rear Suspension Using Nonlinear Neural Networks

The performance of neural networks to be used for identification and optimal control of nonlinear vehicle suspensions is analyzed. It is shown that neuro-vehicle models can be efficiently trained to identify the dynamical characteristics of actual vehicle suspensions. After trained, this neuro-vehicle is used to train both front and rear suspension neuro-controllers under a nonlinear rear preview control scheme. To do that, a neuro-observer is trained to identify the inverse dynamics of the front suspension so that front road disturbances can be identified and used to improve the response of the rear suspension. The performance of the vehicle with neuro-control and with LQ control are compared.     

A Two DOF Model for Jerk Optimal Vehicle Suspensions

Researchers have proposed various active suspension concepts to optimize the tradeoff between ride and handling in passenger vehicles. A few investigators suggested inclusion of the passenger jerk, the derivative of the passenger acceleration, as a measure of ride quality in the performance index. Minimization of a performance index then optimizes both the acceleration and jerk as well as other outputs representing handling quality and design constraints. This approach is called jerk optimal control.

This paper compares two different vehicle models of increasing complexity (the one and two DOF quarter car) using jerk optimal control. Different aspects of suspension performance are investigated, including the structure of the system transfer functions, the structure of the force control laws, and the tradeoffs between the various root mean square (rms) outputs defining system ride and handling performance. Tables compare the numerical results of the two models, allowing predictions of actual vehicle performance.

The results of the two models show the same basic trend for the tradeoff between ride and handling quality: at a constant level of rms passenger acceleration the rms passenger jerk can be reduced significantly, but only at a cost of increased rms tire deflections. In physical terms, a softer ride results in degraded handling performance. For a chosen level of ride improvement, the more realistic two DOF quarter car model predicts more severe degradation of handling. The latter nevertheless predicts a substantial increase in vehicle ride quality is possible through a 55% reduction in jerk. It is expected that actual suspensions could also produce significant increases in ride quality through jerk reduction. Jerk optimal suspensions could find use both in higher end passenger vehicles and in transports for vibration sensitive cargo.     

Structured H2 Optimization of Vehicle Suspensions Based on Multi-Wheel Models

Summary Various control techniques, especially LQG optimal control, have been applied to the design of active and semi-active vehicle suspensions over the past several decades. However passive suspensions remain dominant in the automotive marketplace because they are simple, reliable, and inexpensive. The force generated by a passive suspension at a given wheel can depend only on the relative displacement and velocity at that wheel, and the suspension parameters for the left and right wheels are usually required to be equal. Therefore, a passive vehicle suspension can be viewed as a decentralized feedback controller with constraints to guarantee suspension symmetry. In this paper, we cast the optimization of passive vehicle suspensions as structure-constrained LQG/H2 optimal control problems. Correlated road random excitations are taken as the disturbance inputs; ride comfort, road handling, suspension travel, and vehicle-body attitude are included in the cost outputs. We derive a set of necessary conditions for optimality and then develop a gradient-based method to efficiently solve the structure-constrained H2 optimization problem. An eight-DOF four-wheel-vehicle model is studied as an example to illustrate application of the procedure, which is useful for design of both passive suspensions and active suspensions with controller-structure constraints.     

液压主动悬架的非线性自适应控制

以车身垂直加速度和悬架动行程为控制目标，同时引入非线性高通滤波器和非线性低通滤波器，基于逆向递推(Backstepping)技术，并考虑液压系统的非线性特性及其参数不确定性，提出了一种主动悬架的非线性自适应控制方法。仿真结果表明，在不同的激励信号作用下，都取得了较好的控制效果。     

Optimal Preview Control of Rear Suspension Using Nonlinear Neural Networks

SUMMARY

The performance of neural networks to be used for identification and optimal control of nonlinear vehicle suspensions is analyzed. It is shown that neuro-vehicle models can be efficiently trained to identify the dynamical characteristics of actual vehicle suspensions. After trained, this neuro-vehicle is used to train both front and rear suspension neuro-controllers under a nonlinear rear preview control scheme. To do that, a neuro-observer is trained to identify the inverse dynamics of the front suspension so that front road disturbances can be identified and used to improve the response of the rear suspension. The performance of the vehicle with neuro-control and with LQ control are compared.     

Rollover mitigation for a heavy commercial vehicle

A heavy commercial vehicle has a high probability of rollover because it is usually loaded heavily and thus has a high center of gravity. An anti-roll bar is efficient for rollover mitigation, but it can cause poor ride comfort when the roll stiffness is excessively high. Therefore, active roll control (ARC) systems have been developed to optimally control the roll state of a vehicle while maintaining ride comfort. Previously developed ARC systems have some disadvantages, such as cost, complexity, power consumption, and weight. In this study, an ARC-based rear air suspension for a heavy commercial vehicle, which does not require additional power for control, was designed and manufactured. The rollover index-based vehicle rollover mitigation control scheme was used for the ARC system. Multi-body dynamic models of the suspension subsystem and the full vehicle were used to design the rear air suspension and the ARC system. The reference rollover index was tuned through lab tests. Field tests, such as steady state cornering tests and step steer tests, demonstrated that the roll response characteristics in the steady state and transient state were improved.     

Decentralized neuro-fuzzy control for half car with semi-active suspension system

In this paper, a decentralized neuro-fuzzy controller has been created in order to improve the ride comfort and increase the stability for half car suspension system, which used the magneto-rheological damper as a semi-active device. Firstly, relative gain array and relative disturbance gain methods have been used for deriving a relation between inputs, disturbances and outputs to select pairing with minimum interaction to design a decentralize controller. Secondary, decentralized neuro-fuzzy controllers for front and rear chassis are designed to predict the required damping force taking the acceleration of the sprung mass and desired acceleration obtained by using pole-placement method as inputs. To predict the control voltage required for producing the force predicted by the controller, the inverse neuro-fuzzy model of MR damper has been designed. Simulation by using MATLAB programs has been created. The results show that the ride comforts and vehicle stability have been improved in comparison with the passive system.     

基于虚拟样机技术的扭杆悬架汽车平顺性仿真

运用面向整车系统的数字化虚拟样机技术和大型机械系统动力学仿真软件ADAMS建立了包括前后悬架、轮胎、转向系统及整车的多体动力学模型。用Visual Basic编制路面生成文件，生成不同等级的随机路面。对整车进行了随机输入平顺性仿真分析，把仿真数据输入编制的平顺性仿真程序中，发现结果满足国家规定的汽车平顺性评价标准。     

基于串联型模糊神经网络的汽车半主动悬架的研究

本文建立了五自由度汽车半主动悬架系统模型，提出一种用于汽车悬 半主动振动控制系统的模糊神经网络方法，对半主动悬架 计算机仿真和结果分析，并通过与被动悬架相比较，证明半主动悬架系统在减少振动，提高汽车平一方面要优于被动悬架。     

基于LQR控制的现代客车自适应空气悬架

长期在不良工况的道路上驾驶会降低驾驶员的乘坐舒适性。随着人们对乘坐舒适性需求不断提升,空气弹簧的优势尤为明显。文章提出了一种基于LQR控制策略的自适应空气悬架系统的创新设计方案,提出的LQR控制器采用粒子群算法进行优化。以客车空气悬架为研究对象,采用MATLAB软件对空气悬架系统的被动和自适应动力学模型进行了设计和仿真。仿真结果表明,自适应空气悬架系统在保证车辆稳定性的同时,降低了车辆在随机道路上的最大位移幅值,从而提高了车辆的平顺性。     

基于多目标粒子群优化算法的某轻型商用车操纵稳定性优化研究

针对某轻型商用车稳态回转时侧倾度偏大的问题对其悬架进行优化改进。基于ADAMS/car搭建整车多体动力学模型,通过前悬架反向平行轮跳试验、后悬架理论计算验证了悬架仿真模型的准确性。进行整车稳态回转工况和转向盘中间位置转向工况仿真分析,结果表明,车身侧倾度偏高。为实现操纵稳定性优化分析的流程自动化,提出了基于modeFRONTIER的联合仿真方法。以悬架设计参数为优化变量,以汽车的侧倾度与横摆角速度响应滞后时间为优化目标,采用拉丁超立方试验设计方法拟合得到混合代理模型,并结合多目标粒子群优化算法对悬架系统进行多目标优化,获得了悬架系统优化方案。优化结果显示,在不影响平顺性的前提下,汽车车身侧倾度降低了13.93%,横摆角速度响应滞后时间降低了2.75%,整车操纵稳定性得到了提升。