Piecewise Modeling And Multi-Objective Control Of Magneto-Rheological Semi-Active Suspension

Suspension system is an important part of vehicle chassis,and is a key factor that determines the vehicle ride dynamics.Semi-active suspensions with controllable damping force can effectively improve the comfort performance,and has the advantage of low cost,and has become an important field of suspension technology development.Among different types of semi-active suspensions,magneto-rheological semi-active suspensions have the advantages of fast response and large adjustment range of damping force,but their complicated nonlinear characteristics make controller design difficult.Additionally,traditional suspension control focus on the suppression of sprung mass vibration under vehicle-coupling excitation,which can be characterized by the problem of vertical dynamic control under unknown vehicle-coupling excitation or random excitation.The reason is that road information is difficult to obtain and vehicle maneuver is realized by the driver.Vehicle intellectualization brings more road information and more changes in vehicle manipulation,which brings new research content and opportunities for suspension control.The research work of this dissertation is carried out and put forward under this background,exploring the suspension control problem from two aspects.The first content starts with the nonlinear damping force constraint of magneto-rheological dampers and studies the semi-active suspension control methods with complex road excitation,vehicle load variations,and limited number of sensors.The second content considers the abundant road information and vehicle manipulation methods,and studies the coordination of vehicle longitudinal speed and vertical suspension to improve the comfort performance.The main researches are summarized as follows.The nonlinear characteristics of MR dampers are firstly analyzed and a state-dependent description of the damping force constraint boundaries is obtained.On this basis,a piecewise approximation method of the constraint boundary is proposed to derive a piecewise quarter vehicle model with nonlinear MR damper characteristics.Further,the suspension control problem is transformed to a piecewise H_?control problem.Aiming at the convex optimization methods which depend on state invariant sets and non-zero symmetric constraints,a controller with affine terms is proposed.The control gains which can guarantee H_?performance and system stability is derived based on a global quadratic Lyapunov function(GQLF).Under simulation environments and an experimental test rig,the proposed method is verified and the influence of approximation parameters on control performance is analyzed.Simulation and experimental results show the controller is ef-fective for improving the vehicle comfort performance,and can be applied to embedded vehicle-mounted electronic control unit.Based on the piecewise constrained model of magneto-rheological damping force and the piecewise H_?control,the influence of load variations on suspension control performance and the conservatism caused by the global quadratic Lyapunov function are analyzed.By constructing independent and load-related Lyapunov function for each state partition and designing control gains that varies with the load,a load-adaptive controller design method for magneto-rheological semi-active suspensions is proposed with parameter-dependent quadratic Lyapunov functions.Simulation and experimental results show that the proposed method has better vibration suppression performance and load adaptability when the number of vehicle occupants or cargo load changes.In order to reduce the application cost of automotive electronic control systems and facilitate the application of control methods in actual vehicles,on the basis of the above state feedback control methods,output feedback control methods of magneto-rheological semi-active suspensions based on vehicle-mounted sensors is further explored.The dissertation firstly proposes a supplement and modification for the existing control methods which cannot handle the constraints.By adding control constraint conditions,a design method for the static output feedback control of constrained PWA(piecewise affine)systems is given.Although this method is theoretically feasible,calculation results show that this method has no feasible solution for magneto-rheological semi-active suspensions.The reason of this infeasibility is that the method is based on convex optimization design and a conservative boundary inequality is employed.To this end,this dissertation further explores the design of output feedback controller for magneto-rheological semi-active suspensions from a new view.The basic idea is to analyze the from and structural characteristics of the state and output feedback controllers,and further obtain the output feedback control gains from state feedback controllers through matrix transformation.This method establishes the relationship between the state and output feedback controllers,which is a new research idea and method exploration.The bench and vehicle experiments show that the proposed method has better control performance than passive suspensions and the classic skyhook algorithm,and can be used for real-time control of semi-active suspensions in an embedded electronic control unit.Advances in automotive intelligence and network technologies have brought research opportunities for suspension control to change control methods and control dimensions.The characteristic is that the suspension control is changed from a control with unknown road information or only statistical information to a control with obtained road information,and one control dimension of vertical dynamics is transformed into two control dimensions of vertical and longitudinal dynamics.The changes in control methods and dimensions have caused suspension control change from passive vibration suppression after excitation to active vibration suppression with adjustable excitation,which brings new suspension control problems.In view of the research opportunities brought about by the change of control modes,this dissertation explores the preview semi-active suspension control method by using the information of uneven road ahead.Unlike the widely studied preview control with constant vehicle speed,this dissertation considers the effect of vehicle speed changes on preview information,and discusses how to convert the spatial-domain road information obtained by preview sensors into the time-domain information which can be used by the suspension controller.On this basis,a preview controller that can make full use of the preview data can adapt to the change of preview data length is proposed.For the change of control dimensions brought by automobile intelligence,this dissertation studies the vertical and longitudinal coordination methods to improve the comfort performance.The study is mainly from the perspective of adjusting vehicle speed to reduce excitation and suppressing vertical vibration through suspension control.The idea is to segment the global driving road surface,and optimize the vehicle speed for the area that needs to be adjusted by taking into account the mobility and comfort performance.The proposed method can improve the comfort performance of intelligent vehicles in automatic driving mode,and also provide new ideas and methods for autonomous vehicles to plan and adjust speed according to comfort requirements when driving on uneven roads.