| All-terrain vehicle(ATV)have strong maneuverability and passing ability,compared to conventional vehicles,they can easily drive on various complex terrains such as swamps,mountainous,jungles,snow and so on,was widely used in combat,commodity transport,rescue and relief and forest fire prevention.However,most ATV currently use passive suspensions with fixed stiffness and damping parameters,which limits their offroad performance.The semi-active suspension system based on magnetorheological damper has continuous and controllable damping compared with passive suspension,and has low energy consumption and simple structure compared with active suspension,so it received more and more attention.But the current research of MR suspension for ATV is still in its infancy,and many difficulties to overcome in the real vehicle applications.Firstly,the driving conditions of the ATV are complicated,and the noises in sensor signal leads to inaccurate body state estimation,which limits the performance of controller.Secondly,there existed parameter uncertainties and actuator delay in suspension system,traditional robust control algorithm considers the performance of the system in whole frequency domain,which is highly conservative and cannot specifically suppress the vibration of human sensitive frequency band.In order to solve the above two problems,this thesis has carried out the following work:(1)The range of uncertain parameters such as damping,sprung mass and actuator delay in MR suspension system are analyzed,and a seven-degree-of-freedom full-car suspension model is established.A time domain model of full-car road excitation is established based on track coherence of off-road terrains,and the performance index of MR suspension system is presented.(2)The accuracy of estimation of state feedback variable(sprung mass velocity)is affected by sensor noise in partial measurement signals(sprung mass acceleration)in suspension control system,and the existing frequency domain integration method has poor real-time performance in the application of embedded platform(time delay),an incremental PI adaptive Kalman filtering method based on error correction is proposed.The accuracy and real-time performance of the proposed method for velocity signal estimation are verified by experiments.To solve the problem of high frequency noise introduced by relative displacement sensor signal differential in suspension system,Kalman filter is designed to obtain more accurate relative velocity signal.(3)Consider of the actuator delay and the parameter uncertainty of MR suspension system,and considering the main sensitive frequency band of human body to vibration(4Hz-12.5Hz in vertical direction,0.5Hz-2Hz in horizontal direction),a finite frequency robust controller is designed based on the generalized KYP lemma,the solution condition of frequency bounded robust controller is proposed.Finally,the effectiveness of the proposed controller is verified by Matlab simulation.(4)An ATV MR suspension control system has been developed based on a domestic4×4 light ATV.The state estimation algorithm and control algorithm software are programmed by using the model-based design method,and the road test is carried out.The results show that the semi-active suspension using the finite frequency robust controller can attenuate the vibration of the human sensitive band,and the attenuation of the root mean square of the three-axis weighted acceleration at the driver can reach 22.93 %and 17.12% compared with passive suspension and robust control suspension.The research of this thesis has important theoretical significance and practical reference value for the design and development of MR suspension of ATV. |
PDF Full Text Request