Research On Intelligent Control Algorithm Of Vehicle Semi-active Suspension

Suspension system is a core part to alleviate the car's external interference.Poor control of the car suspension will cause riders carsickness.More importantly,it can cause safety accidents.Semi-active suspension system can adaptively reduce vibration by changing spring rate coefficient or damping coefficient according to the driving status of the vehicle.It has a wide range of applications and strong practical significance to study its control issues.Magnetorheological damper(MRD)has been widely concerned in the semi-active suspension due to the easy realization of variable damping real-time control and low energy input.But there is a problem of large local error in predicting the control current.Meanwhile,it is difficult to improves both the handling stability and the ride comfort in vehicle control.To cure the above problems,the main research contents are as follows:1)A inverse model of MRD is established by a back-propagation neural network and the inverse model is optimizeused by genetic algorithm.2)A force coordinator with front and rear axis control parameters is proposed.Three forces of the coordinator body in the vertical,pitch and roll directions are properly assigned to the four semi-active suspensions.And the forces are output by three parallel fuzzy PID controllers.By coordinating the damping force of the output of the four-wheel suspension,the vibration of the vehicle is controlled.3)A steering strategy is proposed to discuss the steering characteristics of the vehicle.Therelationship is described among wheel angle,difference between yaw rate calculated and target yaw rate,steering characteristics and the front and rear axle roll distribution parameter.In order to verify the effectiveness of the proposed control algorithm,test analysis and generalization verification were carried out.In the continuous uneven and discrete uneven road conditions,experimental analysis of the 1/4 semi-active suspension and the semi-active suspension was conducted.Then the influence of road grade and vehicle speed on vehicle vibration are discussed.Experimental and objective data show:1)After using the proposed algorithm to optimize the MRD inverse model based on BP neural network,the predictive tracking of the control current at the positive and negative inflection points is more accurate and the error is significantly reduced compared with unoptimized reverse model.Under the excitation of two different road surfaces,the single suspension's vertical acceleration,dynamic deflection and dynamic load are improved.2)Under two different roads,the vehicle system is experimented and analyzed,compared with the semi-active suspension control system before optimization,the RMS values of vertical acceleration and pitch acceleration are significantly reduced.The ride comfort of the car is improved.The roll angle acceleration decreases by 29.64% and 24.87% respectively,which improves the steering stability of the vehicle.The above data shows that the proposed force coordinator can effectively adjust the force distribution.3)The front wheel angle and the difference between the calculated value of the yaw rate and the target yaw rate determines the understeer,oversteer,or neutral.The changing trend of the roll distribution parameters is given,which provides a reference for the coordinated control of the whole vehicle and is beneficial to the neutral steering of the whole vehicle.In summary,the MR damper inverse model and control method proposed in this paper are effective for damping under different road grades and different vehicle speeds.