Research On Coordinated Control Of Passenger Cars' Ride And Handling Based On Electronically Controlled Air Suspension

Electronically controlled air suspension system is one of the most advanced automotivesuspension systems at present. Through simultaneously controlling the stiffness of airsprings and the damping of shock absorbers, it can excellently coordinate the ride comfortand handling performance of cars, which is hard to achieve for traditional suspensions. So itis widely accepted that as one of the future trends of automotive suspension systems,especially for intelligent vehicles, electronically controlled air suspension system hasbrought a brand-new and feasible way of improving the performance of vehicles.This Ph.D. dissertation is based on the project of National Natural Science Foundationof China named "Research on Control methods and Key Technology for X-By-Wire Vehicle(No.50775096). Based on the summary of domestic and foreign achievements of researcheson electronically controlled air suspension of passenger cars, this dissertation has made anin-depth study into the intelligent control method of electronically controlled air springs andvariable damping shock absorbers, aiming at the optimal coordination between ride andhandling. The control method is verified by hardware-in-loop experiments. The results ofthis dissertation will give theoretical support to the independent development ofelectronically air suspension systems in China. The main research work is summarized asfollows:(1) Study the characteristics of electronically controlled air springs and variabledamping shock absorbers theoretically and experimentally Aiming to understand the characteristics of electronically controlled air suspensionsystem of passenger cars, electronically controlled air spring and variable damping shockabsorber are analyzed theoretically and experimentally. As for the diaphragm type air spring,the relationship between the stiffness of the air spring and some other parameters is revealedby analyzing the internal gas motion state in detail. In terms of variable damping shockabsorber with electromagnetic valve control, the relationship between the parameters ofshock absorber and damping force is also analyzed deeply. On the basis of theoreticalanalysis, experiments on the characteristics of springs and shock absorbers are conductedand functions of them in active suspension control are designed. Experimental results areused to lay theoretical grounds for studies on control algorithm which is proposed infollowing sections.(2) Build the simplified model for the vehicle with electronically controlled suspensionand establish the evaluation system of the overall performance of the vehicleAs the premise of research on active control algorithms, the dynamics model of quartervehicle with2degrees of freedom is established by Newton's law of motion in order todescribe its vertical vibrational characteristics. On the basis of this, the full vehicle modelwith9degrees of freedom is built considering both vibration and handling characteristics.Specifically, the9degrees of freedom are vertical, yaw and lateral motions of the entirevehicle; roll, pitch and vertical motion of the sprung mass; and vertical motion of the fourunsprung masses. Combining the control function of electronically controlled air suspensionwith its impacts on ride and handling performances, the evaluation system is proposed fromthree aspects: vibration comfort, motion comfort and steering stability. So as to validate theeffectiveness and accuracy of the model, outputs from the model is compared with those ofcommercial dynamics package (CarSim). The results demonstrate that the9-degree-of-freedom model can be used to simulate vehicle's ride and handling performance,and to design the control method.(3) Establish the control method of the suspension system based on Genetic ParticleSwarm Linear Quadratic Gaussian Algorithm As the dominant parameter of Linear Quadratic Gaussian algorithm, the weightingcoefficient of performance plays a significant role in the suspension control system. Due tothe weakness of traditional weighting coefficient choice, Linear Quadratic Gaussianalgorithm integrated with Genetic Particle Swarm is put forward in this dissertation.Considering the advantages and disadvantages of both algorithms, selection operation,crossover operation and mutation operation from Genetic Algorithm are introduced intoParticle Swarm Algorithm to optimize weighting coefficients. In the processing ofoptimization, the weighting coefficients are regarded as optimization variables, while thesum of relative variation ratio of index values with Linear Quadratic Gaussian control isconsidered as objective function. The constraint condition is the dynamic stroke ofsuspensions. Optimal weighting coefficients are found by this algorithm and then applied tothe design of the controller. The validation of the algorithm is proved by the simulationresults from a simplified2-degree-of-freedom vehicle model.(4) Establish the coordinated control method of vehicle's ride and handling synthesizingGenetic Particle Swarm Linear Quadratic Gaussian algorithm and fuzzy control algorithmNot only can the electronically controlled suspension damp the vibration caused byuneven road surface, but it can effectively modify the attitude of vehicles during severemaneuvers. Aiming at various maneuvers on uneven road surface, a coordinated controlmethod of ride and handling performances combining Genetic Particle Swarm OptimizationAlgorithm and Fuzzy control is proposed. The control system is divided into three parts:overall performance control module, roll angle control module and coordination controlmodule. On the one hand, the Genetic Particle Swarm Optimization controller of overallperformances takes vehicle states and disturbance as inputs and suspension forces as outputs.Through this algorithm, variables such as vertical acceleration, pitch angular acceleration,roll angular acceleration and dynamic loads of tires are under control, meanwhile, ensuringthe limited stroke of suspension and minimal suspension forces, which sustainably improvesthe ride comfort. On the other hand, Fuzzy controller takes roll angle and angular velocity asits input, and also suspension forces as its output. The controller improves the handling performance by reducing the roll angle and then the roll movement. The distribution moduleof suspension forces takes the outputs of the other two controllers above as input variables. Itidentifies the current state of vehicles by analyzing steering wheel angle and the lateralacceleration, and then determines the overall force of suspensions. With the help of9-degree-of-freedom vehicle model, the comparative research is conducted in open-loopJ-turn driving scenario and single sine driving scenario on a random-rough road. The resultsof simulations illustrates that the control algorithm can effectively coordinate the ride andhandling performances. When vehicle is on straight running, suspension system is controlledto compensate for the vibration derived from random-rough road. The improvement of rideand handling is completed through overall performance module, and roll angle module isshielded. When vehicle is on steering maneuver, roll angle module is triggered to control rollmotion strongly combined with overall performance module, which improve further ride andhandling of automotive. The coordination module is used to accomplish mathematic andlogic calculation of suspension control forces obtained from respectively two module. Theperformance of this control strategy is then examined and assessed in open-loop J-turndriving scenario and single sine driving scenario on a random-rough road by means ofcomputer simulation. Comparisons to a passive suspension system in terms of vehicle sprungmass vertical acceleration, body roll angle and yaw rate is conducted. Simulation resultsindicate that the integrated control strategy proposed in this paper could effectively enhancevehicle ride comfort meanwhile benefit handling quality and driving safety.(5) Design and build the Hardware-in-Loop test rig, measure characteristic parametersof the suspension and verify the control method on the test rigBased on the study above, the rapid prototyping platform with Matlab/xpc Target isbuilt for the study of suspension control. The Hardware-in-Loop test rig consists of thehardware of electronically control air spring, variable damping shock absorber and solenoidvalve driving mechanism. On the basis of this test rig, performances of vehicles with only airspring controlled, only shock absorber controlled and then both of them are study andcompared with those without suspension control. Not only the accuracy of suspension control strategy is verified, but also the control mode of electronically control air suspensionis determined.So in conclusion, several innovative aspects are proposed by this dissertation as below:(1) Comprehensively considering vibration and handling characteristics of the vehicle, afull vehicle model with9degrees of freedom is built up to achieve co-simulation of ride andhandling. In accordance with the established model, evaluation indexes of vehicleperformances are presented from four new points including vibration comfort, vibrationstability, steering stability and energy saving, which lays the foundations for optimal controldesign.(2) For combating the inaccuracy problem of weighting coefficient caused byunreliability of experience and weakness of Genetic Optimization Algorithm by itself, theLinear Quadratic Gaussian algorithm based on Genetic Particle Swarm is proposed. ParticleSwarm algorithm as the basis of the algorithm is combined with selection, crossover, andmutation operation of Genetic algorithm. By applying it, the weighting coefficient ofperformance indexes is optimized and used to the design of Linear Quadratic Gaussiancontroller.(3) In order to fulfill the requirement of different driving conditions, a control strategyfor the vehicle suspension system which combines Linear Quadratic Optimum control lawwith Fuzzy control algorithm is proposed to improve both ride and handling performance.By adopting two performance control modules and one performance coordinated module, theimprovement of ride and handling in open-loop J-turn driving scenario and single sinedriving scenario on a random-rough road is realized.