Research On Active Control Systems For Rollover Prevention Of Heavy Vehicles

Vehicle rollover is an extremely dangerous lateral motion for people drivingon the road. Especially for heavy vehicles, the loss of life and property caused byheavy vehicle rollover accidents is extensive. In fact, the number of rolloveraccidents is keep increasing year by year among all other traffic accidents, limitingthe development of transportation industry.On the basis of relevant research, the active roll control system for improv ingroll stability and avoiding vehicle rollover is mainly analyzed. The main researchcontents in this thesis are as follows:(1) Based on the simple2DOF vehicle handling model, a3DOF model isestablished by adding the roll motion to the2DOF model. The simple3DOFvehicle model directly and accurately describes basic vehicle chara cteristics inlateral motion, yaw motion and roll motion. Meanwhile, by defining the relevantauxiliary variables, the two-axle and three-axle vehicle3DOF models are extendedto develop a generalized3DOF model for a multi-axle vehicle to apply to thecontrol design.(2) By combining the genetic algorithm and L-M optimization algorithm, theparameters which can not be measured easily and directly, such as vehicle inertias,tire cornering stiffness, the location of gravity center, suspension roll stiffness anddamping, are identified based on the simulation data of TruckSim under acombined sinusoidal and step steering input. Then set up other driving conditionsto validate the results of parameter identification. Finally, the comparison ofTruckSim and Simulink simulation results show that the accuracy of vehicle modelis improved by using the estimated parameters. The transient and steady state ofvehicle motion can be better descried by using the identified parameters.(3) By using the traditional control strategy and the value of lateralacceleration for rollover warning, the rollover prevention control system is built byusing LQR algorithm, which based on a reference vehicle model. Firstly, tables offeedback control data are computed off-line with LQR method. Once the lateralacceleration exceeds the set rollover threshold value, then the control sy stem willcalculate the optimum control moment quickly by looking up the table s. And thenadjust the break pressure of wheels to produce the optimum moment to stable the vehicle. Meanwhile, the whole control system is modeled by using TruckSim andSimulink. The co-simulation results under the fishhook and sinusoidal dr ivingconditions indicate that the designed control system has no interference withnormal driving. However, when the roll stability is dropping during the dr iving, thecontrol system can prevent the vehicle rollover effectively.