Study on Integrated Chassis Control through Neural Network Nonlinear Decoupling for Light Vehicle
|Keywords||Integrated Chassis Control Nonlinear Relative Gain Array DecouplingControl Neural Network|
To improve the performance of the vehicle, more and more electronic controlsystem was used for the chassis control. However, when these systems were designed,the mutual influence between each other was rarely considered. Adding together thesesystems simply, not only lead to the vehicle structure is more complex, increased costand reduce the vehicle performance, but also produce interference and couplinginevitably. Therefore, it is necessary to research the decoupling of chassis controlsubsystem to decrease the coupling effect. In addition, the vehicle system is acomplex nonlinear system, When the vehicle has a larger lateral acceleration, the2-DOF linear vehicle model which is widespread adoption is no longer meet therequirements. But this work condition always leads the vehicle into a dangeroussituation. Therefore, a new model should be established which can describe theresponse of this vehicle nonlinear traveling state.The objective of this dissertation was to investigate the integrated chassis controlbased on the neural network nonlinear decoupling method. Research is mainlyfocused on nonlinear decoupling integrated chassis control between the active steeringand active braking system. The main contents of this dissertation includes:1、The2-DOF nonlinear vehicle modelNonlinear simplified model is a key point in this article, and it is also animportant factor for integrated control. Considering nonlinear characteristics of thetire, a nonlinear simplified tire’s model is established. Based on this model and theprinciple of vehicle system dynamics, the2-DOF nonlinear model is built. By usingMSC.CarSim software, the accuracy of this2-DOF nonlinear model was verified.Finally this nonlinear model is used to obtain the reference model for the controlsystem. 2、The integrated chassis controller designThe controller includes the reference model, decoupling control algorithm, activeyaw moment allocation servo loop and vehicle state observer. Firstly, the coupledcharacteristic of the chassis system is analyzed. Then the ‘Interactor’ algorithm isused to analyze the reversibility of vehicle dynamics system. Using theBristol-Shiskey method and the neural network’s ability of approximate arbitrarynonlinear systems designed the inverse model of vehicle system as pre compensator torealize the decoupling control. The inverse system connect with the vehicle MIMOsystem in series can acquire two equivalent SISO systems. The performance ofdecoupling is proved by an instance, and the results showed that the pre compensatorcan eliminate the interference between active steering and active braking system.Then using PID method design the closed loop controller for the equivalent systems toimprove the characteristics of the dynamic response. Finally, designed the active yawmoment distribution strategy and established the vehicle state observer.3、Offline simulation research of the integrated chassis control systemBuild integrated chassis control offline simulation platform byMATLAB/Simulink and MSC.CarSim. The CarSim was used to establish the vehicledynamics model, tire model and so on, and the controller model was designed byMATLAB/Simulink. Under typical conditions, the integrated chassis controller’sperformance was analyzed using the integrated chassis control offline simulationplatform. The results showed that the controller can eliminate the coupling betweenactive steering and active braking system effectively and elevate the vehicle steeringstability.4、Integrated chassis control hardware in the loop experimental researchThrough analyzing the functional requirements of the test bench, the structurescheme of the hardware in the loop(HIL) experimental platform was determined andCompleted the type selection and installation of actuators and sensors. The integrated chassis control HIL experimental platform based on dSPACE real time simulationplatform was designed. The HIL experiments under typical conditions wereperformed on this platform. These results and the offline simulation results wereconsistent, further verified the chassis integrated controller has excellent controleffects, can elevate the vehicle handling stability effectively.