| With the development of automobile industry, more and more demandsto vehicle active safety performance are taken into account. When vehiclestarts or accelerates on the soft road or the road with low coefficientof friction, the wheels will slip excessively because of the low adhesionthat the road surface provides. It causes the longitudinal and lateraladhesions to further reduce. And the vehicle stability,handling andacceleration will greatly reduce. Traction control system (TCS) iscoming to be a popular feature on many vehicles because it can preventthe driven wheels from slip during vehicle accelerates especially whenvehicle starts , accelerates or corners and improve vehicle'sacceleration stability and handling. TCS, which gets the optimum longitudinal traction by regulating thedriving wheels slip ratio, is a new active safety control system thatwas developed in the middle of 80's. It is to regulate the driving wheelsslip ratio around the demanded slip ratio range so that the road adhesioncoefficient is maximized and lateral stability is achieved. Accordingto this theory, with the focus on the active brake control and enginetorque control, this paper has discussed three control algorithms thatis based on optimal control,sliding mode control and fuzzy neural networkcontrol, simulated these three control algorithms with MATLAB/SIMULINK,and compared the performance with the situation without TCS. This paper mainly includes three parts: (1) The vehicle dynamic models are established, which compriseseveral key parts of vehicle, such as engine,transmission,tyre,andwhole vehicle and etc. Engine model contains two parts, which are staticstate model and dynamic model. The former is modeled according to thespeed characteristic experiment, then polynomial fit these experiment 73data (output torque and engine speed); the latter is modeled bysimplifying the engine dynamic process as once inertia system with atransport delay. In the case of transmission, we regard clutch,gear-box,power transfer,transmission shaft and half shaft as the components whosefunction are only transferring the torque. Tyre model adopts the GIMtheory model. And the whole model employs the seven freedom model ,whichincludes longitudinal displacement,lateral displacement,sway angledisplacement and four wheels rotation movement that encircle their ownaxis. These models are used to reflect the subsystem's influence on thedynamic of whole vehicle and observe the running statement of vehicle.AND the corresponding simulation models are established with SIMULINK,which provide an off-line simulation platform for debugging the controlalgorithms. (2) Three algorithms for TCS are proposed, using sliding mode controltheory,optimal control theory and fuzzy neural network control theory.In the sliding mode control, a switching function is established and thestate variables are the difference between the wheel slip ratio and thedemanded slip ratio and the rate of the difference change. And controlalgorithms with constant reaching law and exponential reaching law havebeen established. In the optimal control, the vehicle equation of stateis made and the output state variable is the driving wheel slip ratio.A linear-quadratic state-feedback regulator is founded to calculate theoptimal gain matrix and the state-feedback law is gotten. In the fuzzyneural network control, a multi-layer feed-forward neural network isdesigned to get a fuzzy neural network. The fuzzy control rules are madedue to the experience in practice. The network is trained with these rulesand the neural network controller is founded. (3) The parameters in the algorithms are selected and thesealgorithms are simulated with SIMULINK, using engine torque control onsingle road surface and active brake control on split-μ road surface74respectively. The results of the simulation...
|
PDF Full Text Request