Four-wheel Steering Control Method Of Vehicle Based On Preview Driver Model

Facing the challenge of resource shortages environmental degradation all of theworld, vehicles energy-saving, environmental protection have become an importantresearch direction worldwide. In order to solve these problems, the use of diverse sources,mobile source of zero discharge of advanced electric technology entered people’s field ofvision. The vehicle safety is focused; vehicle handling stability is one of the mostimportant influencing factors.Four-wheel steering technology, machine vision and a variety of new control methodshave been applied to improve the handling stability and driving safety of vehicle. Thispaper combined machine vision and four-wheel steering to achieve the allocation ofvehicle front and rear steering, simulating and analysising the validity of this method.Firstly, the road information is obtained. In this paper image sensor is used to get laneimage. The vehicle state information is obtained using a sensor on the vehicle.Secondly, the neural network driver model is built. Based on the manipulation ofdynamics establish preview-follow the theory of driver model, select the appropriate roadsimulation and then get the required sample training network. The initial network weightthreshold is got by genetic algorithm for optimization of neural network. The neuralnetwork driver model is established, and validated reasonable and effective of the model.Then, four wheel steering controller is designed. It sideslips the angle minimum asthe objective when designing the feed forward controller for assuring that the vehicle hascertain insufficiency steering; designing of the feedback, applying optimal control theoryto minimize the four wheel steering model output and the ideal reference model outputerror.Finally, the different steering mode of the vehicle model is established. They arecarried on the contrast analysis and evaluated of the respective control stability.