MECHANICAL ENGINEERING

Mechanical Engineering

Mechanical Engineering-Car Simulation

Part 1: How to Build a simulation of a control system for an unmanned automatically guided vehicle in C#

With the social progress and the rapid development of the transportation industry, people have paid more and more attention to both the active and passive safety of the automobile. Among many kinds of active and passive safety systems, the 4WS is one of the most important methods in improving the low-speed mobility and the high-speed handling stability.

Under the circumstance of the same steering angle of front wheel, the 4WS vehicle has a much smaller steering radius than the front-wheel steering (FWS for short) vehicle at low speed because of the inverse direction of the steering angle of the front and rear wheel. So the 4WS vehicle has a good performance in mobility at low speed. Besides, the steering radius of 4WS vehicle is longer than that of FWS at high speed because of the same direction in which the front and rear wheel steer. And when the 4WS is at high speed the yaw rate and lateral acceleration can be decreased to improve the high-speed driving stability [1].

Nowadays, most of the control methods of 4WS are traditional methods. For example, Professor Hu Lisheng and Li Youde presented the method of designing the 2-DOF robust Controller using the robust control theory [2]. With the development of the artificial neural network, people begin to study the 4WS through the artificial neural network theory. The influential factor of the neural network is that it can approximate any nonlinear connection in any given accuracy. So it can be widely used in identification and controlling fields. Lv qiang presented the control method by RBF network, and has discussed the validity of that method. The characteristic of that control method is to use the nonlinear vehicle model which would be learned in order to design the controller having the ability to compensate the error of nonlinear [3]. But this method is an offline control of the neural network. The data used in training the neural network can't represent the true working conditions of the vehicle. So there must be errors.

According to this phenomenon, the neural network direct inverse control with the widespread used BP neural network model is built in this paper. This control method has a much more succinct structure than other neural network control methods and manages to control the object only through the online learning processes of its inverse models. The control accuracy can hit high levels by the online learning.

4WS DYNAMIC MODEL

Only the mobility at low speed and the handling stability at high speed between FWS and the controlled 4WS are analyzed and compared in this paper. In order to simplify the system model, the basic linear 2-DOF vehicle model [4] is used and the change of the tire cornering properties is not considered. The vehicle movement is considered as plane motion parallels the road, so the displacement in Z axis, the circuition of Y axis and X axis ...