| Popis: |
The side slip angle provides a key state value that defines the yawing and lateral motion of a vehicle. Accordingly, would it be possible to detect or estimate the side slip angle of a real vehicle with higher precision, the vehicle could be properly controlled with a more ideal state feedback control. There are two methods of measuring side slip angles. One is a direct method that measures the longitudinal and lateral velocity respectively, using optical speed sensors. The other is an indirect method that integrates the difference between the lateral acceleration/vehicle speed and the yaw velocity. These methods have been generally used for measurements of vehicle dynamics. Although, the only representative method for estimation would be the application of a linear vehicle model, like an observer, precision and reliability are not sufficient in various conditions of production cars. The objective of this paper is to estimate side slip angles with higher precision using an artificial neural network, without using optical speed sensors or an integral calculation method. The description is devoted to a settlement of basic system performance using simulation and to a verification of test results using actual vehicles. |
