A new method to develop an optimal four-wheel steering (4WS) vehicle system will be introduced. Based on a linear two-degree of freedom (2DOF) model, a linear quadratic regulator (LQR) with side slip angle and yaw rate as state feedbacks is obtained to follow a desired model, by which the side slip angle is proportional to the lateral acceleration. In addition stability under disturbance, such as wind gusts, is greatly improved due to autonomous front and rear wheel steering effect by the feedback compensation for vehicle handling performance instead of the driver's feedback compensation. Furthermore, since the side slip angle is practically very difficult to measure, a substitution of an optimal output feedback controller with measurable variables (yaw rate or lateral acceleration) is then proposed to achieve the same performance as the full state feedback control. Such an optimal output feedback controller consists of an artificial neural network (NN), which has the properties of learning, memory and adaptability. Simulation results are given to evaluate this design method and some conclusions are made.
A new approach to the design of optimal four-wheel steering system using supervised neural networks
Ein neues Verfahren zur Auslegung eines optimalen Allrad-Lenksystems durch Überwachung mittels eines neuronalen Netzwerks
1993
16 Seiten, 9 Bilder, 17 Quellen
Aufsatz (Konferenz)
Englisch
Automobilbau , Fahrzeuglenkung , Hinterachse , Vorderachse , Projektentwicklung , Auslegung (Dimension) , Prozessoptimierung , technische Entwicklung , Simulationsmodell , Steuerungs- und Regelungssystem , Rückkopplung , Fahrverhalten , Fahrdynamik , lernender Automat , neuronales Netz , Allradlenkung
A new approach to the design of optimal four-wheel steering system using supervised neural networks
British Library Conference Proceedings | 1993
|Nonlinear Design Approach to Four-Wheel-Steering Systems Using Neural Networks
Online Contents | 1995
|Nonlinear Design Approach to Four-Wheel-Steering Systems Using Neural Networks
British Library Conference Proceedings | 1995
|Nonlinear design approach to four-wheel-steering systems using neural networks
Tema Archiv | 1995
|