Title: Feedforward neuro-controlled active suspension using frequency and time-mixed shape performance index. Part 1: control logic and performance

Authors: Yong-San Yoon, Hyuk Kim

Addresses: Department of Mechanical Engineering, Korea Advanced Institute of Science and Technology, Yusung-Gu Kusung-Dong 373-1, Taejon 305-701, Korea. ' Department of Mechanical Engineering, Korea Advanced Institute of Science and Technology, Yusung-Gu Kusung-Dong 373-1, Taejon 305-701, Korea

Abstract: The objective of this study is to develop a control law of the active suspension for the safety and ride quality improvement using a frequency and time-mixed shape performance index with speed and road-adaptive preview logic. The incorporation of a mixed shape performance index allows one to emphasize or de-emphasize certain variables over specific bands of frequencies. The criteria recommended by ISO2631 and BS6841 provide the frequency weighting curves for human body vibration in relation to ride comfort. A unified strategy of finite optimal feedforward control is developed both for a haversine bump and general highway. In the course of deriving the feedforward control logic, we used the theorem that approximate control solution exists and becomes arbitrarily close to the optimal control solution for a sufficiently large preview time. Through an eight degrees of freedom linear ride model, we show that preview information and complex type performance index can improve the ride quality of the vehicle.

Keywords: active suspension; performance index; preview control; ride comfort; vehicle safety; feedforward control; optimal control; neural networks; control logic; vehicle design; vehicle suspension.

DOI: 10.1504/IJVD.1996.061985

International Journal of Vehicle Design, 1996 Vol.17 No.2, pp.163 - 181

Published online: 28 May 2014 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article