2004 年 70 巻 698 号 p. 2830-2838
We make a semi-active shock absorber on an experimental basis and realize its intelligent control. In the shock absorber, an area of orifice is controllable widely according to the angle of the adjustment shaft which is rotated by a stepping motor. A load cell is only used as a sensor because of practically. We intend to control the high shock absorbing efficiencies of the shock absorber using intelligent control method in spite of rare detum. In the experimental studies, masses and energies of collision things are estimated from resisting force values by means of neural networks. Neuro-like fuzzy control using backpropagation leaning is applied to the shock absorber and high shock absorbing efficiencies are realized for various crushes.