A new vibration isolation of a hydraulic system using a heterosynaptic neural network

Many hydraulic components have non-linearities to some extent. These non-linearities often cause time delays, thus degrading the performance of hydraulic control systems and making it difficult to model them. This paper proposes a new vibration isolation control algorithm that eliminates the need for sophisticated modeling of a hydraulic system. The algorithm is a hybrid-type control scheme consisting of a linear controller and a heterosynaptic neural network controller. Using this control scheme, simulations and experiments were performed for I and 2 degrees of freedom vibration isolation. The hybrid-type controller can isolate the vibration significantly better than can the linear controller. Also, from the weight values of the heterosynaptic neural network, a second-order equivalent differential model of the hydraulic control system can be obtained, which provides useful information such as the stability and the characteristics of the control system. (C)1997 Academic Press Limited.