1996 年 1996 巻 3 号 p. 139-144
A pneumatic servo system has been expected to be applied to contact tasks because of the ability of collision force absorption, delicate force control and so on due to air compressibility. However, the high nonlinearity and friction in the pneumatic system make a satisfactory servo-type force control not easy. The purpose of this study is to develop a pneumatic servo-type force control system which can be applied to the system with unknown and variable parameters and environments by using a neural network controller with the capability of learning and adaptability. The controller is installed in parallel with an usual linear controller. Some experiments confirm that the neural network controller is one of effective approaches to realize the advanced force control of a pneumatic servo system.