1993 年 59 巻 559 号 p. 700-707
Feedforward control of sound and vibration using a neural network-based control system is considered. By using a back propagation method, an adaptive algorithm is derived which enables stable adaption of the neural controller for this purpose, while providing the capacity to maintain causality within the control scheme. The neural network controller is shown to be able to compensate for the introduction of harmonics by the control actuator by producing a control signal, derived from a pure tone reference signal, which contains some level of harmonics. Also, the neural network controller is seen to be able to compensate for a distorted reference signal in a manner superior to that of a linear controller. Furthermore, experiments are undertaken which demonstrate the utility of the proposed arrangement, showing it to be the equal of a linear control system in performance for a linear control problem, and superior for a nonlinear control problem.