1978 年 26 巻 290 号 p. 163-171
The optimal regulator utilizing the future value of disturbance input is investigated, to begin with. It is assumed that disturbances can be measured directly without measurement noise and that future disturbances can be previewed with finite preview length. The optimal control law consists of feedforward of convolution integral with respect to future disturbances in addition to orthodox full state feedback. Then, the theory is applied to an aircraft gust alleviation system of longitudinal mode, in which preview of future disturbance can be realized by locating wind sensors forward of the center of gravity. The performance improvement obtained by using this optimal preview regulator is calculated. Numerical results show its effectiveness, especially in case a controller has non-minimum phase lag.