カルマンフィルタを用いた最適トラッキング制御に関する研究

概要

論文の詳細を見る
In the former report, we proposed a simplified new autopilot system with tracking function using a linearized-maneuvering model. And we demonstrated the actual sea tests of the proposed system. In this paper, we propose an extended new tracking system of the former simplified one. In this system, the ship's position is estimated by using Kalman Filter and using the estimated position, the cross track error from the desired track is estimated. The estimated cross track error are input into the ship's motion estimator and using the Kalman Filter technique, the optimal estimates of ship's maneuvering state variables composed of ship's lateral speed, yaw rate, yaw, cross track error and its changing rate are estimated. At last, the estimated maneuvering variables are input into the LQG controller calculation block and the optimal command of the rudder angle is calculated. In the report, we describe the theory of this tracking system, the result of simulation to determine the filter gain and the actual sea tests.
2008-03-25