フィードバック誤差学習を応用したフレキシブルアームの軌跡制御(機械力学,計測,自動制御)

概要

論文の詳細を見る
In this paper, we propose the controller based on feedback-error-learning and physical properties of mechanical system. We applied the controller to flexible robot arm with two-mass structure. In the control system, at first, the desired trajectory of the tip point of the arm is transform into the desired signal of the driving point of the arm by equation of motion. And then, the driving point is controlled, the trajectory at the tip of the arm is controlled with high precision. And, when the controller has modeling error, the controller learn it with neural networks. So, the controller can control the tip point of the arm without motion sensor. We simulated trajectory control of the tip point of the flexible robot arm. And, we experimented about it. As a result, we obtained the improvement of performance of the arm by the proposed controller.
一般社団法人日本機械学会の論文
2004-07-25