비선형궤환을 이용한 매니퓰레이터의 최적경로제어

Abstract

본 연구에서는 매니퓰레이터의 모델방정식을 각각의 관절에 대해 서로 독립된 선행 감결합 모델방정식으로 변환시킨 후, 경로추종오차를 최소화시켜 주는 최적제어 알고리즘을 제안하였다.

이때 비선형시스템의 선형화를 위해 선형화 및 비선형?槪?을 이용하여, 제어변수들에 대해 서로 독립적인 부시스템들로 재구성시켜 주었다. 이러한 부시스템의 상태방정식은 원점에 이중극점이 존재하기 때문에 부시스템이 불안정하게 되므로 이를 개선하기 위하여 상태궤환에 의한 극 재배치방법을 이용하였다. 또한 매니퓨레이터의 수학적 동특성 모델방정식과 전 부시스템 상태방정식에서 유도된 오차방정식으로부터 최적제어 알고리즘을 도출하여 출력오차를 최소화시켜 주었다.

제안된 알고리즘을 PUMA 560 매니퓰레이터의 3관절에 적용하여 시뮬레이션하여 보았을때 만족할 만한 결과를 얻었다.

In the study, an optimal control algorithm is proposed to reduce the path follow-up error of the manipulators.

The algorithm transfes a nonlinear adnamics model equation of a manipulator to several independent decoupled linear model equation.

The dynamics charateristics of the manipulator are derived from the nonlinear model. For modeling the dynamics of the robot manipulator, a nonlinear feedback method is used and a control variable is reconstructed into the independent subsystems. The subsystems are unstable since their state equations of subsystms have double poles at the origin. Therefore, a pole assignment law is adapted to improve stability. The output error is minimized by applying the optimal control algorithm to the error equation, which is derived frm both the mathmetical dynamics model equation of the manipulator and the state equation of the subsystems.

we have simulated the proposed algorithms to a three-joint PUMA 560 model, and observed very encouraging result.

In the study, an optimal control algorithm is proposed to reduce the path follow-up error of the manipulators.

The algorithm transfes a nonlinear adnamics model equation of a manipulator to several independent decoupled linear model equation.

The dynamics charateristics of the manipulator are derived from the nonlinear model. For modeling the dynamics of the robot manipulator, a nonlinear feedback method is used and a control variable is reconstructed into the independent subsystems. The subsystems are unstable since their state equations of subsystms have double poles at the origin. Therefore, a pole assignment law is adapted to improve stability. The output error is minimized by applying the optimal control algorithm to the error equation, which is derived frm both the mathmetical dynamics model equation of the manipulator and the state equation of the subsystems.

we have simulated the proposed algorithms to a three-joint PUMA 560 model, and observed very encouraging result.