CAJ | 학술논문

自由漂浮空间机器人动力学方程不能被参数线性化并且存在模型不确定性,使得基于数学模型的控制问题变得十分复杂.本文研究了自由漂浮空间机器人的智能控制方法,提出了一个鲁棒的模糊神经网络控制器.首先利用模糊神经网络控制器来逼近理想控制器,然后利用鲁棒控制器对逼近误差进行估计和抑制.根据Lyapunov函数建立的新的网络学习算法保证了系统的稳定性.最后利用该控制器对平面二连杆空间机器人进行了研究.仿真结果表明该智能控制方案是有效的.
자유표부공간궤기인동역학방정불능피삼수선성화병차존재모형불학정성,사득기우수학모형적공제문제변득십분복잡.본문연구료자유표부공간궤기인적지능공제방법,제출료일개로봉적모호신경망락공제기.수선이용모호신경망락공제기래핍근이상공제기,연후이용로봉공제기대핍근오차진행고계화억제.근거Lyapunov함수건립적신적망락학습산법보증료계통적은정성.최후이용해공제기대평면이련간공간궤기인진행료연구.방진결과표명해지능공제방안시유효적.
Due to the model uncertainties and impossibility of linear parameterization of its dynamics, the control of free-floating space robot based on mathematic model is difficult. In this paper, a robust fuzzy neural network (FNN) controller was proposed to deal with the problem. The FNN controller was designed to approximate an ideal controller, and the effect of approximation error was estimated and attenuated by a robust controller. A novel learning method established using Lyapunov function assured the stability of the system. Simulation results of a two-link planar free-floating space robot verify the validity of the proposed robust FNN controller in the presence of uncertainties.