论文信息 - MULTI-LAYER FEEDBACK-ERROR-LEARNING FOR CONTROL OF FLEXIBLE LINK

MULTI-LAYER FEEDBACK-ERROR-LEARNING FOR CONTROL OF FLEXIBLE LINK

1. ABSTRACT In this paper two neural networks using one control approach using Feedback-Error-Learning are combined. The approach is used for neural control of a one arm flexible link robot in order to acquire the inverse dynamic model of the plant. Such systems are characterized as a non-minimum phase system, which is difficult to be controlled by most techniques and by only one neural network, hence this difficulty is overcome by the combination of two neural networks using the Input Delayed approach discussed here.

Luiz Carlos | Areolino de Almeida Neto | Sandoval Góes | Cairo Lúcio | Nascimento Júnior

[1] Mitsuo Kawato,et al. Feedback-error-learning neural network for trajectory control of a robotic manipulator , 1988, Neural Networks.

[2] Mitsuo Kawato,et al. Neural network control for a closed-loop System using Feedback-error-learning , 1993, Neural Networks.

[3] Panos Liatsis,et al. Artificial Neural Networks in Control and Optimization , 1998 .

[4] C.A.L. Bailer-Jones,et al. An introduction to artificial neural networks , 2001 .