论文信息 - A hybrid learning architecture based on neural networks for adaptive control of a walking machine

A hybrid learning architecture based on neural networks for adaptive control of a walking machine

Online learning of complex control behaviour of autonomous mobile robots is one of the current research topics. In this article a hybrid learning architecture based on self-organizing neural networks for online adaptivity is presented. The hybrid concept integrates different learning methods and task-oriented representations as well as available domain knowledge. The proposed concept is used for reinforcement learning of control strategies on different control levels on a walking machine.

Winfried Ilg | Karsten Berns | Thomas Mühlfriedel

[1] Jean-Arcady Meyer,et al. Coordination of Leg Movement in Walking Animals , 1991 .

[2] Holk Cruse. Coordination of leg movement in walking animals , 1991 .

[3] José del R. Millán,et al. Learning efficient reactive behavioral sequences from basic reflexes in a goal-directed autonomous robot , 1994 .

[4] Vijaykumar Gullapalli,et al. Learning Control Under Extreme Uncertainty , 1992, NIPS.

[5] Sebastian Thrun,et al. Lifelong robot learning , 1993, Robotics Auton. Syst..

[6] Long-Ji Lin,et al. Reinforcement learning for robots using neural networks , 1992 .

[7] Richard S. Sutton,et al. Neuronlike adaptive elements that can solve difficult learning control problems , 1983, IEEE Transactions on Systems, Man, and Cybernetics.

[8] Roderic A. Grupen,et al. Robust Reinforcement Learning in Motion Planning , 1993, NIPS.

[9] Karsten Berns,et al. A learning architecture based on reinforcement learning for adaptive control of the walking machine LAURON , 1995, Robotics Auton. Syst..