Mobile Robot Control Using BP Based Adaptive Critics
The paper presents results of experimental analysis of backpropagation and adaptive critic based control architecture. A number of experiments were performed with the application of a given method for mobile robot navigation. The number of variations of actor and critic neural networks were tested, and results show the ability of algorithm to deal with as many as six layers of units in networks. Observed ability of algorithm to deal with very big learning rates and simple experimental method for stabilization of algorithm are also discussed in the paper.
KeywordsMobile Robot Actor Network Mobile Robot Navigation Adaptive Critic Complex Neural Network
Unable to display preview. Download preview PDF.
- 1.Yves Chauvin and David E. Rumelhart, editors. Backpropagation: Theory, Architectures, and Applications. Lawrence Erlbaum Associates, 1995.Google Scholar
- 2.W. Thomas Miller III, Richard S. Sutton, and Paul J. Werbos, editors. Neural Networks for Control. MIT Press, 1996.Google Scholar
- 3.Omid Omidvar and David L. Elliot. Neural Systems for Control. Academic Press, 1997.Google Scholar
- 4.Danil Prokhorov and Don Wunsch. Adaptive critic designs. IEEE Transactions on Neural Networks,1997. http: //www.acil.ttu. edu/.Google Scholar
- 5.Richard S. Sutton and Andrew G. Barto. Reinforcement Learning (An Introduction). MIT Press, 1998.Google Scholar
- 6.Paul J. Werbos. Stable adaptive control using new critic designs. Technical report, National Science Foundation, Room 675, Arlington, Virginia, USA, 1998.Google Scholar
- 7.P.J. Werbos. The Roots of Backpropagation. Wiley, 1994.Google Scholar
- 8.David A. White and Donald A. Sofge, editors. Handbook of Intelligent Control. Van Nostran Reinhold, 1992.Google Scholar