论文信息 - Robot-resident neural networks: A step beyond computer simulation

Robot-resident neural networks: A step beyond computer simulation

A simple, commercially available robot platform is presented in this paper with suggested modifications for adding a neural network controller and a better sensory system. The robot, called HexWalker, is a six-legged, servo-driven, battery-powered platform. We found this platform easy to modify with both hardware and software changes. The software changes can be programmed from a standard PC and downloaded into the EEROMS of the robot. This robot provides a tangible model of sensory/motor integration which is useful in both teaching and research.

Frederick J. Bremner | William Collins | Sarah E. Scott

[1] C. Gallistel. The Organization of Action: A New Synthesis , 1982 .

[2] H. J. Chiel,et al. A lesion study of a heterogeneous artificial neural network for hexapod locomotion , 1989, International 1989 Joint Conference on Neural Networks.

[3] Randall D. Beer,et al. A Dynamical Systems Perspective on Agent-Environment Interaction , 1995, Artif. Intell..

[4] K. Pearson,et al. Characteristics of Leg Movements and Patterns of Coordination in Locusts Walking on Rough Terrain , 1984 .

[5] Mary P. Czerwinski,et al. Microsaint for Unix Systems Micro Analysis and Design 4900 Pearl E. Cir., #201E Boulder, CO 80301 303/442–6947 , 1994 .

[6] Frederick J. Bremner,et al. Shades of Braitenberg’s Vehicle 5 , 1995 .

[7] K. Pearson. The control of walking. , 1976, Scientific American.

[8] Randall D. Beer,et al. Evolving Dynamical Neural Networks for Adaptive Behavior , 1992, Adapt. Behav..

[9] V. Braitenberg. Vehicles, Experiments in Synthetic Psychology , 1984 .