Real-time control of walking
A distributed neural network architecture for hexapod robot locomotion
Neural Computation
Behavior-based artificial intelligence
Proceedings of the second international conference on From animals to animats 2 : simulation of adaptive behavior: simulation of adaptive behavior
Artificial neural nets for controlling a 6-legged walking system
Proceedings of the second international conference on From animals to animats 2 : simulation of adaptive behavior: simulation of adaptive behavior
Evolving dynamical neural networks for adaptive behavior
Adaptive Behavior
Walknet—a biologically inspired network to control six-legged walking
Neural Networks - Special issue on neural control and robotics: biology and technology
The Handbook of Brain Theory and Neural Networks
The Handbook of Brain Theory and Neural Networks
Reflex-oscillations in evolved single leg neurocontrollers for walking machines
Natural Computing: an international journal
Robotics and Autonomous Systems
Evolved neural reflex-oscillators for walking machines
IWINAC'05 Proceedings of the First international conference on Mechanisms, Symbols, and Models Underlying Cognition: interplay between natural and artificial computation - Volume Part I
Robotics and Autonomous Systems
| Hi-index | 0.00 |
Classical engineering approaches to controlling a hexapodwalker typically involve a central control instance that implementsan abstract optimal gait pattern and relies on additionaloptimization criteria to generate reference signals forservocontrollers at all the joints. In contrast, the gait of theslow-walking stick insect apparently emerges from an extremelydecentralized architecture with separate step pattern generators foreach leg, a strong dependence on sensory feedback, and multiple, inpart redundant, primarily local interactions among the step patterngenerators. Thus, stepping and step coordination do not reflect anexplicit specification based on a global optimization using arepresentation of the system and its environment; instead they emergefrom a distributed system and from the complex interaction with theenvironment. A similarly decentralized control at the level of singleleg joints also may explain the control of leg dynamics. Simulationsshow that negative feedback for control of body height and walkingdirection combined with positive feedback for generation ofpropulsion produce a simple, extremely decentralized system that canhandle a wide variety of changes in the walking system and itsenvironment. Thus, there is no need for a central controllerimplementing global optimization. Furthermore, physiological resultsindicate that the nervous system uses approximate algorithms toachieve the desired behavioral output rather than an explicit, exactsolution of the problem. Simulations and implementation of thesedesign principles are being used to test their utility forcontrolling six-legged walking machines.