Hexapod gait control by a neural network
N. Porcino · 1990
The consideration of neurophysiological data from invertebrate nervous systems and various theories of motor control leads to a robust and biologically plausible architecture for a neural controller for hexapod locomotion. It is an open question as to whether the gait generation is the result of peripheral sensory input or whether it is a function of central control. The controller proposed attempts to reconcile the two arguments by using simple reflexes like those observed in the locust to generate the basic swing-stance cycle and contralaterally and ipselaterally inhibitory central pattern generators to affect coordination of the stepping patterns. When this system is modeled using a network of biologically realistic neurons, it generates walking patterns which respond adaptively to the environment. The patterns generated correspond well to data found in the physiological literature