The study of locomotion by finite state models

Abstract

A methodology to derive finite state models of legged locomotion is outlined. Background data for model derivation are joint angle functions and gait diagrams. The method is used to describe the walking of the cat in terms of an abstract automaton. The main features of finite state descriptions of legged locomotion are described. Such models are presenting locomotion invariants of a species in explicit form. It is emphasized that finite state models provide insight into structural features of motor control organization such as decomposition into subsystems, interaction between centralized and decentralized control, and the role of control levels. The finite state model of locomotion can be helpful in suggesting experiments pertinent to the study of motor control and interpretation of experimental results.