Analogue neuronal network in a biomorphic machine: Modelling and simulation in ADAMs

Abstract

This paper presents a study of the adaptive navigation behaviour in a five-degree-of-freedom biomorphic machine driven by an analogue neuronal network. The analogue neuron is introduced and its mathematical model derived. The biomorph, Quadrupedal Uncontrolled Locomotion (QUL 1.4), is fabricated for study. The analogue neuron ring controller and mechanics are described. A computer model is created in Automatic Dynamic Analysis of Mechanical Systems (ADAMs) software and successful forward gaits in simulations match the observed results. Obstacle climbing over a cylindrical object is simulated and results are shown. The final part of the paper deals with three levels of coupling within QUL 1.4. A distance modulation approach is introduced as a simple method of identifying lever points in biomorphic machines. © 2010 The Institute of Measurement and Control.