A Novel Analog CMOS Cellular Neural Network for Biologically-Inspired Walking Robot

Abstract

We propose a novel analog CMOS circuit that implements a class of cellular neural networks (CNNs) for biologically-inspired walking robots. Recently, a class of autonomous CNNs, so-called a reaction-diffusion (RD) CNN, has applied to locomotion control in robotics. We have introduced a novel RD-CNN, and implemented it as an analog CMOS circuit by using multiple-input floating-gate (MIFG) MOS FETs. As a result, the circuit can operate in voltage-mode. From the results on computer simulations, we have shown that the circuit has capability to generate stable rhythmic patterns for locomotion control in a quadruped walking robot.