Hardware Implementation of a GMDH Controller for Mobile Robot Obstacle Following/Avoidance

Abstract

Proportional, integral, and derivative (PID) controllers have been widely adopted for industrial applications. However, these controllers are not very efficient for non-linear systems. Artificial neural networks (ANN) based on the Group Method of Data Handling (GMDH) have great potential to replace the PID controllers due to their polynomial structure, allowing complex non-linear systems to be controlled. This work presents a hardware architecture of a GMDH network applied to speed control of a mobile robot platform. The proposed GMDH controller was implemented using a 16-bit floating-point arithmetic representation and was mapped on a Zynq7020 device. A hardware-in-the-loop based on the Universal Direct Memory Access methodology was developed to validate the proposed circuits, allowing for performance comparisons between a classical PID and the GMDH controllers for different simulated scenarios.