A Review of Artificial Neural Network-Based Controllers for Mobile Robot Dynamics Control

摘要

Artificial neural networks (ANNs) are adaptive systems employed in many applications such as solving complex nonlinear mathematical functions, identification, data classification, control, and others. The utilization of ANNs in the field of control systems has significantly enhanced the performance of controllers, surpassing that of traditional controllers such as proportional-integral-derivative (PID) controllers. This advancement has transformed these controllers into adaptive control. These adaptive controllers are utilized to control the mobile robot and handle its nonlinearity property and friction uncertainty in its mobility area. Therefore, this paper aims to introduce the comparison of various types of ANNs-based controllers used to practically control the dynamics of the mobile robot. Moreover, multiple experimental tasks are applied to the mobile robot to investigate the performance of each adaptive controller through two performance indices namely; integral absolute error (IAE) and mean absolute error (MAE). The experimental tasks included the set-point change, parameter uncertainty, and measurement error disturbance.