Analysing dynamics of two-wheel mobile robot for real-time development of chatbot

Abstract

This study demonstrates effectiveness of artificial neural networks (ANNs) for controlling highly nonlinear and multi-variable systems, i.e., two-wheel mobile robot (TWMR). The study considers real-time control of two-wheel mobile robot (TWMR) incorporating PID and ANFIS-based switching controller. The study aims at controlling these highly nonlinear systems on four different granular surfaces ranging from diameter 2 mm to 4 mm which makes the task more challenging. The performance of the controllers has been monitored considering settling time for chassis angle and wheel displacement. The results clearly show better performance of ANFIS-based switching controller compared to conventional PID controller. The dynamic model of TWMR has been further adopted to develop a real-time model of a movable chatbot as highlighted in the study.