Real-Time Based ANFIS Control of Two Wheel Robotic (TWR) System

Abstract

This article presents design and control of a highly nonlinear and multivariable two-wheeled robotic (TWR) system. A model of TWR system has been designed and controlled using an Arduino UNO microcontroller equipped with a MPU-6050 sensor. The real-time data for chassis angle and chassis angular velocity was obtained through Arduino Uno controller and was used for training of an Adaptive neuro fuzzy inference system (ANFIS) controller. ANFIS controller has been designed using only seven gbell shape membership function which solved the problem of fuzzy rule explosion by reducing the number of fuzzy rules from 256 to 49 compared to ANFIS controller proposed in one of the previous studies of author. Moreover, through proposed technique it was successfully possible to stabilize the chassis angle and chassis angular velocity within 2.0 sec and 2.5 sec respectively within acceptable overshoot and steady state error limits.