Performance Analysis of a Light Weight Ground Robotic Vehicle by Implementing Adaptive Neuro-Fuzzy Inference System (ANFIS)

Abstract

Automated Guided Vehicles (AGVs) are widely used as delivery agents and for material transportation in factories, hospital environment, and other facilities. Conducting performance tests on AGVs has the potential to ratify and improve the efficiency, and reliability of the system. However, published studies on performance analysis focused on classical metrics for such evaluation. In this study, the emphasis is on the performance evaluation of a developed lightweight AGV using the Adaptive Neuro-fuzzy Inference System (ANFIS). The developed line following AGV is flexible, intelligent, and nifty, and can be accessed wirelessly, and controlled by an operator. It was programmed to avoid collision with the help of a proximity sensor attached. The performance test was conducted by drawing black lines on a plain surface for easy navigation of the AGV. A series of experiments was carried out by using realistic test variables like the navigation pattern of AGV, test accuracy, energy efficiency, obstacle avoidance, task accomplishment, and others. Sensitivity analysis was done using the ANFIS surface plot. The total system intelligence (TSI) obtained for the different trials are 76%; 79%; 80%; 81%; 79% and 81 %, for the first, second, third, fourth, fifth, and final trials respectively. The preeminent observable performance was the fourth and sixth trials, obtained at 81 %. The outcome of the investigation reveals that the ANFIS model is an efficient soft computing technique capable of performing TSI tests of AGVs with a high degree of accuracy. The model is also recommended in AGV platooning.