High-Performance of Mobile Robot Behavior Based on Intelligent System

Abstract

This paper presents behavior based for leader follower mobile robots using fuzzy logic and YOLO convolutional neural network (CNN). The designed system consists of two behaviors, formation and go to goal. Using a web camera fixed on the roof, the YOLO network is trained to identify the locations of mobile robots, while a computer vision algorithm is used to calculate the desired distances and angles required to achieve the formation and go to goal behaviors. The mobile robots’ speed and direction are managed by a fuzzy logic controller to accomplish the formation and go to the goal point. The novelty of this paper comes from its integration of YOLO for real-time mobile robot detection through a roof-mounted webcam, computer vision for calculating distances and angles, and fuzzy logic for controlling the speed and heading of the mobile robots. Furthermore, using YOLO and computer vison instead of traditional distance sensors, such as LiDAR or ultrasonic sensors give us better understanding of the environment, since it can identify not only distances and angles but also the type and identity of objects in the workspace area. This can aid in the decision-making process for a mobile robot’s autonomous navigation. The system is implemented without inter communication between the leader and follower mobile robots. The practical results show the validity of the presented approach.