Behavior Based Tracking for Human Following Robot

Robots are becoming part of our everyday life and help to ease the tasks that humans perform with difficulty. Human following while carrying loads is such a challenging task for service robots, as the robot needs to traverse in an unknown dynamic environment with the crowd, obstacles, varying lighting, floor conditions, and different structural components. Therefore, it requires several techniques: uniquely identifying the target and real-time tracking without losing the target often, avoiding static and dynamic obstacles in the path while maintaining a safe distance that does not appear threatening to the person, adapting to different walking speeds, and the target attraction-based acceleration. In this research, a robust vision-based target detection system that detects a custom-designed tri-colored belt with short initialization time, and a computationally less complex Robot Control Architecture was proposed using Fuzzy logic and Subsumption architecture to achieve these goals. The robust performance of the proposed approach is illustrated by the experimental results on a real-world robot which maintains accuracy, hardware cost as well as simplicity of the system and ensures that the robot follows the target person stably, smoothly, and safely.