Research on People Following Technology Based on Multiple Sensors

With the continuous expansion of application scenarios for service robots, person-following technology has become a key capability for enabling human-robot interaction and intelligent control.This paper focuses on a multisensor fusion-based person-following system and systematically reviews its core technologies, including target recognition, feature extraction, path planning, and obstacle avoidance.First, the features of typical followable targets are categorized into five aspects: visual appearance, geometric structure, motion trajectory, multi-modal perception, and pose estimation.Representative methods such as deep learning, skeleton extraction, and trajectory prediction are discussed.Next, the mainstream path planning strategies and obstacle avoidance mechanisms are reviewed.Finally, an analysis of existing prototype systems reveals that current person-following technology still faces challenges in occlusion handling, real-time embedded computing, and ensuring user experience and comfort..