From passive perception to active control and collaborative planning: A system-level survey on UVA visual moving target tracking

Relying on high maneuverability, unmanned aerial vehicles (UAVs) have become an important frontier direction in the field of moving target tracking. Due to their high adaptability, UAVs equipped with multi-sensor visual tracking technology can execute various tasks in hazardous environments. However, different application requirements and scenarios lead to different tracking focuses. In consideration of the tracking requirements of UAV-based tracking, this paper reviews the current research status of UAV moving target tracking. First, according to the different demands for target tracking across various scenarios, this paper proposes a classification method for target tracking technologies, categorizing them into a three-class evolutionary framework consisting of passive perception, active control, and collaborative planning, based on the deep coupling logic of perception-decision-control. Subsequently, based on different application scenarios, existing UAV moving target tracking algorithms are classified and summarized. Finally, this paper comprehensively reviews multidimensional evaluation metrics and proposes a forward-looking developmental roadmap oriented toward scalable, highly reliable, and swarm-intelligence tracking architectures.