Vision-Based Multi-UAV Cooperative Target Tracking and Control Technology

With the extensive deployment of unmanned aerial vehicles (UAVs) in various fields, effectively countering "low-altitude, slow, and small"UAVs, especially in security-sensitive zones like civil aviation airports, has emerged as a critical challenge. Traditional electromagnetic interference based countermeasures risk disrupting aviation communications and flight safety, making UAV-based tracking and interception a safer alternative. However, small UAVs used in such missions face constraints in power and payload capacity, limiting the use of high-precision sensors like millimeter-wave radar and LiDAR, which are inadequate for tracking highly dynamic targets. To address this, this study focuses on vision sensor-based multi-UAV cooperative target tracking and motion control strategies for aerial small UAV interception. The primary contributions include: (1) A bearing-only target tracking algorithm using the Extended Kalman Filter and a perception-optimized adaptive control algorithm to resolve depth information limitations in multi-UAV visual tracking. By integrating the error covariance matrix from cooperative Kalman filtering, we formulate a joint estimation error minimization framework for real-time optimal target state estimation. Concurrently, a motion control optimization function dynamically adjusts UAV observation positions and geometric configurations to maximize estimation accuracy. (2) A multi-UAV cooperative perception and control platform incorporating hardware (airframes, sensors, flight controllers, onboard computers) and software (perception, planning, communication modules). Vision sensors are calibrated, and task-specific datasets train target detection models. Extensive experiments validate the proposed algorithms' effectiveness and reliability in real-world scenarios, demonstrating robust performance in complex environments.