Distributed Inertial–Visual Fusion System with Real-time Evaluation for Multi-target Dynamic Positioning in Team Sports

Addressing the challenges of high-speed motion, occlusion, and complex interactions in multi-target positioning for team sports, this study proposes a distributed inertial–visual fusion measurement and control system with real-time performance evaluation. An adaptive weight allocation framework dynamically adjusts IMU–vision contributions based on motion states and environmental conditions, enabling high-precision fusion of heterogeneous data. A cooperative positioning strategy incorporating formation constraints and relative positional relationships enhances robustness through inter-target information sharing. A comprehensive accuracy–latency–stability evaluation model supports reliable operation via online monitoring and early warning.