Research on cooperative navigation for UAV swarms integrating enhanced YOLOv5 with inertial ranging information

Addressing the insufficient navigation accuracy of UAV swarms in complex environments, this study proposes a multi-source perception method integrating improved YOLOv5 with inertial odometry data. The enhanced YOLOv5 improves robustness in detecting small targets and occluded scenarios, while inertial measurement modeling and data fusion strategies enable high-precision relative position estimation. Building upon this foundation, a distributed collaborative navigation algorithm is designed and validated through simulations and field tests. Results demonstrate superior performance over single-sensor approaches in positioning accuracy, convergence time, and robustness, indicating promising engineering applicability.