Resonant Beam Enabled Relative Localization for UAV Swarm

The relative localization (RL) of unmanned aerial vehicle (UAV) swarms has gained considerable traction within the realm of UAV formation control. However, most existing RL methods face the challenge of balancing accuracy and computational power in unknown environments. The Resonant Beam (RB) system, which features energy-concentrated and selfaligned transmission can be used for positioning with centimeterlevel accuracy. In this study, we improved the RB system through a double complementary metal-oxide-semiconductors (CMOS) design and proposed a corresponding calculation method to simultaneously obtain the relative position and relative attitude angle, making it adapt to RL problems of swarms. Building upon these advancements, we design a UAV system to solve the three-dimensional RL problems. For large-scale swarms and indirect RL situations, we propose an RB-featured direct-indirect link optimization algorithm considering dynamic correction to improve positioning accuracy and meet real-time requirements. The simulation results show that our root mean square error (RMSE) is within 4cm when the UAV reaches a maximum distance of 30m and is not adjacent to the baseline UAV. This represents a notable 24.2% improvement over the shortest path algorithm. Compared with using LiDAR only for RL, our method offers higher accuracy and achieves millisecond-level speed without requiring GPU computing power.