Instantaneous rotational attitude estimation of the UAV airframe based on micro-Doppler LiDAR

This is the authors' abstract. We don't add key points for this paper.

Conventional micro-Doppler extraction methods face difficulty in accurately estimating the instantaneous micro-motion attitude of the UAV airframe when the vehicle performs variable-speed or direction-changing maneuvers. To address this problem, this paper proposes a coherent laser micro-Doppler detection method based on improved particle swarm optimization. A dynamic model of variable-speed airframe rotation is first established. Then, a non-prior parameter estimation framework is used to perform nonlinear fitting of the instantaneous rotational frequency and attitude angles. Simulations compare and analyze the time-frequency characteristics of global and local scattering under different variable-speed conditions. The simulation results show that, compared with conventional methods, the proposed approach reduces the average estimation error of rotational frequency by 82.5% and that of attitude angle by 89.6%. Finally, a 1.2 km UAV detection experiment is conducted using a visible-light-guided micro-Doppler LiDAR. The instantaneous rotational frequency and attitude of a real UAV airframe are successfully acquired, with average relative errors below 9% and 0.6%, respectively. This work provides a theoretical and technical foundation for short-term flight trajectory prediction and behavior perception of maneuvering UAVs.