Unmanned aerial vehicles (UAVs) have emerged as a versatile platform for a wide range of civil applications, and offer flexibility and convenience in performing various tasks, like mapping, investigation, and patrolling. Accurate navigation information such as velocity and position is required to ensure the flight safety of UAVs. However, the global navigation satellite system (GNSS) is easily denied in urban environments due to the occlusion of tall buildings, which results in navigation information loss. Inertial navigation systems (INSs) are also hard to be relied on for long-scale autonomous navigation as its error accumulates over time. High-precision and independent velocity measurement methods will be beneficial for the navigation of UAVs. The laser Doppler velocimeter (LDV) has been applied to the integrated navigation of land vehicles improving localization accuracy. There is a bottleneck for LDV deployment on UAVs due to the limited working distance of LDV which is typically restricted to only a few meters. However, UAVs often operate at flight heights of dozens of meters, posing a challenge for LDV integration and utilization.

The carrier-to-noise ratio (CNR) of LDV is analyzed concerning coherent Doppler wind lidar since both of them are coherent detection systems. For the same LDV, the CNR can be improved by reducing the size of the probe beam spot on the target. On this basis, we propose a solution to the bottleneck of implementing airborne LDV by an optical transformation system to extend the working distance of LDV. The optical transformation system comprises a concave lens and a convex lens. By passing the Gaussian probe beam through this system, the size and location of the beam waist can be adjusted by varying the distance between the two lenses. Simulation and experiments show that the size of the transformed waist can be reduced without changing the transformed location of the waist by a probe beam with a larger waist size in the optical transformation system. Before the airborne LDV prototype is assembled, the parameters of the transformation system are optimized through simulation, while considering the size and weight of the LDV. The focuses of the concave and convex lenses are chosen to be -100 mm and 600 mm respectively. Before being input into the optical transformation system, the probe beam is expanded to 10 mm by an 8× expander. The entire LDV system is constructed with a sturdy cage structure, with four metal rods serving as the core skeleton to ensure the coaxial alignment of the optical transformation system. As the attitude of UAVs changes over time, a micro-electromechanical (MEMS) INS has been employed to measure and track these variations in UAV attitude. Additionally, the quality factor of the Doppler signal is defined as the ratio between the amplitude of the Doppler frequency and the mean amplitude of the base in the frequency domain, and it is adopted to represent the CNR in experiments.

After designing, a single-beam airborne LDV prototype is fabricated with a working distance of 50 m and a 10 m depth of field. The spot diameter of the probe beam at 50 m is 0.34 mm. The quality factor has been measured to be 3685 at the working distance of 50 m and remains above 800 throughout the entire depth of field. The depth of field is enough to prevent signal loss and a 110-second flight experiment is conducted with a UAV as the carrier. The velocity measured by the prototype is corrected for the pitch angle recorded by MEMS INS, and the corrected velocity is basically consistent with the velocity recorded by global position system(GPS) as a reference. The entire measurement maintains a high Doppler signal quality factor. However, the accuracy of MEMS INS is insufficient. Although the airborne LDV provides highly precise velocity components along the direction of the probe beam, the accuracy of velocity worsens after correction for pitch angle. The utilization of a two-beam LDV can alleviate this problem since the two beams have different angles for the ground. The velocity can be accurately determined by measuring the Doppler frequency and the angle between the two beams.

The airborne LDV is designed and tested through simulation and experiments. We verify the feasibility of airborne LDV which has promising applications in UAV-integrated navigation. Our study lays a foundation for the development of multi-beam onboard LDV in the future. The acquisition of all accurate velocity components can significantly improve the navigation and localization of UAVs.