How Lateral Doppler Resolution Pinpoints UAV Swarms with Sub-Meter Precision

Nanjing, China, Source:

The rise of unmanned aerial vehicle swarms has created a complex puzzle for modern radar systems. When multiple drones operate in tight formations and execute rapid maneuvers, traditional radar technologies often struggle to distinguish individual units. Limited cross-range resolution means that signals from closely spaced targets can overlap, making it difficult to track each drone separately. A recent study published in Sensors by researchers from the Army Engineering University of PLA offers a compelling solution to this persistent challenge.

This research introduces a robust method for localizing maneuvering UAV swarms by leveraging lateral Doppler resolution. The core innovation lies in integrating three key components: lateral-Doppler scaling, data-driven estimation of equivalent cross-range reference velocity, and acceleration-hypothesis backprojection. By addressing the limitations of radars with small apertures, this approach aims to separate targets that conventional systems might miss.

The technique operates by estimating reference velocity through the maximization of a regularized spatial-concentration criterion. This allows the system to adapt to target-dependent motion by performing pixel-wise maximum fusion over an acceleration grid. In practical tests involving a fixed formation of 15 targets with acceleration scales ranging from 1.8 to 9 m/s², the method demonstrated remarkable precision. Conditional localization root mean square error values ranged between 0.81 and 1.15 meters, while single-scene detection rates achieved between 0.87 and 1.00.

These results stand out when compared to traditional range–Doppler processing and second- or third-order Taylor-Chirplet methods, particularly at higher acceleration scales. Further validation through Monte Carlo experiments, which included randomized noise and Swerling-I scattering models, confirmed the superiority of the proposed method. Over per-pulse signal-to-noise ratios from -18 to -6 dB, maximum fusion yielded higher matched-detection probabilities and lower miss-penalized RMS errors compared to mean fusion and zero-acceleration baselines.

The study also characterized computational trade-offs and operating ranges by sweeping through acceleration grids, pulse repetition frequencies, and reference velocities. As defense and security sectors increasingly rely on drone technology for surveillance, logistics, and combat operations, the ability to accurately detect and localize these assets becomes critical. This research provides a vital technological leap forward, offering a reliable framework for distinguishing individual drones within a swarm even under challenging dynamic conditions.

By enhancing the situational awareness capabilities of radar systems, this innovation contributes significantly to the broader global trend of integrating advanced signal processing techniques into next-generation defense infrastructure. The findings suggest a future where radar systems can maintain clarity and precision even in the most crowded and fast-moving aerial environments.