How Radar Works for Drone Detection

Why Radar Is the Foundation of C-UAS Detection

Every serious counter-UAS architecture starts with radar. Electro-optical sensors can't see through cloud. Acoustic arrays top out at a few hundred meters in ambient noise. RF detection requires the adversary to be transmitting. Radar is the only sensor modality that provides long-range, all-weather, passive-target detection — which is why the U.S. Army chose radar as the primary cueing layer for both the Coyote interceptor and the M-SHORAD system.

The challenge is that radar was originally designed for aircraft with radar cross-sections (RCS) measured in square meters. A Group 1 UAS — anything under 55 pounds — presents an RCS between 0.001 and 0.1 m², comparable to a large bird or a metallic balloon. That physics constraint drives every design decision in a dedicated counter-UAS radar.

Radar Fundamentals: How Electromagnetic Detection Works

The Range Equation

The received power from a radar return follows the radar range equation:

P_r = (P_t × G² × λ² × σ) / ((4π)³ × R⁴ × L)

Where P_t is transmitted power, G is antenna gain, λ is wavelength, σ is target RCS, R is range, and L captures system losses. The R⁴ term is brutal: double the range, receive 1/16th the power. For a target with 0.01 m² RCS versus one with 1.0 m² RCS, detection range drops by a factor of 3.2 (the fourth root of 100). This is why KURFS — built from the AN/MPQ-64F1 Sentinel airspace surveillance radar — can detect an aircraft at 40 km but may only reliably detect a DJI Phantom-class drone at 8–12 km under favorable conditions.

Pulse vs. Continuous Wave

Pulsed radar transmits short bursts of energy and listens for returns in the silence between pulses. Pulse repetition frequency (PRF) determines the unambiguous range and velocity windows. High PRF provides good velocity resolution but can create range ambiguities. Low PRF avoids range ambiguity but aliases fast targets. Modern systems use staggered PRF or multiple waveforms simultaneously.

Continuous Wave (CW) radar transmits constantly and receives on a separate antenna (or a circulator). FMCW (frequency-modulated continuous wave) sweeps the transmit frequency linearly, allowing both range and velocity extraction from the beat frequency. FMCW is compact, inexpensive, and increasingly common in commercial drone detection systems. The tradeoff: CW systems struggle with simultaneous multi-target resolution and have limited maximum range due to transmitter-receiver isolation requirements.

Pulse-Doppler: The Standard for Airspace Surveillance

Pulse-Doppler combines the range resolution of pulsed waveforms with Doppler velocity processing. By coherently integrating received pulses, the system applies a Fast Fourier Transform (FFT) to detect frequency shifts caused by target motion. A target moving at velocity v produces a Doppler shift:

f_d = (2 × v × cos θ) / λ

For a 10 GHz (X-band) radar, a drone flying at 20 m/s produces a Doppler shift of approximately 1.3 kHz — well within the processing capability of modern digital receivers. The cos θ term is critical: a target flying perpendicular to the radar line-of-sight produces zero Doppler shift and can fall into the clutter notch.

AESA: Electronic Beam Steering Changes Everything

Active Electronically Scanned Array (AESA) radar replaced mechanical dish rotation with a grid of individually controllable transmit/receive (T/R) modules. By adjusting the phase of each element, the beam can be steered anywhere in the antenna's field of view in microseconds — without moving parts.

For counter-UAS, this enables two critical capabilities:

Staring mode: Instead of scanning a 360° volume and revisiting each sector every rotation period (typically 4–12 seconds for legacy systems), an AESA can allocate its dwell time adaptively. If a threat is detected at bearing 045°, the radar can immediately dedicate additional dwells to that sector while maintaining coarser coverage elsewhere. The Giraffe 1X uses this approach to achieve high track continuity on slow-moving Group 1 UAS that would be missed between scans by rotating systems.

Simultaneous multi-function: A single AESA aperture can generate multiple independent beams, some performing search while others track. IBCS exploits this by fusing AESA track data from multiple radars across a network, constructing a composite track picture that no single node could achieve alone.

Micro-Doppler: The Signature That Exposes Rotors

A drone's fuselage contributes one Doppler signature based on its translation velocity. But its spinning rotors contribute additional frequency components — sidebands around the main Doppler return — corresponding to rotor tip velocity and blade flash rate. This is the micro-Doppler signature.

For a quadcopter with 20 cm diameter rotors spinning at 6,000 RPM, tip velocity reaches approximately 63 m/s. This produces micro-Doppler sidebands spanning ±8 kHz at X-band — detectable as distinctive spectral structure absent from bird returns or aircraft returns.

The challenge is algorithmic: extracting micro-Doppler reliably in the presence of ground clutter, multi-path, and competing targets requires time-frequency analysis (short-time FFT, wavelet decomposition, or neural network classifiers trained on measured drone signatures). KURFS uses proprietary discrimination algorithms developed by Raytheon. DroneShield's RfPatrol radar augments RF detection with micro-Doppler classification to reduce false alarms below operationally relevant thresholds.

The Clutter Problem: Why Low Altitude Is Hard

Ground clutter — radar returns from terrain, buildings, trees, and precipitation — occupies the same Doppler bins as slow-moving UAS flying at low altitude. A Group 1 drone flying at 50 m AGL at 15 m/s may be indistinguishable from moving tree canopy in a 15 knot crosswind without sophisticated processing.

Legacy Moving Target Indicator (MTI) filters cancel clutter by subtracting successive pulse returns, revealing moving targets. But MTI also cancels targets moving at the blind speed — when Doppler shift aliases to zero or a multiple of PRF. Adaptive clutter cancellation (space-time adaptive processing, STAP) uses multiple antenna elements and pulse integration together to suppress clutter while preserving target returns even near blind speeds.

The AN/TPS-80 G/ATOR employs STAP processing to maintain detection of low-slow-small targets in cluttered environments — a requirement driven by experience in Iraq and Afghanistan where insurgent drones exploited terrain masking aggressively.

3D vs. 2D Radar Architecture

Legacy air defense radars often used separate search and height-finder radars — one for azimuth/range, one for elevation. This slowed reaction time and created track handoff gaps. Counter-UAS radars must be 3D from the start.

Stacked-beam 3D radar uses multiple simultaneous receive beams at different elevation angles, stacked vertically. Returns are compared across beams to compute elevation angle without requiring additional transmit time. LSTAR uses this approach to provide altitude data on artillery shells — the same beams work for UAS detection.

AESA with elevation scan uses the same phased array to scan in both azimuth and elevation. This provides flexibility — beams can be directed to specific altitude layers — but requires more dwell time than stacked-beam since the antenna must physically point to each elevation sequentially.

For C-UAS, 3D coverage is non-negotiable. A UAS detected in 2D at range and azimuth cannot be handed off to a fire control system — whether a Coyote launcher or a directed energy weapon — without elevation data.

System Examples: KURFS, LSTAR, Giraffe 1X

KURFS (Ku-band Radio Frequency System) operates in Ku-band (12–18 GHz) rather than X-band, providing shorter wavelength and thus better RCS sensitivity for small targets. Its 360° staring AESA provides continuous coverage without scan-revisit gaps. KURFS was specifically developed to cue Coyote interceptors, and the KURFS-Coyote pairing has been the cornerstone of U.S. Army C-UAS since its combat debut in the Middle East circa 2019.

AN/TPQ-50 LSTAR operates in X-band and was originally a counter-rocket, artillery, and mortar (C-RAM) radar. Its ability to detect and track small ballistic objects translates reasonably well to Group 1–2 UAS, though its scan rate creates revisit gaps that KURFS avoids. LSTAR is deployed widely because it is already in the Army inventory — C-UAS capability is often added as a software update rather than a new procurement.

Giraffe 1X (Saab) operates in X-band with a software-defined AESA that can be configured for air surveillance, UAS detection, or ground surveillance. Its staring-mode capability — dwelling on a sector rather than rotating — makes it particularly effective against slow, low-flying UAS that would fall through the cracks of a rotating system. Sweden, the UK, and several NATO partners have fielded Giraffe 1X specifically for forward base C-UAS protection.

What Limits Detection of Small UAS

Detection range for Group 1 UAS is governed by three hard constraints:

1. RCS physics: Small plastic and composite airframes simply do not reflect much energy. Counter-UAS radars compensate with higher transmitted power, lower noise figures, and longer integration times — all of which cost size, weight, and power (SWaP).

2. Clutter floor: In cluttered environments, the limiting factor is not noise but competing returns from the environment. No amount of transmitter power overcomes a clutter floor that is stronger than the target return.

3. Discrimination false alarm rate: A radar that detects everything also triggers on birds, balloons, and blowing debris. Operators can sustain only a finite rate of false alarms before alert fatigue degrades response quality. Micro-Doppler discrimination and multi-sensor fusion (optical, RF, acoustic cueing) are the primary tools for controlling false alarm rate without sacrificing probability of detection.

The next frontier is cognitive radar — systems that adapt their waveform, PRF, and beam scheduling in real time based on the environment and detected threats. DARPA and Army Research Laboratory programs are advancing these concepts, with the goal of achieving order-of-magnitude improvements in small UAS detection range without corresponding increases in SWaP.