In recent years, the rapid advancement of small unmanned aerial vehicle (UAV) technology, including improvements in autonomous navigation, obstacle avoidance, hovering capabilities, and video display systems, has led to widespread applications in both industrial and consumer sectors. However, this proliferation has also increased the risks associated with UAVs, such as accidental malfunctions and potential terrorist attacks. These incidents pose significant challenges to public security and management, prompting governments worldwide to seek effective countermeasures. Anti-UAV defense systems (AUDS) have emerged as a critical solution, typically integrating radar, electro-optical sensors, and radio frequency jammers to detect, track, identify, and neutralize rogue drones. Among these components, radar plays an indispensable role due to its all-weather, all-day, long-range, and wide-area surveillance capabilities, independent of the target’s electromagnetic characteristics. This paper explores a novel radar detection system designed specifically for anti-UAV operations, addressing the unique challenges posed by “low, slow, and small” (LSS) targets. We propose a digital circular phased array radar system that overcomes the limitations of conventional one-dimensional mechanically scanned phased arrays. Through detailed analysis, simulation, and field testing, we demonstrate the feasibility and effectiveness of this new anti-UAV radar paradigm.
The detection of UAVs using radar is notoriously difficult due to their LSS characteristics. These targets typically fly at low altitudes (below 1,000 meters), move at slow speeds (often less than 20 m/s), and have small physical sizes with weak radar cross-sections (RCS). For instance, consumer-grade multi-rotor drones can have an RCS as low as 0.01 m², making them susceptible to being淹没 in strong clutter backgrounds. The technical challenges in anti-UAV radar detection can be summarized in several key aspects. First, low-altitude flight leads to long-range detection blind spots caused by Earth curvature and multipath effects, and spatially, these targets are proximate to ground clutter like buildings and terrain. Second, slow speeds result in Doppler frequencies that are closely aligned with clutter spectra, complicating motion detection. Conventional radars often have minimum detectable velocities above 10 m/s, which may fail to capture slower UAVs. Third, in urban environments, clutter-to-noise ratios can exceed 80 dB, with strong reflections from structures reaching 100–120 dB, making clutter suppression arduous and leaving residual noise that masks weak UAV signals. Fourth, UAVs are non-inertial targets with erratic flight patterns, increasing the likelihood of track loss if data update rates are insufficient. Existing anti-UAV radars commonly employ one-dimensional planar phased arrays with mechanical azimuth scanning and electronic elevation scanning. While cost-effective and mature, this体制 lacks flexibility in beam steering, supporting only track-while-scan (TWS) modes without true track-and-search (TAS) capabilities. This limitation can compromise tracking continuity and accuracy for agile UAVs. To address these issues, our research introduces a digital circular phased array radar system that enables full 360-degree electronic scanning, uniform performance across all directions, and high-data-rate TAS functionality, thereby enhancing anti-UAV detection and tracking efficacy.
To better illustrate the characteristics of typical UAV targets, we present a summary in Table 1. This table highlights the differences between fixed-wing and multi-rotor drones, emphasizing their small RCS, low flight altitudes, and slow speeds, which collectively contribute to the detection challenges in anti-UAV operations.
| Characteristic | Fixed-Wing UAV | Multi-Rotor UAV |
|---|---|---|
| Typical Size | Wingspan 1–2 m | Wheelbase ≤ 0.6 m |
| Radar Cross-Section (RCS) | ≤ 0.1 m² | ≤ 0.01 m² |
| Flight Altitude | 50–1000 m | ≤ 500 m |
| Flight Speed | Typical: 15–20 m/s, Max: 50 m/s | Safe: 5–20 m/s, Max: 25 m/s |
The proposed anti-UAV radar system is based on a digital circular phased array antenna, which facilitates two-dimensional electronic scanning in both azimuth and elevation. Unlike linear arrays, a circular array consists of antenna elements uniformly distributed along a ring, allowing for symmetric beamforming and eliminating mechanical rotation delays. This design is particularly advantageous for anti-UAV applications, as it enables rapid beam steering to track multiple dynamic targets simultaneously. The core principle involves digital beamforming (DBF), where signals from selected antenna elements are combined with precise phase shifts to synthesize a directional beam. For a circular array with M elements arranged on a circle of radius R, the angular position of the m-th element is given by:
$$ \theta_m = \frac{2\pi m}{M}, \quad m = 0, 1, 2, \dots, M-1 $$
When forming a beam in the direction of an incoming signal with azimuth angle φ, a subset of N elements (typically N ≈ M/3 to M/4) is used to create a narrow beam. The element directivity pattern is often modeled with a cosine function to account for directional sensitivity:
$$ f(\theta_k, \varphi) =
\begin{cases}
\cos(\theta_k – \varphi), & |\theta_k – \varphi| \leq \pi/2 \\
0, & |\theta_k – \varphi| > \pi/2
\end{cases} $$
To achieve low sidelobes, the selected elements are projected onto a diameter perpendicular to φ, with the projection distance for the k-th element expressed as:
$$ x_k = R \cos(\theta_k – \varphi) $$
The beamformed output signal is then computed by applying a window function \( w_k \) (e.g., a Taylor or Chebyshev window) to control sidelobe levels, combined with phase compensation for the path differences. For a set of N elements indexed from i to i+N-1, the output is:
$$ S_{\text{out}} = \sum_{k=i}^{i+N-1} f(\theta_k, \varphi) w_k S_k \exp\left(-j\frac{2\pi x_k}{\lambda}\right) $$
where \( S_k \) is the received signal from the k-th element, \( \lambda \) is the wavelength, and the exponential term accounts for the phase shift due to the projection distance. This digital beamforming approach allows for flexible and adaptive beam shaping, which is crucial for suppressing clutter and enhancing target detection in anti-UAV scenarios. The system can operate in two scanning modes: single-beam azimuth scanning, where the beam is electronically steered by cycling through element weight vectors, and multi-beam scanning, which generates multiple simultaneous beams for higher data rates at the cost of increased hardware complexity. For our anti-UAV radar, we primarily employ single-beam scanning with optimized windowing to balance beamwidth and sidelobe performance.
To validate the beamforming performance of the circular array, we conducted digital simulations using MATLAB. The radar operates at a frequency of 6 GHz (S-band), with an array radius of 25 cm and 48 antenna elements. For beam synthesis, we select N=12 contiguous elements and apply a -27 dB Taylor window to suppress sidelobes. The simulated transmit beam pattern in the horizontal plane is shown in Figure 1 (though not explicitly referenced, it illustrates the concept), with a beamwidth of approximately 7.6° and a maximum sidelobe level of -14.5 dB. The receive beam pattern, under similar conditions, achieves a beamwidth of 8.8° and a maximum sidelobe level of -25 dB. These results confirm that the digital circular array can produce narrow, low-sidelobe beams suitable for discriminating UAV targets from clutter. The uniformity of the beam across all azimuth angles is a key advantage for consistent anti-UAV coverage, unlike mechanically scanned systems that may suffer from performance variations at different orientations.
The engineering implementation of the anti-UAV radar system involves several key components, as outlined in the block diagram (Figure 2). The system leverages direct digital synthesis (DDS) for precise phase control in the transmit path and digital beamforming in the receive path. In transmit mode, the radar control processor calculates the desired beam direction and determines the corresponding set of active antenna elements and their initial phase values. These parameters are fed to a multi-channel signal generator, which uses DDS to produce RF signals with the specified phases for each selected channel. The signals are then amplified by transmit/receive (T/R) modules and radiated through the corresponding column-fed antenna elements, synthesizing a focused beam in the intended direction. This process enables rapid azimuth scanning without mechanical movement, essential for tracking fast-maneuvering UAVs in anti-UAV operations.
In receive mode, echoes are captured by the same antenna elements and passed through low-noise amplifiers, filters, mixers, and analog-to-digital converters (ADCs) in the T/R modules. The digitized I/Q signals are processed using DBF algorithms to form sum and difference beams for the specified azimuth direction. This involves applying the phase shifts and window functions as derived in the earlier equations. The beamformed data is then sent to a signal processor for pulse compression, coherent integration, and detection. A data processor further refines the target tracks through correlation and filtering algorithms, outputting trajectory points to a display for operator monitoring. The entire system is designed for portability and real-time performance, making it suitable for field deployments in anti-UAV defense networks. Key parameters of the radar prototype are summarized in Table 2, highlighting its capabilities in addressing LSS targets.
| Parameter | Value |
|---|---|
| Operating Frequency | S-band (6 GHz) |
| Array Configuration | Circular phased array with 48 elements |
| Beamforming Method | Digital beamforming (DBF) |
| Scanning Coverage | 360° azimuth, ±30° elevation |
| Beamwidth (Azimuth) | Approx. 8° |
| Sidelobe Level | < -25 dB (receive) |
| Detection Range for UAVs | Up to 5 km (for RCS ≥ 0.01 m²) |
| Tracking Update Rate | Up to 2 Hz (TAS mode) |
| Clutter Suppression | Pulse Doppler processing with MTI |
Field trials were conducted to evaluate the performance of the anti-UAV radar prototype against real-world drone threats. The tests involved a consumer-grade DJI Phantom 4 UAV, which has an RCS of about 0.01 m², flies at altitudes below 400 m, and speeds under 10 m/s—representative of LSS targets. The radar was configured in a low-altitude detection mode optimized for anti-UAV operations, using high pulse repetition frequency (PRF) Doppler processing to enhance sensitivity to slow-moving targets. In search mode, the radar scans the airspace with a 5-second update rate, and upon detecting a potential UAV, it automatically switches to TAS mode with a 1-second update rate for precise tracking. During the trials, the radar successfully detected the DJI Phantom 4 at distances exceeding 3 km and maintained stable tracking, as evidenced by continuous trajectory plots on the display. The target tracks (shown in blue in Figure 3) demonstrated the system’s ability to handle non-inertial maneuvers without loss, a critical requirement for effective anti-UAV defense. Moreover, the radar’s outputs were used to guide electro-optical sensors and RF jammers for target identification and neutralization, validating its integration into a comprehensive AUDS.
The effectiveness of the anti-UAV radar system can be further analyzed through its signal processing gains. For a UAV with RCS σ at range R, the received signal power \( P_r \) is given by the radar range equation:
$$ P_r = \frac{P_t G_t G_r \lambda^2 \sigma}{(4\pi)^3 R^4 L} $$
where \( P_t \) is transmit power, \( G_t \) and \( G_r \) are transmit and receive gains, \( \lambda \) is wavelength, and L represents system losses. The circular array’s high gain and adaptive beamforming improve \( G_t \) and \( G_r \), while pulse-Doppler processing enhances the signal-to-clutter ratio (SCR). The minimum detectable velocity \( v_{\text{min}} \) is influenced by the Doppler resolution \( \Delta f_d \):
$$ v_{\text{min}} = \frac{\lambda \Delta f_d}{2} $$
where \( \Delta f_d = 1/T_{\text{dwell}} \), with \( T_{\text{dwell}} \) being the beam dwell time. By employing longer dwell times in focused tracking modes, our anti-UAV radar achieves \( v_{\text{min}} \) values as low as 2–3 m/s, enabling detection of very slow drones. Additionally, the system’s clutter suppression capability is quantified by the improvement factor (IF) of the moving target indicator (MTI) filter, which can exceed 40 dB for urban environments. This performance is crucial for distinguishing UAVs from background clutter in dense settings.
In conclusion, this paper presents a novel anti-UAV radar detection system based on a digital circular phased array, offering significant advantages over traditional one-dimensional scanned radars. The system addresses the core challenges of LSS target detection through full electronic scanning, high data rate tracking, and uniform azimuthal performance. Digital beamforming allows for flexible and low-sidelobe beam synthesis, enhancing clutter suppression and target discrimination. Simulation and field test results confirm the practicality and efficacy of this approach for anti-UAV applications. Future work will focus on networking multiple circular array radars to form a distributed anti-UAV defense grid, improving coverage and resilience against swarm attacks. Additionally, advanced machine learning algorithms could be integrated for automatic target recognition and threat assessment, further strengthening anti-UAV capabilities. As UAV threats continue to evolve, innovative radar technologies like the one proposed here will play a vital role in safeguarding airspace security.