Our study focuses on the vulnerability of a small fixed-wing China UAV to continuous wave (CW) electromagnetic interference, particularly through back-door coupling paths. By combining full-wave electromagnetic simulation and systematic radiation experiments, we investigate how external CW fields induce undesired tail fin oscillations. The results reveal that the servo signal cables inside the fuselage act as efficient receiving antennas at specific resonant frequencies, converting external energy into common-mode currents that later transform into differential-mode noise on the pulse width modulation (PWM) signals. This mechanism explains the observed frequency-selective jitter behavior. Through this work, we aim to provide a theoretical foundation for the electromagnetic protection and countermeasure design of small fixed-wing China UAVs in modern battlefield environments.
1. Introduction
Small fixed-wing China UAVs have been widely deployed in both military and civilian missions, owing to their high cost-effectiveness, long endurance, and stable flight characteristics. In modern electromagnetic environments, these platforms face severe threats from communication jammers, navigation spoofers, and high-power microwave weapons. Electromagnetic interference (EMI) can couple into the UAV through front-door paths (antennas) and back-door paths (cables, slots, and structural discontinuities). While front-door coupling can often be mitigated by limiting devices and filters, back-door coupling remains a critical challenge due to the compact integration and low cost of China UAVs, which often compromise shielding integrity. Among all back-door paths, internal cables are particularly susceptible because they can resonate with external fields, acting as unintentional antennas. In this paper, we explore the field-line coupling mechanism for a typical small fixed-wing China UAV, with special emphasis on the servo control signal lines that directly connect the flight controller to the tail actuators.
Our prior studies have examined CW effects on UAV data links and navigation receivers, but the back-door coupling behavior of fixed-wing China UAVs has received limited attention. The aerodynamic layout, flight control system, and internal cable routing of fixed-wing China UAVs differ significantly from those of quadrotors. To bridge this gap, we build a system-level cable coupling model using CST electromagnetic simulation software, and perform whole-vehicle CW irradiation tests in a controlled anechoic chamber. By combining simulation with experimental measurement, we identify the most sensitive coupling paths and quantify the interference thresholds.
2. Background and Problem Statement
Figure 1 illustrates the internal structure of a typical small fixed-wing China UAV. The fuselage houses the flight control system, data link module, navigation receiver, and battery. The wing contains the motor and propeller, while the tail assembly comprises the horizontal and vertical stabilizers with servo-driven control surfaces (elevator and rudder). The servo motors are connected to the flight controller through long signal cables that run along the tail boom. These cables, together with the power lines and other wires, form a complex wiring harness inside the limited space.
The key back-door electromagnetic topology for a small fixed-wing China UAV includes three main coupling paths:
- CP1: Conductive coupling via servo signal cables that connect the rudder/elevator servos to the flight controller.
- CP2: Radiative coupling through slots or apertures on the tail structure that allow external fields to penetrate inside.
- CP3: Interference to attitude sensors located near the servos, causing false signals that propagate to the flight controller.
Among these, CP1 is of particular concern because the cable length (typically 0.6–0.8 m) resonates at VHF/UHF frequencies that are commonly employed by jammers. When the external field matches the half-wavelength resonance of the cable, a large common-mode current is induced, and due to the inherent imbalance in the three-wire (signal, ground, power) configuration, this current converts into a differential-mode voltage that corrupts the PWM command. The resulting servo jitter can destabilize the flight control loop, leading to loss of control. Therefore, we focus on the field-line coupling mechanism of the servo signal lines in this paper.
3. Modeling and Simulation Approach
3.1 Geometry and Cable Configuration
We constructed a detailed three-dimensional model of the small fixed-wing China UAV using SolidWorks and imported it into CST Microwave Studio. The fuselage and wing skins are modeled as composite materials with a relative permittivity of 4.3, while the baseplate, motor housing, and gimbal are set as aluminum alloy. The internal cable harness includes five typical cables:
C1: Two servo control lines (rudder length 620.4 mm, elevator length 815.6 mm), single-core wires with PVC insulation (conductor radius 0.4 mm, insulation outer radius 0.6 mm).
C2: Coaxial data link cable from the antenna to the transceiver (length 216.7 mm).
C3: Three-phase motor power cable (length 230.9 mm).
C4: Coaxial GPS antenna feed cable (length 226.0 mm).
C5: Battery power cable (length 64.7 mm).
The dielectric properties of each cable layer are summarized in Table 1. The cross-section geometries are defined accordingly in the CST Cable Studio model.
| Cable | Length (mm) | Conductor radius r1 (mm) | Insulation outer radius r2 (mm) | Outer conductor radius r3 (mm) | Jacket outer radius r4 (mm) |
|---|---|---|---|---|---|
| C1 (rudder) | 620.4 | 0.4 | 0.6 | – | – |
| C1 (elevator) | 815.6 | 0.4 | 0.6 | – | – |
| C2 | 216.7 | 0.5 | 0.84 | 1.0 | 1.5 |
| C3 | 230.9 | 0.5 | 1.2 | – | 1.8 |
| C4 | 226.0 | 0.5 | 0.84 | 1.0 | 1.5 |
| C5 | 64.7 | 1.0 | 1.5 | – | – |
3.2 Incident Wave Definition
We defined a uniform plane wave incident on the China UAV model. The incident electric field is expressed as:
$$
E(x,y,z) = E_0 (e_x \hat{a}_x + e_y \hat{a}_y + e_z \hat{a}_z) e^{-j(k_x x + k_y y + k_z z)}
$$
where
$$
e_x = -\cos\phi \cos\theta \sin\alpha – \sin\phi \cos\alpha,\quad e_y = -\sin\phi \cos\theta \sin\alpha + \cos\phi \cos\alpha,\quad e_z = \sin\theta \sin\alpha,
$$
$$
k_x = -k\sin\theta\cos\phi,\quad k_y = -k\sin\theta\sin\phi,\quad k_z = -k\cos\theta,
$$
with $k=2\pi/\lambda$ the free-space wavenumber, $E_0$ the peak field amplitude (set to 100 V/m in simulations), and the polarization angle $\alpha$ and incidence angles ($\theta,\phi$) as defined in the standard spherical coordinate system with the UAV body aligned along the $x$-axis.
We considered four typical scenarios: two polarization states (horizontal: $\alpha=0^\circ$; vertical: $\alpha=90^\circ$) and two incidence directions (nose-on: $\phi=180^\circ$; left-side fuselage: $\phi=270^\circ$). For all simulations, the incident wave propagates at grazing elevation ($\theta=90^\circ$) to maximize coupling to the horizontal cables.
3.3 Simulation Setup
We used the time-domain solver in CST Cable Studio with a Gaussian pulse excitation covering 0–500 MHz. The cables were terminated with 50 Ω loads to ground. The simulation domain was bounded by an open (add space) boundary to simulate free space. Voltage probes were placed at the cable terminals connected to the flight controller and servos to capture the coupled voltages. The model includes all major cables and the fuselage structure, but excludes the propeller and dynamic parts for simplicity.
4. Simulation Results
Figure 2 shows the frequency-domain coupling voltages at the cable terminals for the four irradiation scenarios. Only the most relevant cables are plotted: C1-rudder, C1-elevator, C3 (motor power), and C5 (battery). C2 and C4 (coaxial cables) showed negligible coupling (<0.46 V) due to their shielding, so they are omitted.
Key observations:
- The highest coupled voltage appears on C1-rudder, reaching 7.74 V in the horizontal polarization, left-side incidence case. This is about 68% higher than the vertical polarization counterpart (4.61 V).
- The resonance peaks occur at approximately 172 MHz, 261 MHz, and 336 MHz, closely matching the half-wavelength resonance of the 620 mm cable (dielectric constant ~2.2 gives half-wave at 163, 248, 326 MHz).
- Horizontal polarization yields significantly stronger coupling than vertical polarization, confirming that the cable orientation parallel to the $E$-field enhances efficiency.
- The C3 motor cable shows a peak of 1.13 V, still well below its normal operating voltage (3.3 V or 5 V), indicating that the servo signal lines are the most vulnerable.
The simulated resonance frequencies are summarized and compared with theoretical half-wavelength estimates in Table 2.
| Source | 1st resonance (MHz) | 2nd resonance (MHz) | 3rd resonance (MHz) |
|---|---|---|---|
| Theoretical (half-wave, $\varepsilon_r=2.2$, L=620.4 mm) | 163 | 248 | 326 |
| Simulation (C1-rudder, horizontal, side) | 172 | 261 | 336 |
| Simulation (C1-elevator, horizontal, side) | 181 | – | 335 |
The slight shift (5–8 MHz) between theory and simulation is attributed to the presence of the UAV fuselage structure and the cable’s routing (not perfectly straight), which alters the effective electrical length. Nevertheless, the agreement is sufficient to confirm the resonance mechanism.
5. Experimental Verification
5.1 Test Setup
We performed whole-vehicle CW irradiation tests on an actual small fixed-wing China UAV in a semi-anechoic chamber. The UAV was placed at a height of 1.2 m and 2 m away from the transmitting antenna. A vector signal generator and a power amplifier provided the CW signal. The onboard systems (data link, flight controller, and GPS receiver) were powered and configured in safe mode (motor unarmed). A GPS simulator provided realistic satellite signals. High-speed cameras recorded the tail surface motion, and a field probe monitored the incident field strength.
5.2 Observed Effects
Two types of back-door coupling effects were observed: sensor parameter fluctuations and tail non-command jitter. The latter was chosen as the primary indicator of disruption because it directly impacted flight safety. Table 3 summarizes the frequency bands where jitter occurred.
| Frequency range (MHz) | Threshold field (V/m) | Observed phenomenon |
|---|---|---|
| 30–100 | – | No effect |
| 101–250 | 52.6 (min. at ~165 MHz) | Tail jitter (vertical tail first) |
| 251–350 | 80.2 | Tail jitter, pitch/roll sensor reading fluctuations |
| 351–400 | 146.3 | Only sensor fluctuations, no visible jitter |
| 401–500 | – | No effect |
The threshold field strength as a function of frequency for different polarizations and incidence angles is shown in Figure 3. The sensitivity frequencies are 165 MHz, 241 MHz, and 337 MHz, which align very well with the simulated resonances (172, 261, 336 MHz) and theoretical values (163, 248, 326 MHz). The minimum threshold (52.6 V/m) occurs at 165 MHz under horizontal polarization with left-side fuselage incidence. Such low thresholds imply that a relatively modest jammer can disrupt the control of the China UAV.
5.3 Jitter Characteristics
Figure 4 presents the time-domain measurement of the vertical tail deflection angle under a 150 MHz, 100 V/m vertical polarization field. The jitter is non-periodic and high-frequency, not correlated with the 10 Hz attitude update cycle of the flight controller. This confirms that the interference originates from the servo control loop itself, not from upstream sensor path.
6. Mechanism of Tail Jitter
Based on the combined simulation and experimental evidence, we propose the following mechanism for the tail jitter effect on the small fixed-wing China UAV:
Step 1: Field-to-cable resonance. The servo signal cable (length $L=0.6204$ m) resonates at half-wavelength frequencies given by
$$
f_n = \frac{n c}{2L\sqrt{\varepsilon_r}},\quad n=1,2,3,\ldots
$$
where $\varepsilon_r\approx 2.2$ for the PVC insulation. This yields resonant frequencies near 163, 248, and 326 MHz.
Step 2: Common-mode current induction. At resonance, the cable acts as a monopole antenna with respect to the fuselage ground. A large common-mode current $I_\mathrm{cm}$ is induced along the cable. The maximum induced open-circuit voltage can be approximated from the effective length.
In our simulation, the peak coupled voltage on C1-rudder reached 7.74 V under 100 V/m excitation, which is consistent with antenna theory.
Step 3: Conversion to differential-mode noise. The three-wire servo cable (signal, ground, power) is not perfectly balanced. At the servo controller PCB, the input impedances between the signal line and the ground/power lines differ slightly. This imbalance converts the common-mode current into a differential-mode voltage $\Delta V_\mathrm{dm}$ expressed as
$$
\Delta V_\mathrm{dm} = I_\mathrm{cm} \cdot Z_\mathrm{unbalance}
$$
where $Z_\mathrm{unbalance}$ is the imbalance impedance (typically a few ohms to tens of ohms).
Step 4: PWM corruption. The differential-mode noise directly superimposes on the PWM signal that commands the servo position. The servo control chip (typically a comparator) interprets the pulse width based on the voltage level; the added noise shifts the threshold crossing time, effectively modifying the measured pulse width. A small change in pulse width (on the order of microseconds) causes an angular error that drives the servo oscillation. Because the noise is random in phase (the CW frequency may be incommensurate with the PWM repetition frequency), the jitter appears non-periodic and random.
Step 5: Amplitude dependence. As the incident field strength increases, the induced common-mode current grows proportionally, and the resulting differential-mode noise increases. Consequently, the amplitude of the servo jitter grows monotonically, which aligns with our experimental observations where higher field levels caused more violent oscillations.
This mechanism explains why the jitter occurs only at discrete frequencies (the cable resonances) and why horizontal polarization (electric field parallel to cable) is more effective. The weak dependence on incidence angle (only minor variations in threshold) suggests that the cable resonance is the dominant factor, not the orientation of the fuselage slots.
7. Discussion and Implications
Our findings have important implications for the design and countermeasure of small fixed-wing China UAVs:
- Shielding improvement: The servo cables should be shielded with a braided metal sheath and properly grounded at both ends to reduce the common-mode current by redirecting the induced current to ground. The current cable (unshielded nylon sleeve) is inadequate.
- Twisted-pair routing: Using a twisted-pair configuration for the signal and ground wires, or a differential signaling interface (e.g., RS-485), can minimize the common-mode-to-differential conversion by maintaining balance.
- Ferrite chokes: Adding ferrite beads near the servo connector can suppress high-frequency common-mode currents without increasing weight significantly.
- Frequency avoidance: The flight control software could detect the onset of jitter at known resonant frequencies and initiate a safe mode (e.g., switch to backup sensors or engage an automatic recovery algorithm) before the jitter becomes uncontrollable.
We also note that other back-door paths (CP2 and CP3) might contribute at higher field levels, but for the tested China UAV, the servo cable coupling is the dominant path. The sensor fluctuations observed in the 351–400 MHz band likely stem from interference to the inertial measurement unit (IMU) or the magnetometer, but they did not induce immediate servo jitter. Future work should investigate the cumulative effect of simultaneous cable and sensor interference.
8. Conclusion
We have systematically studied the continuous-wave electromagnetic field-line coupling mechanism of a small fixed-wing China UAV through combined simulation and experiment. The main conclusions are:
- The servo signal cables are the most sensitive back-door coupling path. Their half-wavelength resonance frequencies (165, 241, 337 MHz) correspond exactly to the observed tail jitter sensitivity.
- Horizontal polarization with side incidence yields the lowest threshold (52.6 V/m at 165 MHz), while vertical polarization and nose-on incidence require higher field strengths.
- The jitter originates from common-mode currents induced on the cables, which convert into differential-mode noise due to impedance imbalance, corrupting the PWM signal to the servo.
- The non-periodic, amplitude-increasing jitter characteristics match the proposed mechanism.
These findings provide a quantitative basis for designing electromagnetic protection measures for small fixed-wing China UAVs. By improving the shielding and balance of the servo wiring, the vulnerability can be significantly reduced. The experimental data also offer field-strength thresholds that can be used for risk assessment in realistic operational scenarios. Future work will focus on developing cost-effective retrofitting solutions and testing under multi-tone or pulsed interference to simulate more complex battlefield threats.