High Power Microwave (HPM) technology has emerged as a transformative countermeasure against unmanned aerial vehicles (UAVs), particularly micro-small fixed-wing platforms that pose significant challenges in modern conflict scenarios. Our team conducted a dynamic disruption experiment using a narrowband HPM experimental system to assess its effectiveness against a specific micro-small fixed-wing UAV. The objective was to investigate the real-time flight behavior under high-power electromagnetic pulse (HEMP) irradiation and to identify the underlying mechanisms leading to loss of control. The experimental results clearly demonstrate that narrowband HPM can rapidly disrupt the flight control system of a fixed-wing UAV, causing severe attitude deviations and eventual crash. This paper presents the experimental system design, the measured flight data, and a detailed analysis of the coupling processes that led to the UAV’s loss of control.
The proliferation of micro-small fixed-wing UAVs in intelligence, surveillance, precision strike, and swarm operations has exposed vulnerabilities in conventional air defense systems. These UAVs are difficult to detect and neutralize due to their small radar cross-section, low altitude, and maneuverability. Among various countermeasure technologies, HPM weapons offer distinct advantages: they deliver effects at the speed of light, have a high cost-effectiveness against swarms, and possess unlimited magazine capacity. However, the complexity of HPM interaction with UAV electronics—especially through “back-door” coupling paths via cables, slots, and apertures—necessitates experimental validation. Static effect tests alone cannot fully predict dynamic flight responses. Our work fills this gap by conducting a live-flight experiment with a narrowband HPM system, recording onboard flight data before and after irradiation, and analyzing the root cause of the UAV’s failure.
The experimental system, depicted conceptually in Figure 1, comprised a target acquisition subsystem and an HPM transmitter subsystem. The target acquisition subsystem utilized a radar and optoelectronic tracker to detect and precisely track the UAV, providing real-time guidance data. The HPM transmitter subsystem was based on relativistic vacuum electronics, including a charging power supply, a high-voltage pulse driver (Marx generator), an HPM source, a radiating antenna, control and monitoring equipment, and a vacuum molecular pump group. The charging power supply delivered up to 50 kV using constant-current charging to the high-voltage energy storage capacitors of the pulse driver. The Marx generator produced a high-voltage pulse of several hundred kilovolts, which excited the HPM source to generate gigawatt-level electromagnetic pulses radiated by the antenna. The vacuum system maintained a low-pressure environment to enhance power handling capability.
Table 1 summarizes the key parameters of the narrowband HPM experimental system, obtained through calibration and measurement. The output voltage of the pulse driver exhibited a linear relationship with the charging voltage, as confirmed by water-load tests. A typical output waveform at 520 kV is shown in Figure 2 (not reproduced here per instruction). The radiated waveform was well-formed and the spectrum was clean, indicating stable operation in the desired mode. The system’s ability to operate in burst mode with repetition rates was verified, meeting the requirements for the disruption experiment.
| Parameter | Value / Description |
|---|---|
| Charging voltage range | 0 – 50 kV |
| Pulse driver output voltage range | 300 – 600 kV |
| Typical output voltage (test) | 520 kV |
| Estimated peak power (at antenna) | ~1 GW |
| Center frequency (narrowband) | S-band (classified) |
| Pulse width | ~100 ns |
| Repetition rate | Up to 100 Hz (burst) |
| Antenna type | Conical horn with vacuum window |
| Insulating gas in Marx generator | SF₆ or N₂ |
| Protection circuit | Anti-reverse high-voltage diode and resistor network |
The experimental target was a commercially available micro-small fixed-wing UAV equipped with a complete flight control system including sensors (gyroscope, accelerometer, magnetometer, barometer, airspeed sensor), a central processing unit (CPU), and actuators (elevator, rudder, aileron servos). The UAV had a wingspan of about 1.2 m and a maximum takeoff weight of approximately 3 kg. It was programmed to fly a pre-planned route at a cruise altitude of 120 m above ground level (AGL). The test site was selected in an uninhabited area to ensure safety during potential uncontrolled descent. The UAV launched from a point approximately 15 km away from the HPM system, and its position was continuously monitored by the ground control station. When the UAV entered the designated irradiation zone, the target acquisition subsystem locked on and commanded the HPM transmitter to fire a single burst of pulses. The flight data logger onboard recorded all parameters at a rate of 50 Hz, which were later extracted after recovery of the wreckage.
Upon HPM irradiation, the UAV immediately exhibited abnormal flight behavior. The optoelectronic tracker captured a sudden change in attitude, followed by a rapid spiral descent. Approximately 10 seconds after the irradiation event, the UAV impacted the ground on a hillside. The wreckage was located using coordinates provided by the tracking system, and the onboard data storage module was retrieved intact. Figure 3 (not shown) presents the altitude versus time profile from takeoff to impact. The UAV climbed to the cruise altitude of about 120 m in 67 seconds, corresponding to a climb rate of approximately 4.4 m/s. After HPM irradiation, the altitude decreased sharply at an average rate of 20 m/s, far exceeding the normal descent rate of 2–3 m/s during landing. This indicates a non-controlled fall. The difference of 78 m between the impact point and the launch point is attributable to the topographical slope of the hillside.
To understand the attitude dynamics, we analyzed the roll angle and pitch angle time histories, plotted in Figure 4 (not reproduced). At the moment of HPM irradiation, the roll angle jumped from 13.6° to 3.9° instantaneously, then exhibited a nearly linear decrease to -14.9° over the next few seconds, followed by step-like changes to -66.3°. The total roll excursion was 79.9°, far beyond the normal control range (±30°). Similarly, the pitch angle changed abruptly from -3.8° (slightly nose-down for cruise) to 21.5° immediately after irradiation, then linearly increased to 32.3°, and eventually reached a maximum of 43.6° before slightly retreating to 37.1°. These rapid and large-amplitude changes in attitude are consistent with erroneous actuator commands from the flight control system.
Further evidence comes from the magnetic heading angle, shown in Figure 5 (not shown) alongside altitude. After HPM exposure, the heading angle began to oscillate with an approximate period of 1.25 seconds, completing about 8 full cycles during the 10-second descent. This oscillatory behavior, combined with the rapid altitude loss, confirms that the UAV was in an uncontrolled spiral descent—totally unlike a normal landing pattern. The spiral motion is a direct consequence of asymmetric elevator and rudder deflections. Figure 6 (not shown) overlays altitude with elevator and rudder deflection angles. At the instant of HPM irradiation, the elevator deflection angle jumped from its normal trim value of -5.9° to -25° and remained locked at that value for the remainder of the flight. This “jamming” of the elevator indicates that the servo was receiving a constant maximum deflection command, likely a result of CPU malfunction. The rudder deflection angle changed from 0.1° to 9.7° in a step, then increased nearly linearly to 16.9°, followed by an irregular jump to 21.5° and a subsequent slow decrease. The rudder behavior was chaotic, reflecting corrupted control signals.
To quantify the observed changes, Table 2 lists the key flight parameters at the moment of HPM irradiation and at selected subsequent times. The data clearly demonstrate the immediate and profound impact of the HEM pulse on the fixed-wing UAV’s flight dynamics.
| Time (s) | Altitude (m) | Roll (°) | Pitch (°) | Heading (°) | Elevator Defl. (°) | Rudder Defl. (°) |
|---|---|---|---|---|---|---|
| 0 (reference, before irradiation) | 120 | 13.6 | -3.8 | 180 | -5.9 | 0.1 |
| 0.02 (first sample after pulse) | 119 | 3.9 | 21.5 | 185 | -25.0 | 9.7 |
| 2.0 | 80 | -14.9 | 32.3 | 210 | -25.0 | 16.9 |
| 5.0 | 20 | -50.2 | 43.6 | 270 | -25.0 | 14.0 |
| 8.0 | -30 (below launch) | -66.3 | 37.1 | 180 | -25.0 | 21.5 |
The root cause analysis focuses on the coupling mechanisms of the HEM pulse into the UAV’s avionics. Two primary paths exist: “front-door” coupling through antennas of the GPS receiver and communication link, and “back-door” coupling through cables, apertures, and slots in the airframe. In our experiment, the GPS receiver and data link did not show signs of permanent failure—telemetry continued briefly after irradiation, and the onboard data logger recorded all parameters until impact. If the GPS had been disrupted, the UAV would have lost position hold but not necessarily suffered severe attitude upset. The data link disruption would cause communication loss but not directly control the servos. Therefore, the observed actuator jamming and chaotic deflections must be attributed to back-door coupling that directly interfered with the flight control CPU. The CPU, typically a microcontroller or a system-on-chip, is sensitive to high-frequency electromagnetic fields that induce voltages on internal traces and package leads.
The back-door coupling process can be modeled in three stages, as illustrated in Figure 7 (not shown): (1) free-space propagation of the HEM pulse from the antenna to the UAV, (2) coupling into internal cables and wiring harnesses through slots and apertures in the fuselage, and (3) further coupling into the CPU’s digital and analog circuits via the connected cables. The overall transfer function can be expressed as:
$$ V_{\text{induced}}(t) = \int_{0}^{t} \left[ H_1(\tau) * H_2(\tau) * H_3(\tau) \right] E_{\text{inc}}(t-\tau) d\tau $$
where \( E_{\text{inc}} \) is the incident electric field at the UAV location, \( H_1 \) represents the propagation path loss (dependent on distance \( R \) and arrival angle \( \theta \)), \( H_2 \) represents the cable coupling transfer function (dependent on frequency \( f \), polarization \( p \), and \( \theta \)), and \( H_3 \) represents the circuit response function (also dependent on \( f \), pulse width \( \tau_w \), and \( p \)). In practice, the nonlinear behavior of semiconductor junctions in the CPU makes precise prediction extremely difficult. However, the experimental data provide clear evidence that the induced interference was sufficient to corrupt the CPU’s instruction execution, causing it to output erroneous servo commands. The elevator servo received a constant full-down command, while the rudder servo received a varying, non-periodic command—likely the result of a register being stuck or a control loop being broken.
The UAV’s flight control system architecture is typical for small fixed-wing UAVs: a sensor suite feeds data into a CPU running a stabilization and navigation algorithm, which in turn generates pulse-width modulated (PWM) signals for the servos. Under normal operation, the elevator and rudder deflections are modulated to maintain desired pitch and yaw rates. The observed lock of the elevator at -25° (full down) suggests that the CPU’s output register for that channel was forced to the maximum value, while the rudder channel received a corrupted sequence of duty cycles. The lack of recovery after 10 seconds indicates that the CPU was not merely transiently upset but entered a persistent fault state, possibly due to latch-up or single-event functional interrupt (SEFI).
Additional evidence supporting the back-door coupling hypothesis comes from the fact that the power supply battery, sensors (gyroscope, accelerometer, barometer, magnetometer), and propulsion motor continued to function normally after the flight—the data logger recorded valid values from all sensors until impact, and the motor was still running when the wreckage was examined. This rules out bulk destruction of the UAV’s electronics. The selective vulnerability of the CPU is consistent with its high sensitivity to induced transient voltages on input/output lines and its lack of adequate filtering or shielding on the servo control lines.
Our findings underscore the importance of electromagnetic hardening of the flight control CPU for fixed-wing UAVs against HPM threats. While front-door protection (e.g., frequency rejection in antenna paths) can mitigate some risks, back-door coupling remains a formidable challenge that requires careful design of cable routing, shielding, and circuit-level hardening (e.g., voltage clamps, transient suppressors, and redundant voting logic). The experimental methodology presented here—dynamic flight testing with onboard data recording—provides a realistic assessment of HPM effectiveness that static bench tests cannot replicate. It also yields threshold data essential for weapon system engineering and vulnerability analysis.
In conclusion, we have experimentally demonstrated that narrowband HPM can disrupt a micro-small fixed-wing UAV by causing the elevator to jam at a large negative deflection and the rudder to exhibit irregular large-angle oscillations. These erroneous actuator responses originated from the back-door coupling of the HEM pulse into the flight control CPU, leading to corrupted control outputs. The UAV rapidly entered an uncontrolled spiral descent and crashed. The research emphasizes the need for improved electromagnetic protection of flight control processors in fixed-wing UAVs. Future work should focus on real-time monitoring of the CPU’s internal states during HPM exposure to better understand the disruption mechanism and develop effective countermeasures. Additionally, extending the experiments to different UAV platforms and HPM parameters (e.g., broadband pulses, different polarizations) will further validate the findings and support the development of robust HPM counter-UAV systems.