This paper presents TriSweep, a simulation framework for four-drone swarm electromagnetic side-channel analysis of masked embedded cryptographic devices at standoff distances of 0.25–1.5 m. The following subsections establish the threat landscape that motivates the work, the technical challenges it addresses, and the framework's design and key results.
Contributions
A four-drone EM-SCA platform design: Drone A (Anchor), Drone B (Mask Probe), Drone C (Cipher Probe), and Drone D (Accumulator). All results use real ASCAD datasets and a simulated standoff noise model; no hardware has been fabricated.
A swarm consensus protocol for EM-optimal repositioning via distributed Fisher information maximization, 200 ms cycles.
A two-stage inter-drone clock synchronization protocol targeting 10 ns jitter.
Profiling-trace cross-correlation alignment: key rank reduces from 89 to 21 on the 100-sample-jitter ASCAD dataset.
Drone D second-order combining via a centered product of Drones B and C streams, reducing simulated key rank to a value close to 20.
First published aerial, multi-node, autonomous EM-SCA framework with hardware second-order mask cancellation design.
Related Work
This section surveys four bodies of literature that directly motivate TriSweep: classical EM side-channel attacks, second-order analysis, deep learning-based profiling, and mobile threat models. The survey closes with a structured comparison positioning TriSweep against recent work.
System and Threat Model
This section establishes the physical signal model underlying TriSweep and the threat assumptions under which the framework is evaluated. The signal model calibrates simulated standoff SNR to the real ASCAD dataset; the threat model defines adversary capabilities and operational constraints.
Physical Signal Model
EM power received by a drone at a distance r is modeled using free-space path loss. Signal-to-noise ratio (SNR) with N coherent receivers scales as N squared, providing the theoretical foundation for multi-drone combining gains.
EM Target Detection and Localization
Each drone independently scans and flags the peak-power frequency; the ground station resolves a three-way consensus; time difference of arrival (TDOA) from synchronized GPSDO timestamps triangulates the target. The localized position seeds the Fisher information optimization for optimal repositioning.
Second-Order Combining
Drone D computes the centered product of Drones B and C, canceling the mask m without knowledge of its value. The innovation over prior work is physical spatial separation: the two leakage windows are captured by dedicated drones rather than extracted algorithmically from one trace.
The combined log-likelihood is computed from the centered product of the two probe streams.
Profiling-Trace Alignment
Before template construction, profiling traces for each desync dataset are aligned via cross-correlation against the ASCAD_Masked mean, with alignment window sizes of 50 and 100 samples for the two desync variants.
Results
This section presents simulation results across five experimental campaigns: EM leakage characterization, SNR vs. standoff distance, four-drone ablation, statistical validation, distance sweep, cross-dataset combining, desync validation, and CNN profiling attack comparison. All key-rank values are simulated using real ASCAD traces with the physics-based noise model.
Key-Rank vs. Distance: Full Four-Drone System
Four-drone (A+B+C+D) performance at five standoff distances. ASCAD_Masked: rank 20 at 0.25 m to 25 at 1.5 m, confirming Drone D second-order compensates SNR loss. ASCAD_Desync100 shows inverted distance ordering because per-trace attack-phase alignment is not yet applied.
Key-Rank vs. Trace Count: Three-Drone CNN Baseline
The three-drone single-channel CNN baseline shows at 0.25 m rank reaches 24 within 10,000 traces; rank degrades monotonically with distance confirming the noise model.
Coherent Combining Gain
At 1.0 m standoff: 1-drone yields rank 49; 2-drone yields rank 26; 3-drone yields rank 19 — consistent with theoretical N-squared SNR gain.
Cross-Dataset Drone Combining
Using heterogeneous templates (Drone A from ASCAD_Masked, Drone B from ASCAD_Desync50, Drone C from synthetic fallback), the four-drone result (rank 92) is substantially worse than the homogeneous case (rank 20), confirming matched profiling templates are required for effective second-order cancellation.
Multi-Dataset Validation and Desync Results
Key rank across all real ASCAD datasets after profiling-trace alignment. Four-drone second-order combining on desync variants requires per-trace attack-phase alignment; those results are reserved for future work.
The primary result is rank 20 with four drones on ASCAD_Masked. Alignment reduces Desync100 two-drone rank from 89 (unaligned) to 21, demonstrating that cross-correlation alignment compensates 100-sample jitter. The B-C product degrades on desync variants because per-trace attack-phase alignment was not applied.
Discussion
This section examines the TriSweep framework from four perspectives: design trade-offs inherent in the four-drone architecture, operational and environmental factors that will affect physical deployment, two fundamental limitations that bound the current simulation-only evaluation, and future research directions.
Design Trade-Offs
Drone count vs. combining complexity. Adding collector drones increases coherent SNR gain linearly but adds inter-drone synchronization burden and Wi-Fi mesh latency. Four drones balance the N-squared gain from three coherent collectors against the coordination overhead of Drone D; beyond four drones, synchronization latency is projected to exceed the 200 ms repositioning budget under the current 50 Hz heartbeat protocol.
Standoff distance vs. SNR budget. Each doubling of standoff distance costs 6 dB SNR under the free-space path-loss model; three-drone combining recovers only 9.5 dB. The net SNR deficit therefore grows with distance, making 1.5 m the practical ceiling under the current architecture — beyond this point, additional traces are the only recovery mechanism.
Operational and Environmental Factors
Physical deployment introduces a class of factors absent from the current Gaussian noise model, each of which is expected to degrade effective SNR relative to the simulation predictions.
Propeller EMI and vibration. BLDC motor controllers generate broadband RF emissions across the 1–500 MHz capture band. Vibration-induced mechanical jitter couples into the IQ sample stream as additional timing noise beyond the GPSDO and cross-correlation budget. Both effects are unmodeled and represent the largest expected gap between simulated and physical results.
Wind and hover instability. Even a 5 km/h crosswind induces centimeter-scale lateral displacement at the hover distances of interest. Displacement from the optimal standoff position degrades the SNR according to the physical signal model; periodic repositioning mitigates slow drift but cannot compensate for rapid gusts within the 200 ms update cycle.
Drone positioning and swarm geometry. The Fisher information maximization assumes accurate relative positioning from the visual-inertial odometry (VIO) system. At 0.25 m, a 2 cm positioning error represents an 8% standoff uncertainty, propagating directly into the SNR model. Outdoor GPS multi-path and magnetic interference from the target's power supply can degrade VIO accuracy below the sub-centimeter lab specification.
RF interference and co-channel leakage. Urban RF environments introduce co-channel interference across the capture band that is correlated neither with the AES computation nor with the swarm's Wi-Fi control channel, but that can raise the effective noise floor above the Gaussian model calibrated in a shielded lab. Notch filtering and adaptive gain control on the low-noise amplifier (LNA) are expected mitigations.
Detection and operational security. A drone hovering at 0.25 m is visually detectable at close range and is audible from the propeller noise. Practical deployment would require elevated standoff (greater than 1 m), reduced propeller RPM, and timing during low-activity periods, all of which trade against effective SNR and trace collection rate.
Limitations
TriSweep is a simulation framework bounded by two fundamental limitations:
No physical hardware. All results use a free-space Gaussian noise model calibrated to the ASCAD dataset; no drone has been built or flown. The operational factors are qualitatively identified but not quantified; physical experiments are required to bound the real-world performance gap.
Second-order combining requires matched, aligned data. The centered product degrades when profiling and attack traces are not co-aligned; the two-channel CNN additionally over-fits at the current training scale on clean masked data. Per-trace attack-phase alignment and cross-validated CNN training are required before Drone D combining can be considered experimentally validated.
Frequently Asked Questions
What is TriSweep? TriSweep is a simulation framework for a four-drone swarm that performs electromagnetic side-channel analysis on masked cryptographic devices at standoff distances from 0.25 to 1.5 meters.
How does the four-drone system cancel the cryptographic mask? Drone B captures the mask leakage, Drone C captures the masked ciphertext leakage, and Drone D computes their centered product to cancel the mask without knowing its value.
What key rank does TriSweep achieve on the ASCAD_Masked dataset? The four-drone system achieves a key rank of 20 on ASCAD_Masked at 0.25 m standoff distance using 10,000 profiling traces.
Has TriSweep been tested with physical drones? No, TriSweep is currently a simulation-only framework. All results use real ASCAD datasets with a physics-based Gaussian noise model; no drone hardware has been fabricated.