Event cameras provide high-temporal-resolution visual sensing that is well suited for observing fast-moving aerial objects; however, their use for drone trajectory prediction remains limited. This work introduces an event-only drone forecasting method that exploits propeller-induced motion cues. Propeller rotational speed are extracted directly from raw event data and fused within an RPM-aware Kalman filtering framework. Evaluations on the FRED dataset show that the proposed method outperforms learning-based approaches and vanilla kalman filter in terms of average distance error and final distance error at 0.4s and 0.8s forecasting horizons. The results demonstrate robust and accurate short- and medium-horizon trajectory forecasting without reliance on RGB imagery or training data.
Figures & tables
Fig. 1 : Overview of the proposed trajectory forecasting up to 0.4 s into the future. Blue denotes the current drone position and past trajectory. Future ground-truth positions are shown in green, and predictions in orange.
Fig. 2 : Overview of the proposed event-based trajectory forecasting framework.
Fig. 3 : Qualitative trajectory forecasting results under challenging conditions. The current ground-truth position and past trajectory are shown in blue; future ground-truth positions are shown in green; predicted future positions are shown in orange. Results are illustrated across highly dynamic motion, night-time, indoor, and rain scenarios.
Method
Input
ADE (px) ↓
FDE (px) ↓
RGB
Event
0.4s
0.8s
0.4s
0.8s
LSTM
✗
✗
171.9
353.3
53.04
96.44
Transformer
✗
✗
124.8
292.2
46.74
87.68
CNN + Transformer
✓
✗
124.0
280.9
45.92
83.86
CNN + Transformer
✗
✓
151.7
337.0
58.06
100.6
CNN + Transformer
✓
✓
121.3
281.5
45.23
85.78
TABLE I : Trajectory forecasting results on the FRED dataset. ADE and FDE are reported in pixels.
Fig. 4 : Distribution of forecasting errors at 0.4 s and 0.8 s horizons over 54 test cases. Top: ADE, Bottom: FDE. The box represents the interquartile range (25th–75th percentile), whiskers denote the 5th–95th percentile, and the orange colored line indicates the median.
We study short- and mid-horizon UAV bounding-box forecasting on the FRED event-camera dataset. We use a constant-velocity Kalman filter over a full center-size box state as a strong physical baseline, and train a residual model to predict acceleration-like corrections from recent box history, filtered state features, and local event representations. This simple residual formulation consistently improves over the Kalman baseline, with event-conditioned models giving the strongest results among the evaluated methods. We further show that part of the residual target is predictable from anchor position and velocity alone, indicating that canonical FRED results can reflect both visual evidence and dataset-specific motion priors. To analyze this effect, we introduce decorrelated subsets as a diagnostic stress test, showing that event-conditioned residual models retain useful predictive signal even when measured position- and velocity-based shortcuts are weakened.
Per Nyblom, Hannes Ovrén, David Gustafsson
Swedish Defence Research Agency (FOI), Linköping, Sweden
Safety-critical perception for unmanned aerial vehicles and rotating machinery requires microsecond-latency tracking of fast, periodic motion under egomotion and strong distractors. Frame-based and event-based trackers drift or break on propellers because periodic signatures violate their smooth-motion assumptions. We tackle this gap with HelixTrack, a fully event-driven method that jointly tracks propeller-like objects and estimates their rotations per minute (RPM). Incoming events are back-warped from the image plane into the rotor plane via a homography estimated on the fly. A Kalman Filter maintains instantaneous estimates of phase. Batched iterative updates refine the object pose by coupling phase residuals to geometry. To our knowledge, no public dataset targets joint tracking and RPM estimation of propeller-like objects. We therefore introduce the Timestamped Quadcopter with Egomotion (TQE) dataset with 13 high-resolution event sequences, containing 52 rotating objects in total, captured at distances of 2 m / 4 m, with increasing egomotion and microsecond RPM ground truth. On TQE, HelixTrack processes full-rate events (approx. 11.8x real time) faster than real time and microsecond latency. It consistently outperforms per-event and aggregation-based baselines adapted for RPM estimation.
Radim Spetlik, Michal Pliska, Vojtěch Vrba +1
Faculty of Electrical Engineering, Czech Technical University in Prague
Detecting fast-moving objects, such as unmanned aerial vehicle (UAV), from event camera data is challenging due to the sparse, asynchronous nature of the input. Traditional Discrete Fourier Transforms (DFT) are effective at identifying periodic signals, such as spinning rotors, but they assume uniformly sampled data, which event cameras do not provide. We propose a novel per-pixel temporal analysis framework using the Non-uniform Discrete Fourier Transform (NDFT), which we call Drone Detection via Harmonic Fingerprinting (DDHF). Our method uses purely analytical techniques that identify the frequency signature of drone rotors, as characterized by frequency combs in their power spectra, enabling a tunable and generalizable algorithm that achieves accurate real-time localization of UAV. We compare against a YOLO detector under equivalent conditions, demonstrating improvement in accuracy and latency across a difficult array of drone speeds, distances, and scenarios. DDHF achieves an average localization F1 score of 90.89% and average latency of 2.39ms per frame, while YOLO achieves an F1 score of 66.74% and requires 12.40ms per frame. Through utilization of purely analytic techniques, DDHF is quickly tuned on small data, easily interpretable, and achieves competitive accuracies and latencies to deep learning alternatives.