Event-Only Drone Trajectory Forecasting with RPM-Modulated Kalman Filtering

TL;DR

This paper presents an event-only drone trajectory forecasting method using RPM-modulated Kalman filtering, leveraging event camera data to predict fast-moving drone paths accurately without training data.

Contribution

It introduces a novel RPM-aware Kalman filter that directly extracts propeller rotational speeds from event data for improved drone trajectory prediction.

Findings

Outperforms learning-based methods in accuracy.

Achieves robust short- and medium-horizon forecasts.

Operates without RGB imagery or training data.

Abstract

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.