Comparison of particle filtering and extended Kalman filtering for acoustic-based tracking of low-flying aircraft

Abstract

This presentation summarizes a comparative study of particle filtering and extended Kalman filtering applied to acoustic-based tracking of high-speed, low-flying aircraft. A distributed network of small acoustic arrays, each reporting real-time azimuthal bearings and elevation angles to acoustic sources to a central processing unit, are synthesized into real-time three-dimensional tracks. The primary challenge of the problem is the significant propagation delay between the source and the receivers. Both tracking methods are applied to simulated and field test data and reveal nearly equivalent tracking performance in all cases as along as a sufficient number of particles are utilized. Computational requirements of the extended Kalman filter are significantly less than the particle filters.