Real time processing of networked passive coherent location radar system

Abstract

A Passive Coherent Location (PCL) Radar system, consisting of spatially distributed transmitters and receivers is currently being integrated at the University of Cape Town (UCT). The paper investigates the feasibility of real-time processing of PCL system signals using Graphic Processing Units (GPUs), specifically a study of two distinct clutter cancellation algorithms: ECA (Extensive Cancellation Algorithm) and NLMS (Normalised Least Mean Square). Clutter cancellation is the most computationally demanding part of PCL signal processing. This investigation compares the processing speed-up achieved by GPU over CPU implementations, with very encouraging results.