Wideband Compression for Radiolocation

Abstract

The radiolocation of arbitrary transmitters is often accomplished by combining pre-detection radio frequency (RF) snapshots and making time-difference-of-arrival (TDOA) and frequency-difference-of-arrival (FDOA) measurements. This requires transporting these snapshots from two or more geographically separated sensors to a common processor where the combined measurements can be made. The available communications bandwidth between these sensor nodes can be a limiting factor in overall system cost and performance. This research presents data compression techniques designed for RF snapshots: lossless methods suitable for any wideband RF sensing application, and lossy methods designed to maximize the quality of TDOA/FDOA measurements and therefore radiolocation accuracy for a given compression ratio. The lossless techniques leverage time and frequency-domain sparseness to transform the snapshots into lower-entropy representations. The lossy techniques leverage the non-uniform contributions of the RF components to TDOA/FDOA measurement quality to emphasize the most important components. In support of this, the general precision of TDOA/FDOA measurements is revisited for low-SNR intercepts. All techniques are demonstrated on actual wideband RF intercepts and simulated examples and their performance is compared to readily available compression tools and relevant published algorithms.