High-Fidelity Adaptive Compression for Cognitive Spectral Monitoring

A key aspect of the cognitive radio paradigm for maximizing spectral usage is spectral monitoring. However, continuous monitoring of a wide frequency band with high sensitivity and spectral resolution can generate unmanageable amounts of data. By applying an adaptive compression algorithm to measured ultrawideband spectral data we show that it is possible to obtain spectral data compression ratios of 10, 000 or more, allowing the datarate to be reduced below 1 Mbps, sufficient for streaming to disk, broadcasting over a wireless link, or performing further processing. A key aspect of the approach is the fidelity with which the spectral signals are preserved. We show compression results applied to measured data, and discuss an architecture for implementing the data compression algorithms in real-time.