Sensor Network Data Compression Library

"The BMDO is tasked with the integration of leading edge and emerging technologies into an effective acquisition policy for missile defense. A missile defense system must potentially track thousands of objects with many networked sensors and dataprocessors. To support this need, the BMDO requires advanced network-centric data compression algorithms for networks of sensors. To address this demand, Broadata Communications, Inc. (BCI) proposes to develop a novel Sensor Network Data CompressionLibrary (SNDCL). In particular, the SNDCL will be based on innovative data compression algorithms developed in 2001 by Caltech researchers Effros and Zhao. They have shown 20-30% additional compression performance on network sensors, compared to encodingeach sensor signal independently. BCI's SNDCL will extend the code design methods by incorporating BCI's proprietary Nearest Neighbor partitioning scheme, which can speed up the design and optimization of the compression codes by embedding them into a setof software libraries that can be used for compressed data transmission from a large number of network sensors. The BCI SNDCL will also address the practical issues of determining the level of correlation of the data sources, and monitoring the performanceof the codes over time to determine whether any codebook adjustments are needed. In addition to BMDO's sensor network transmission applications, this technology is also applicable to many milita