Networked ISAC: Rate-Distortion Analysis for Efficient Map Compression in Cooperative Sensing

This paper explores data compression’s key role in cooperative sensing within integrated sensing and communication (ISAC) networks, where range-angle maps generated at each base station (BS) are shared with a fusion center (FC). Efficient compression schemes minimize network overhead while ensuring accurate target detection and localization. Motivated by this challenge, we propose three novel compression approaches tailored for a network of multiple-input multiple-output (MIMO)-orthogonal frequency division multiplexing (OFDM)-based mono-static sensors: i) excision filtering (EF) for map sifting, ii) principal component analysis (PCA) for dimensionality reduction, and iii) quantization for efficient encoding. To localize both point-like and extended targets from fused range-angle maps, we exploit the density-based spatial clustering of applications with noise (DBSCAN) clustering algorithm. Given that DBSCAN requires careful tuning of its clustering parameters, we introduce an artificial intelligence (AI)-driven method to optimize these settings dynamically. A comprehensive rate-distortion analysis evaluates the network’s localization performance under varying compression levels. Key metrics—including bit rate, generalized optimal sub-pattern assignment (GOSPA) error, missed detection rate, and false alarm rate—provide a holistic assessment that balances localization accuracy with network overhead.