Distributed detection with quantization based on Neyman-Pearson criterion

Decentralized fusion detection has been a hot topic in multisensor network research for decades, among which decentralized quantitative fusion detection is a focus because of a lighter communication burden brought by quantization. In this paper we propose a quantization strategy for distributed detection according to the Neyman-Pearson criterion. The detection probability given certain false alarm rate is maximized in our quantization criterion when searching for local quantization thresholds and fusion rules. Numerical results show that this method outperforms the Bhattacharyya distance based quantization method in detection performance.