On some unresolved issues in finding optimum distributed detection schemes

Optimum distributed detection under the Neyman-Pearson (NP) criterion is considered for a general case with possibly dependent observations from sensor to sensor. The focus is on the parallel architecture. New necessary conditions are presented that relate the threshold used in the NP-optimum fusion rule to those used in the NP-optimum sensor rules. These results clearly illustrate that the necessary conditions for NP optimality have exactly the same form as those for Bayes optimality. Based on these conditions, a new algorithm for finding NP optimum distributed detection schemes is developed. The algorithm allows randomization at the fusion center, which we show is generally needed to achieve optimality. The algorithm allows one to attempt to optimize the fusion rule along with the sensor rules or to find the best schemes among those using each of a set of fixed possible fusion rules.

[1]  D. Warren,et al.  Optimum quantization for detector fusion: some proofs, examples, and pathology , 1999 .

[2]  Krishna R. Pattipati,et al.  An algorithm for determining the decision thresholds in a distributed detection problem , 1991, IEEE Trans. Syst. Man Cybern..

[3]  Ramanarayanan Viswanathan,et al.  Optimal Decision Fusion in Multiple Sensor Systems , 1987, IEEE Transactions on Aerospace and Electronic Systems.

[4]  M. Barkat,et al.  Decentralized CFAR signal detection , 1989 .

[5]  Amy R. Reibman,et al.  Optimal Detection and Performance of Distributed Sensor Systems , 1987 .

[6]  Po-Ning Chen,et al.  Likelihood ratio partitions for distributed signal detection in correlated Gaussian noise , 1995, Proceedings of 1995 IEEE International Symposium on Information Theory.

[7]  H. Vincent Poor,et al.  An Introduction to Signal Detection and Estimation , 1994, Springer Texts in Electrical Engineering.

[8]  Paul B. Kantor,et al.  Counterexamples in distributed detection , 1992, IEEE Trans. Inf. Theory.

[9]  Pramod K. Varshney,et al.  Distributed detection with multiple sensors I. Fundamentals , 1997, Proc. IEEE.

[10]  Rick S. Blum.  Necessary conditions for optimum distributed sensor detectors under the Neyman-Pearson criterion , 1996, IEEE Trans. Inf. Theory.

[11]  E. Lehmann.  Testing Statistical Hypotheses , 1960 .

[12]  Rick S. Blum,et al.  Unexpected properties and optimum-distributed sensor detectors for dependent observation cases , 2000, IEEE Trans. Autom. Control..

[13]  Jerzy Neyman,et al.  The testing of statistical hypotheses in relation to probabilities a priori , 1933, Mathematical Proceedings of the Cambridge Philosophical Society.

[14]  Peter Willett,et al.  The suboptimality of randomized tests in distributed and quantized detection systems , 1992, IEEE Trans. Inf. Theory.

[15]  E. S. Pearson,et al.  On the Problem of the Most Efficient Tests of Statistical Hypotheses , 1933 .

[16]  Rick S. Blum,et al.  Distributed detection with multiple sensors I. Advanced topics , 1997, Proc. IEEE.

[17]  Rick S. Blum,et al.  The good, bad and ugly: distributed detection of a known signal in dependent Gaussian noise , 2000, IEEE Trans. Signal Process..

[18]  Ramanarayanan Viswanathan,et al.  Optimal distributed decision fusion , 1989 .

[19]  Yong In Han,et al.  Randomized fusion rules can be optimal in distributed Neyman-Pearson detectors , 1997, IEEE Trans. Inf. Theory.