Optimized Bernoulli Trial Technique for M Out of N Binary Integration of Radar Signals

• Published in: 2008 IEEE Aerospace Conference pp. 1 - 8
• Main Authors: Golestani, Y., Mallean, G.
• Format: Conference Proceeding Journal Article
• Language: English
• Published: IEEE 2008
• Subjects: Clutter; Computational modeling; Detectors; Distributed computing; Event detection; Frequency; Lagrangian functions; Polynomials; Radar detection; Signal to noise ratio
• ISBN: 1424414873; 9781424414871
• ISSN: 1095-323X; 2996-2358
• DOI: 10.1109/AERO.2008.4526428

Binary integration is used in radar systems where M successful events out of N trials represent the detection of a target where multiple pulse-repetition frequencies (PRFs) are used. If the probabilities of the single events are equal, the probability of M events occurring out of N can be represented by a binomial distribution. If, however, switching the PRF at each event causes differing clutter to interfere with the signal, the event probabilities are not binomially distributed but are represented by a more complicated polynomial. An optimization procedure is presented that minimizes the radar resources needed for the M out of N binary integration of detector outputs with unequal signal-to-interference noise ratios (SINRs). The amount of coherent integration for each PRF output is optimally adjusted to yield a minimum total integration time over the batch of PRFs. The optimized integration time is computed for a number of different sequences of detector output SINRs and compared to the results from a binomial distribution.