Bayesian probability theory applied to the problem of radar target discrimination

The task of discriminating among a set of N known targets based on their radar returns is viewed as a problem of information processing, calling for a full application of probability theory. Two distinct problem areas are investigated. First, Bayesian probability theory is used to derive an expression for an enhanced discrimination waveform which, in the two-target case, maximizes the log odds in favor of one target over the other. Numerical results are provided which show that best discrimination, in the simple two-target case, occurs when the incident waveform has its energy concentrated near the frequency where the difference in the impulse response of the two targets reaches a maximum. Second, probability theory is used to discriminate among a set of targets based on their high-range-resolution radar returns. Example calculations show that for the four-target case the Bayesian algorithm identifies the unknown target correctly greater than 90% of the time for signal-to-noise ratios as low as 2 (3 dB).<>