Bayes Information-theoretic Radar Waveform Design and Delay-Doppler Resolution for Extended Targets

Abstract

In this paper, we consider the problem of information-theoretic waveform design for active sensing systems such as radar for extended targets. Contrary to the popular formulation of the problem in the estimation-theoretic context, we are rather interested in a Bayes decision theoretic approach where a target present in the environment belongs to two or more classes whose priors are known. Optimal information theory based transmit waveforms are designed by maximizing mutual information (MI) between the received signal and the target impulse response, resulting in a novel iterative design equation. We also derive signal to noise ratio (SNR) maximization based waveforms. In an effort to quantize the benefits of such a design approach, the delay-Doppler ambiguity function of information-theoretic waveforms are presented and is compared with Barker codes of similar time-bandwidth product. It is found that the ambiguity function of information-theoretic waveforms has very sharp main lobe in general and excellent time autocorrelation properties in particular.