An Analysis of Radar in Terms of Information Theory and Physical Entropy.

摘要

A radar is basically a measuring instrument for extracting information from a received signal voltage. The amount of information that can be extracted is ultimately limited by the noise background. The concept of measurable information has been quantitatively developed in the literature and provides a means by which substantial insight may be acquired into the generalized nature of optimum radar processing. The presented work exhibits an interpretation of radar information consistent with intuitive views of radar processing as well as with formal definitions of information theory. The approach has led to the following specific results. A unified description of the mechanism by which both detection capability and parameter estimation are optimized simultaneously. A two-dimensional matched filter is described in which the ideal radar processing is seen to consist of a twodimensional involution in which the signal's input energy area in the time-frequency plane is compressed into an interior area. An equivalence between the amplitude detection threshold commonly employed in radar and the energy threshold employed in physics for setting the reliability of a measurement. The development of a new detection procedure based upon a defined detection information and the employment of the NeymanPearson procedure in the threshold-crossing rate domain rather than in the standard amplitude domain. (Author)