A survey on the maximum entropy method and parameter spectral estimation

D. M. Lin, E. K. Wong

    Research output: Contribution to journalReview article


    This paper presents a survey on the maximum entropy method and parameter spectral estimation. High-quality spectral estimation of a random signal can be defined as the process of abstracting as much frequency information as possible from measurement data. Information theory provides a theoretical foundation. In order to make use of some a prior knowledge, on needs to choose a proper model in either the time domain or the frequency domain. Some real world disturbances, such as noise interference and signal fluctuation, greatly affect the resulting estimates. In this paper, the history of the development of many algorithms and methods in maximum entropy spectral estimation is reviewed. Most of the existing spectral estimation methods are systematized in terms of signal models, with emphasis on the eigenanalysis method. Also, some considerations for software and hardware performance, such as the information bank and the lattice filter, are discussed.

    Original languageEnglish (US)
    Pages (from-to)41-135
    Number of pages95
    JournalPhysics Reports
    Issue number2
    StatePublished - Sep 1990


    ASJC Scopus subject areas

    • Physics and Astronomy(all)

    Cite this