Joint filtering of signal parameters having continuous and discrete states

Authors

  • G. L. Dedkov
  • V. M. Terent’ev

DOI:

https://doi.org/10.3103/S073527271990070159

Abstract

In works which consider the synthesis of algorithms for joint estimation of discrete and continuous signal- parameter states, decisions with regard to the discrete parameters are frequently adopted via threshold comparison of likelihood ratios conditioned by estimates of the continuous parameters. As a result, due to difficulties in obtaining separate estimates of the filtering performance for discrete and continuous parameters, the analysis and synthesis of complex radio-engineering aggregates is complicated and the possibility of unifying the implementation base of the mentioned algorithms is reduced. It is possible to overcome the indicated shortcomings while using a single device for representing and estimating discrete and continuous parameter states. The basis for this can be found in [1–3] where a description is given of discrete and continuous parameters using unified models in the form of stochastic difference (differential) equations.

References

SHIRYAEV, A.I. Stochastic nonlinear-filtering equations for step-wise Markovian processes. Problemy Peredachi Informatsii, vol. 11, no. 3, pp. 3-17, 1986.

SEGALL, A. Stochastic processes in estimation theory. IEEE Trans. Inf. Theory, vol. 22, no. 3, pp. 275-286, 1976. DOI: https://doi.org/10.1109/TIT.1976.1055559.

TIKHONOV, V.I. Nonlinear Transformations of Random Processes [in Russian].Moscow: Radio i Svyaz, 1986.

ROZANOV, Y.A. Theory of Renewing Processes [in Russian]. Moscow: Nauka, 1974.

Published

1990-07-15

Issue

Section

Brief Communications