Chandrasekhar-type filter for a wide-sense stationary signal from uncertain observations using covariance information (Q1826647)

The authors consider the problem of estimating a scalar signal process \(z(k)\) from an observation process \(y(k)\) related with the signal by the observation equation \[ y(k)= u(k)z(k)+ v(k) \] perturbed by a multiplicative noise \(u(k)\) and by an additive noise \(v(k)\). It is assumed that \(u(k)\) is a sequence of independent Bernoulli random variables with \(P[u(k)= 1]= p\), \(P[u(k)= 0] = 1- p\), i.e., \(u(k)\) determines the presence or absence of the signal in the observations. The authors construct a Chandrasekhar-type recursive algorithm for the linear filtering estimator and compare it with a Riccati-type algorithm.