Optimal phase parameter estimation of random amplitude linear FM signals using cyclic moments

This paper considers the problem of estimating the phase parameters of a linear FM signal which is modulated by a random process and is embedded in additive noise. In particular, we consider the use of cyclic moments and derive variance expressions for the phase parameter estimates for all values of the lag parameter of the second order cyclic moment, τ. It is seen that the accuracy of the phase parameter estimates depends greatly on τ. This allows the definition of an optimal value of τ, in the sense that it minimises the phase parameter estimation variance