Fast parameter estimation algorithms for linear FM signals

This paper describes computationally efficient algorithms for estimating the parameters of a complex linear FM signal in white Gaussian noise. Algorithm I deals with the high signal to noise ratio (SNR) case (above -4dB for a s y m p totically long signals). Algorithms I 1 and 111 are for the low SNR range. Algorithm 11 is based on the maximum likelihood (ML) approach, but takes advantage of the numerical conditioning of the parameter estimation problem to yield a computationally efficient algorithm. Algorithm 111 also has its roots in the ML procedure, but uses a suboptimal implementation to avert the difficult 2D search procedure. Simulations are provided to illustrate performance.