Improving detection of linear FM signals in noise using adaptive lattice filters

The probability density function of the first reflection coefficients of adaptive lattice filters is experimentally characterised. Specifically, for small significance levels, the gradient noise at the adaptive coefficients can be modelled as a coloured Gaussian noise process. These coefficients are then used to detect linear FM signals in white Gaussian complex noise