Joint time delay and Doppler stretch estimation using wideband signals

The authors discuss the application of the wavelet transform. The theoretical performance is derived. The joint maximum likelihood estimation is presented and the Cramer-Rao lower bound (CRLB) obtained. The joint estimation using the wavelet transform is shown to be more robust, and under high signal to noise ratio it is unbiased and the estimation variance is close to the CRLB for a large variety of signals. All these results are verified by computer simulations. It is demonstrated that the estimation accuracy depends on the signal structure, and the optimum signal structure is strongly related to the form of its wavelet transform.<<ETX>>