A DPTF algorithm for the time-delay estimation in the reflected environment

Abstract

In passive localization, time-delay estimation (TDE) is an important technology. A common super-resolution approach for TDE is the generalized phase spectrum (GPS). However, because of the reflection in the environment, the phase of the cross-spectrum will fluctuate dramatically, rendering the GPS approach useless. This research work presents a Dual-Path Transfer Function (DPTF) model for the phase of cross-spectrum in the reflected environment to overcome the issue. The DPTF model is a four-dimensional optimization problem with a well-defined objective function. In comparison to recent methods, it obtains a higher resolution in the simulation. After that, a Fast DPTF algorithm is proposed. In the Fast DPTF, the gradient is used to find the initial parameters, and subsequently, the more accurate values are obtained by iteratively decreasing the gradient fluctuating range. The fast method can reduce the computational time to one-fifth of the original one. The method's robustness in the reflected environment is demonstrated in a dark-chamber experiment.