Polynomial Fitting and Interpolation Method in TDOA Estimation of Sensors Network

Abstract

Time difference of arrival (TDOA) estimation is the key part of TDOA positioning technology in sensors network, whose result will directly affect the positioning accuracy of the signal source. Aiming at the problem that the accuracy of TDOA estimation is greatly limited by the system sampling frequency, this article proposes an improved TDOA estimation method combining polynomial fitting and interpolation. Based on the classical generalized cross correlation (GCC) method, the peak location of the correlation function is adjusted slightly by polynomial fitting. Furthermore, linear interpolation is used to process the original signal to improve the time resolution between the samples, so as to improve the system sampling frequency indirectly and promote the estimation accuracy to the subsampling level. At the end of the article, the proposed method is validated by simulation and some measured data from a certain place in Lanzhou. The results show that the polynomial fitting method can precisely determine the peak location of the GCC correlation function, especially after interpolating 2–4 points in the original signal, the overall TDOA estimation accuracy can be improved by 0.2%–3.6%, which may provide good reference information for practical engineering application of TDOA positioning technology.