Narrowband Communication Signal Source Precision Tdoa Estimation

This dissertation focuses on the TDOA (time difference of arrival) estimation problem which has various and important applications in the fields of communication, radar, sonar and so on. Corresponding to time delay estimation of narrowband signals, a lot of research has been done in this paper. In the considered narrowband signals scenario, a practical system model of time delay estimation is derived. In the following, the errors caused by receiver system imperfections and the frequency-domain TDOA estimation method are developed and analyzed in detail. Several theoretically and practically valued algorithms are proposed and derived. Computer simulations and their results are given to verify the efficiency of these algorithms. The main contribution of this dissertation can be concluded as follows:(1) Firstly, ideal system model of TDOA estimation is analyzed, and time/frequency-domain estimation methods are summarized. Then, it is shown how imperfections of the receiver system model will affect the considered correlation-based time delay estimations, and the system errors include the limited acquisition interval, sampling frequency reference error and oscillator frequency error. At last, the adjusting algorithms for system model error are proposed.(2) Based on narrowband signals, a TDOA estimation model is proposed. In the following, the advantages of frequency-domain estimators and estimation algorithm for narrowband signals are analyzed in detail. For narrowband signals, significant time-saving can be achieved in FFT since the number of output samples required to be computed is small. It can be used in frequency-domain estimators and computation can be efficiently reduced.(3) A novel acoustic source localization approach utilizing multi-path information in a reverberant environment is proposed. Making use of environment information, the TDOA of multi-path can be identified. By constructing virtual receivers, the TDOA caused by multi-path can be used for localization, and a closed-form solution to the acoustic source location can be obtained by only two receivers.