DOA estimation based on cross terms of spatial time-frequency distribution matrices

This paper introduces the theoretical basis of spatial time-frequency distributions in direction-of-arrival (DOA) estimation, and then proposes a new method through analysis to non-stationary signals´ time-frequency distributions and extraction of their cross terms. The work done by the previous researchers almost based on paying attention to the auto terms information and suppressing of cross terms, and this paper examines the signal cross terms from a new perspective. Study found that cross terms between signals do contain rich information, and using cross terms information alone can calculate out the direction of the source signals. In particular, when the strength between the signals is different, the auto terms information of the weak signal have a certain possibility can not be detected. Therefore, extraction of cross terms information to estimate the DOA of signals is much easier. Simulation experiments verified these characteristics.