Improved blind separation algorithm for overlapping secondary surveillance radar replies

The secondary surveillance radar (SSR) is a transponder system used in air-traffic control. With the increase in air-traffic, replies from airplanes may overlap in time at ground station receivers, which results in loss of all replies for classic receivers. Blind source separation algorithms were proposed to separate such a mixture by the properties of SSR replies. Two known algebraic algorithms, the Manchester decoding algorithm (MDA) and the multishift zero-constant modulus algorithm, have the best performance but they still have performance degradation in cases with either small overlapping ratios or equal residual carrier frequencies. In this paper, we propose a modified subspace intersection method based on signed URV decompositions to preprocess the received data matrix for MDA. The proposed algorithm works on three successive time slots with the target time slot locating in the center and significantly improves the performance of MDA in all cases. Especially, in cases with small overlapping ratios, it touches the upper bound of the performance it can achieve. It shows the most stable performance compared with previous algorithms.