Antenna array likelihood modulation classifier for BPSK and QPSK signals

Automatical classification of the modulation type of a received signal is an indispensable first step in many military communication systems. Typical military wireless communication environments have low signal-to-noise ratio (SNR). As a result, the classification accuracy is low or a long observation time is required to ensure an acceptable accuracy. We develop a modulation classifier using a uniform linear array to improve classification accuracy of BPSK and QPSK signals at low SNR. The proposed classification method is based on the generalized likelihood ratio test where an additive white Gaussian noise environment is assumed. The likelihood functions depend on the unknown phase shift between received signals due to the spatial separation of the antenna elements. Computationally efficient two-step processing is proposed to estimate the unknown phase shift. It can significantly reduce the computation compared with a direct estimation method, which applies exhaustive numerical searching on the phase shift because it does not have any closed-form solution. Simulations show that the proposed classifier improves the performance considerably. An added advantage of our algorithm is that the transmitted symbol sequence is also demodulated in conjunction with the classification process and no subsequent demodulation is necessary for the data symbols.