Low Complexity MIMO Modulation Classification via Distribution Test Ensembles

Abstract

Modulation classification in MIMO systems needs to overcome its complexity demand to be viable in a real world environment. The number of mathematics operations required by the state-of-the-art maximum likelihood classifier grows exponentially with the number of transmitting antennas and the orders of candidate modulations. In this paper, we propose a low complexity MIMO system modulation classifier by combining multiple magnitude-based distribution tests using a simple Multi-layer perceptron. The resulting solution provides good classification performance in AWGN channel while outperforming the maximum likelihood classifier in fading channels. Moreover, its computational complexity is much lower and does not scale with the number of transmitting antennas or the orders of candidate modulations.