Blind modulation classification algorithm based on machine learning for spatially correlated MIMO system

Spatial correlation is a decisive factor for pragmatic multiple-input multiple-output (MIMO) system, simultaneously bringing about some problems in the received signal modulation identification respect. In this study, the authors focus on blind digital modulation identification in the spatially correlated MIMO system and deliver a robust signal recognition algorithm based on extreme learning machine (ELM) and higher order statistical features for MIMO signal identification without a priori knowledge of the channel and signal parameters. The superiority of ELM lies in random selections of hidden nodes and ascertains output weights analytically, which result in lower computational complexity. Theoretically, this algorithm has a tendency to supply excellent generalisation performance at staggering learning rate. Further, the simulation results indicate that the ELM could reap a perfectly acceptable recognition performance and thus provides a solid ground structure for tackling MIMO modulation challenges in low signal-to-noise ratio.