Theoretical characterisation of memory polynomial systems with in-phase-quadrature imbalanced Gaussian inputs: application to orthogonal frequency division multiplexing error rate analysis

This study addresses the problem of theoretical characterisation of memory polynomial systems driven by complex-valued Gaussian signals with real and imaginary components imbalanced with respect to those of the original information signal. In various practical applications, both effects are observed jointly. The authors prove a theorem about a representation of the distorted signal in the form of the sum of two uncorrelated components, one of which is a linearly distorted version of the information signal. The derived results are applied to orthogonal frequency division multiplexing error rate analysis.