A Real-Valued 4D Memory Polynomial Algorithm for Mixer Modeling

Paper i proceeding, 2021

In this paper, we propose a real-valued multi-input memory polynomial (MP) for behavioral modeling of a mixer, which considers nonlinearity introduced by in-phase, quadrature and the local oscillator (LO) inputs. By splitting the I and Q data of the stimulus and LO signal, we use a 4-input 2-output structure of the real-valued MP model. The experimental validation based on data measured from a real mixer confirms the effectiveness of our proposed model. Compared with the Volterra series in the literature, the real-valued MP has similar modeling accuracy with much lower complexity,