Nonuniform memory polynomial behavioral model for wireless transmitters and power amplifiers

This paper presents a low complexity memory polynomial model suitable for behavioral modeling of wireless transmitters and power amplifiers exhibiting memory effects. This model takes advantage of the fading property of memory effects and separately adjusts the nonlinearity order in each of the memory polynomial branches. An efficient algorithm is reported to successively identify the nonlinearity order required in each of the model branches. Experimental validation carried on a 3G Doherty power amplifier prototype driven by two multi-carrier WCDMA signals shows that the proposed model reduces the required number of coefficients by at least 30% while maintaining the modeling accuracy. This complexity reduction is highly desired in emerging applications where the operating conditions change rapidly and thus requires fast adaptation.