Skip to Main Content
This paper proposes a new identification technique of Wiener or Hammerstein models applied to model the behavior of wideband RF transmitters. A dynamic exponential weighted moving average (DEWMA) algorithm was firstly applied to the raw signals sampled at the transmitter input and output to deduce the static AM/AM and AM/PM curves that are attributed to the memoryless nonlinear behavior. These memoryless curves were then utilized to deduce an intermediate set of data used in the identification of the dynamic linear sub-model. To model the frequency response and/or the memory effects, a finite impulse response (FIR) filter topology could be employed. The validation of the modeling approach was carried out using a 60-Watt GaAs FET push pull amplifier operating between 1930 MHz to 1990 MHz, which was applied by a two-carrier WCDMA.