Neural Network and Memory Polynomial Methodologies for PA Modeling

FPGA-based test bed for measurement of AM/AM and AM/PM distortion and modeling memory effects in RF PAs

“…The phase-to-amplitude transformation simplifies the phase measurements by converting the phase distortion into a proportional amplitude change. If two sine waves of angular frequency ω, amplitude A ¼V PEAK , and phase ϕ are considered, the difference in phase between the two waveforms, assuming both of them are cosine waves, can be obtained by (4). Similarly, for two sine waves, (5) gives the value of V diff .…”

“…These conditions necessitate the modeling of the memory effects added by the PA. For instance, several kinds of behavioral modeling approaches have been explored, such as memory and memoryless models [1,2], polynomial models [3], and neural networks [4,5]. Other models are classic Volterra series [6][7][8] or variations of Volterra series models [9] that provide a proper method of modeling the nonlinear order and memory effects.…”

Modeling memory effects in RF power amplifiers applied to a digital pre-distortion algorithm and emulated on a DSP-FPGA board

“…That way, to mitigate undesired intermodulation and distortion effects, recent advances related to linearization techniques have been developed, mainly to compensate memory effects induced by real PAs [2,3]. These issues have been partially solved introducing modeling techniques for PAs, such as: memory or memoryless models [4,5], polynomial models [6], neural networks [7,8], others techniques taking into account memory effects based on Volterra series [9][10][11][12][13], etc., all of them providing a proper way to capture the nonlinearity order and memory depth.…”

Behavioral model on transmitter power amplifier and IQ imbalance

“…The neural network has also been applied to model the behavior of power amplifier nonlinearity and memory effects, either in feed-forward structure [59][60][61][62][63]68] or in 39 feedback [64][65][66] structure. However, these models do not consider IQ imbalance.…”

“…Besides polynomial based models, there is another category of power amplifier behavioral modelneural network [59]. It has several inner stages and many inner-states, and the final output is determined by both the input and the inner states.…”