Optimized Low-Complexity Implementation of Least Squares Based Model Extraction for Digital Predistortion of RF Power Amplifiers

Selection of compressed training data for RF power amplifier behavioral modeling

“…The result is a more general model of the power amplifier. The multiplication operations according to the formula in [1] and execution time by a 2.4GHz Core i5 CPU with 8GB DDR to calculate the coefficients are shown in Table. I. 14-bit standard LS 14-bit compressed model a nonlinear dynamic system.…”

“…Since modern communication standard signals are not truly periodic, it is not possible to simply select a short section of consecutive data to represent the whole dataset. Guan et al proposed an algorithm to use multiple short training data to build multiple small matrices for behavioural modeling [1]. However, this method does not reduce the total number of training samples and cannot guarantee the robustness for a nonlinear system with memory.…”

Low-Complexity Stochastic Optimization-Based Model Extraction for Digital Predistortion of RF Power Amplifiers

“…e n u n u n (9) Owing to the requirements of conventional extraction techniques, the vast majority of modern DPD behavioral models are designed to ensure a linear relationship between the model output and its coefficients. For these linear-inparameters models, it follows that, for each sample, the error measured at the model output has a linear relationship with the model coefficients.…”

“…The LS algorithm offers high accuracy and fast convergence, but it uses complex matrix multiplications and inversions which require substantial hardware resources to execute. Furthermore, the complexity of these matrix operations increases with the number of coefficients employed in the DPD model and the number of samples used in the extraction process [9]. As operating bandwidths in wireless communication systems continue to increase, the nonlinear behavior of the PAs becomes more complicated.…”

Green Communications: Digital Predistortion for Wideband RF Power Amplifiers

“…Table VI summaries the main contributions in the litera- [28], [30], [91] because of their well-known performance in estimating the coefficients of linear-in parameter models. In order to reduce the standard LS complexity and computational latency for physical real-time coefficient extraction and updating, several approaches are also reported in the literature, such as QR-RLS [64] for LUT-based DPDs, the optimized LS [66] for behavior-model-based DPDs, and adaptive basis function based DPD [86]. As the signal bandwidth increases, extracting the DPD parameters using reduced bandwidth becomes a challenging problem.…”