Memetic Algorithm for Digital Pre-Distortion of Wimax Power Amplifier Based on a Memory Polynomial Model

Digital Pre Distortion (DPD) is one of the fastest and cost effective digital processing technique to mitigate the distortions in Power Amplifier (PA) caused by nonlinearity and memory effects. This paper presents a general DPD architecture for wideband PA systems based on a direct learning approach with constrained feedback bandwidth. The proposed scheme is based on evaluating the Power Spectral Density (PSD) of the PA and DPD output signal. The coefficients of PA and DPD are optimized iteratively in order to minimize the output PSD around the pre specified frequency. In this work, a stochastic optimization algorithm called Memetic Algorithm (MA) is used to optimize the coefficients of PA and DPD. This approach encourages scientific discussions and technological innovations toward an effective solution for reducing DPD system implementation complexity and cost. The study evidences prove that MA produces promising results for Worldwide Interoperability of Microwave Access (WiMAX) signal.