| OFDM technology has been widely used because of its high spectrum efficiency and robustanti-interference ability. But the biggest drawback of OFDM signal is high PAPR, whichmakes it so sensitive to the nonlinear characteristic of the power amplifier intransmitter，causing nonlinear distortion. So, to reduce the nonlinear distortiongenerated by PA to OFDM signal, it is necessary to compensate the nonlinearcharacteristic of PA by linearization techniques. So far, many linearization methodshave been proposed, such as feed-forward, feed-back, LINC, and DPD. Particularly,DPD is considered to be the most promising linearization technique and will beintroduced in details.The paper analyzes the nonlinear and memory characteristics of PA based on theOFDM system, besides, elaborates the basic principle of DPD, and focuses on LUT–based adaptive DPD and polynomial-based adaptive DPD. For whether consideringthe memory effect of the power amplifier, four kinds of DPD methods are researchedrespectively, including memoryless LUTs, memory LUTs, memoryless polynomialand memory polynomial, and the simulation results verify the DPD techniques havecompensated PA’s nonlinearity. LUT method is easy to implement, but it will takeup more RAM memory cells, and has slow convergence rate. Polynomial methodhas faster convergence rate, but when the higher order polynomial pre-distorter, thecomplexity of the system also significantly increased, and the pre-distortion effect isnot ideal. So, an improved LUT-based adaptive pre-distortion is proposed, whichcombines LUT method and polynomial method. In this scheme, to make theconvergence rate of LUT faster, firstly, obtaining a PA’s pre-filter with thepolynomial method to compensate PA’s nonlinear characteristics, and then updatingthe contents of LUTs adaptively. And simulation results show that the scheme canimprove the effect of pre-distortion, besides，the speed of convergence of thealgorithm and save hardware resources are improved, so its effectiveness is verified. |
PDF Full Text Request