Mitigation of Mach Zehnder modulator nonlinearity in millimeter wave radio over fiber system using digital predistortion

In this era of multiscreen generation, with connected devices per person escalating dramatically, the transmission of uncompressed videos and tons of data over wireless networks have driven the wireless networks to migrate from lower radio frequency to higher mm wave frequency band. The standard 802...

Full description

Saved in:
Bibliographic Details
Main Author: Duraikannan, Shankar
Format: Thesis
Language:English
Published: 2017
Online Access:http://psasir.upm.edu.my/id/eprint/70222/1/FK%202017%20116%20IR.pdf
http://psasir.upm.edu.my/id/eprint/70222/
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Universiti Putra Malaysia
Language: English
Description
Summary:In this era of multiscreen generation, with connected devices per person escalating dramatically, the transmission of uncompressed videos and tons of data over wireless networks have driven the wireless networks to migrate from lower radio frequency to higher mm wave frequency band. The standard 802.11ad recommends the usage of 7 GHz unlicensed frequency spectrum at 60 GHz. The spectrum with low multipath impairment, suffers a high channel attenuation that demands mixed architecture of radio and fiber for enhancement of coverage distance. Radio over Fiber (RoF) using Mach Zehnder Modulator (MZM) is the most widely adapted architecture for mm-wave generation. However, the architecture with low insertion loss, power consumption, and dispersion effects suffers the effect of MZM nonlinearity that significantly limits the performance of RoF system. This thesis proposes an I/Q channel separated coherent optical OFDM transmission system at 60 GHz, that employs mm-wave generation by optical frequency up-conversion using cascaded dual drive MZM (DD-MZM) and dual parallel MZM (DP-MZM) architecture at the transmitter and with coherent optical detection at the remote antenna unit. The first stage DD-MZM generates a carrier suppressed odd harmonics of the input optical signal from the laser diode modulated by RF signal. The second stage DP-MZM followed by the Gaussian optical band pass filter (GOBPF) that passes the desired (fifth) harmonic of the optical signal at its output, generates I/Q channel separated OFDM baseband modulated optical signal. The coherent detection of the modulated optical signal received at the Remote Antenna Unit (RAU) produces the 60 GHz mm-wave that is transmitted wirelessly to the Mobile Unit (MU). The theoretical and simulation analysis of the techniques for 16QAM/OFDM signal is performed. The simulation results in an Error Vector Magnitude (EVM) of 10 percent and 13 percent at 10km and 80 km respectively, a reduced power penalty of 2 dB/km at 80 km and enhanced data rate of 40 Gbps with only 10 GHz signal bandwidth that clearly indicates the accuracy of the technique in mm-wave radio signal generation and transmission over fiber. Further with I/Q channel separation, harmonic distortion due to intermediate frequency translation is reduced along with the reduced computational and circuit complexity. However, with coherent optical orthogonal frequency division multiplexing adopted to achieve multi-gigabit transmission the system becomes sensitive to nonlinear distortions induced by MZM. Therefore, this thesis further analyses the modulator nonlinearity and proposes an adaptive digital pre-distortion (DPD) to mitigate the MZM modulator nonlinearity. The proposed adaptive digital pre-distortion is based on memory polynomial (MP) model with indirect learning architecture (ILA) where the predistorter is modeled as an inverse polynomial model of the nonlinear RoF system. The predistorter is the copy of the training filter that is connected as the post distorter to the nonlinear RoF system. The coefficient computation is performed using recursive prediction error method (RPEM) algorithm which shows a dominant spectral regrowth reduction and in-band distortion reduction with reduced complexity compared to the commonly used slow converging, least mean square algorithm. The RoF system with and without the DPD is simulated and the results demonstrate that the MZM nonlinearity is compensated using the proposed adaptive DPD and substantially improves the performance of the system in terms of Adjacent Channel Leakage Ratio (ACLR) and EVM. The ACLR is improved by 10 dB and the EVM is reduced from 13 percent to 0.06 percent at 80 km.