Adaptive optical feedforward linearization of optical transceiver for radio over fiber communication link

Abstract

With the tremendous growth in numbers of mobile data subscribers and explosive demand for mobile data, the current wireless access network need to be augmented in order to keep up with the data speed promised by the future generation mobile network standards. Radio over fiber technology (RoF) is a cost effective solution because of its ability to support numerous numbers of simple structured base stations by consolidating the signal processing functions at the central station. RoF systems are analog systems where noise figure and spurious free dynamic range (SFDR) are important parameters in an RoF link. The nonlinearity of a laser transmitter is a major limiting factor to the performance of an RoF link, as it generates spurious spectral components, leading to intermodulation distortions (IMD), which limit the achievable SFDR of the analog RF wave transmissions. The device nonlinearity can be mitigated through various linearization schemes. The feedforward linearization technique offers a number of advantages compared to other techniques, as it offers good suppression of distortion products over a large bandwidth and supports high operating frequencies. On the other hand, feedforward linearization is a relatively sensitive scheme, where its performance is highly influenced by changing operating conditions such as laser aging, temperature effect, and input signal variations. Therefore, for practical implementations the feedforward system has to be real-time adaptive. This thesis aims to develop an adaptive optical feedforward linearization system for radio over fiber links. Mathematical analyses and computer simulations are performed to determine the most efficient algorithm for the adaptive controller for laser transmitter feedforward linearization system. Experimental setup and practical measurement are performed for an adaptive feedforward linearized laser transmitter and its performance is optimized. The adaptive optical feedforward linearization system has been modeled and simulated in MATLAB Simulink. The performances of two adaptive algorithms, which are related to the gradient signal method, such as least mean square (LMS) and recursive least square (RLS) have been compared. The LMS algorithm has been selected because of its robustness and simplicity. Finally, the adaptive optical feedforward linearization system has been set up with digital signal processor (DSP) as the control device, and practical measurement has been performed. The system has achieved a suppression of 14 dB in the third order IMD products over a bandwidth of 30 MHz, in a two-tone measurement at 1.7 GHz.