Minimization of Feedback Loop Distortions in Digital Predistortion of a Radio-Over-Fiber System With Optimization Algorithms

This article proposes the linearization of an intensity modulation/direct detection radio-over-fiber (RoF) link with feedback loop. The goal is to carry out the predistortion process in a real scenario, in which the output signals are a few kilometers far from the baseband unit (BBU). First, the feedback loop is considered ideal, so the output signals are captured at the remote radio head side. Then, the feedback loop is taken into account, and both the input and output signals are captured at the BBU side. Applying optimization algorithms, such as Fibonacci, Golden, or Powell, it is possible to seek the optimal attenuation value within only a few iterations, which minimizes the distortion of the feedback loop. Experiments are carried out in a RoF system with 10 and 25 km length fiber within the long-term evolution (LTE) standard. Measured results show how with a proper choice of the attenuation it is possible to reach analogous results regarding to an ideal feedback loop in terms of adjacent channel power ratio, the output signal power and error vector magnitude.