Adapting Optical Amplifier Response Estimation to Consider Non-Flat Input Signals

In optical communication systems, optical amplifiers play a fundamental role. Due to the current challenge to adapt the optical network to be more dynamic, it is necessary to adapt the optical amplifiers so that their operation is autonomous and favors transmission quality. Thus, the ACOP (Adaptive Control of Operating Point) problem arises, whose objective is to adapt the amplifier configuration according to the state of the network. In the solutions in the literature, it is essential to have a set of amplifier operating points, called the power mask. There are techniques capable of estimating the amplifier's behavior at points not defined in the power mask. However, few of them consider non-flat input signal power spectrum, which is common in real-world scenarios. This work proposes adapting a previous optical signal estimation technique to consider power masks created with non-flat signals. The results show that this inclusion improved the performance of the original technique. For example, considering the median of the maximum error, the reduction was from around 4 dB (original technique) to around 0.2 dB (proposed adaptation). Furthermore, the adaptation obtained similar results to another technique in the literature that considers non-flat signals and is based on artificial neural networks.