Implementing Intelligent Controls into Gravitational Wave Detectors

Advanced LIGO is a dual-recycled Michelson interferometer and its ability to detect gravitational waves hinges on the precise characterization of its multiple degrees of freedom. These are typically divided into two primary subsystems, Length Sensing and Control (LSC) and Angular Sensing and Control (ASC). Applying feedback to these systems is essential for maintaining each optical cavity's proper resonant conditions. However, achieving both robustness and low noise is challenging due to cross-couplings between the degrees of freedom. Such interactions can lead to significant noise contributions below 25 Hz. Intelligent control theory is a technique which utilizes machine learning algorithms to optimize classical feedback systems. We show how Independent Component Analysis (ICA) can be used to optimize both the LSC and ASC subsystems. This method aims to mitigate control-induced noise, improve low-frequency sensitivity, and expand LIGO's detection capabilities.