Reduction of Beam Pointing Error on Free Space Optical Communication links using Tree Based Machine Learning Multi-Output Regression Model

Free space optical (FSO) communications have the potential to be one of the most essential technologies for solving the high-bandwidth demands of communications between satellites and ground stations. However, a line-of-sight link is required between the transmitter and the receiver for effective establishment of communication link. In this paper, we examine the FSO beam pointing error systems under controlled weak turbulence using a robust $\mathcal{H}_{\infty}$ control law. The goal is to maintain the transmitter-receiver line, which refers to keeping the optical beam’s center as close as possible to the receiving aperture’s center while staying within the boundaries of the permitted disturbance attenuation level. Based on this, we use a tree-based machine learning technique to predict gain value of $\mathcal{H}_{\infty}$ norm optimization problem guaranteeing a stable closed loop pointing error.