Free Space Optical Communication (FSOC) has become a popular wireless communications technique for providing and supporting optical high-speed bandwidth data transmission for telecommunication and computer networking. FSOC is expected to supplement traditional Radio Frequency (RF) technologies and successfully aid in removing congestion from the overly crowded RF spectrum, and its optical fiber communications.

FSO system performance is highly dependent on channel conditions, wherein background noise poses a significant problem, even in the absence of weather and/or atmospheric turbulence. Transmitted signals are significantly affected by background noise (e.g., thermal, shot noise, dark currents) primarily on the receiver side, which leads to system performance deterioration. Such effects are often described by Additive White Gaussian Noise (AWGN). This phenomenon affects the communication link and can hinder the accurate detection of information.

The work reported in this thesis investigated the addition of generated AWGN to single FSOC links, as well as the extraction of noise signals at the receiver end via a subtraction method. Test results demonstrated that AWGN can be extracted from an FSOC signal when standard deviation and noise signal mean are estimated using a Gaussian Mixture Model (GMM). Outcomes show there is approximately 80% cross-correlation when compared with an original Pseudorandom Binary Sequence (PRBS) signal.