Experimental Demonstration of Mahalanobis Distance Correlation Acquisition for OWC Spread Spectrum Communication Systems

Abstract

Optical wireless communication spread spectrum systems are known for their robustness and ability to resist interference. However, they can be impacted by non-Gaussian noise, leading to challenges such as clock synchronization failures and increased false alarm rates. To tackle this issue, this study transitions from examining signal correlation values to focusing on their distribution characteristics. The approach involves utilizing an adaptive Gaussian mixture model to establish the distribution of background noise correlation values and employing the weighted Mahalanobis distance for effectively differentiating the spread spectrum signal. Through experiments conducted at a signal-to-noise ratio (SNR) of -15 dB, the proposed algorithm demonstrates a significant enhancement in signal detection accuracy, surpassing 58%.