Comparative Performance Analysis of Analog and Digital Optical Wireless Communication Through Water-Filled Pipes Using Visible-Light LEDs

This paper presents a comparative investigation of analog and digital optical wireless communication (OWC) through a water-filled polyvinyl chloride (PVC) pipe sealed at both ends with optically transparent PMMA windows using visible-light light-emitting diodes (LEDs) operating at four wavelengths: blue (475 nm), green (528 nm), yellow (583 nm), and red (625 nm). A unified experimental and simulation framework is developed to evaluate continuous (analog) and modulated (digital) transmission under varying water-filling conditions of 0%, 25%, 50%, 75%, and 100%. In the analog analysis, the received optical power is measured to quantify absorption, refraction, and scattering effects based on the Beer–Lambert model. For digital transmission, pseudo-random binary sequence (PRBS7) on–off keying (OOK) modulation is employed to assess signal-to-noise ratio (SNR), analytically estimated bit-error rate (BER), eye diagrams, and information-theoretic channel capacity upper bounds. The presented digital performance metrics are intended for relative trend analysis and physical interpretation, rather than demonstrating a fully optimized or practically deployable communication system. The results show that the 50% water-filled condition introduces the most severe impairment due to repeated air–water interface interactions, which cause additional Fresnel losses, angular redistribution, and enhanced multipath scattering. In contrast, fully submerged operation provides a more homogeneous propagation medium, leading to improved transmission behavior. Among the tested wavelengths, the red LED exhibits the highest relative robustness under the present baseline implementation. The reported digital metrics should be interpreted as comparative indicators for trend analysis and physical interpretation of channel-dependent degradation, rather than as evidence of reliable or optimized communication performance.