Скачать или смотреть Machine Learning-Based Underwater Visible Light Communication

The research proposes a novel methodology for Underwater Visible Light Communication, mainly focusing on two aspects of an underwater communication system; the design and construction of an underwater communication model and improving the performance of the communication system according to the channel condition. The proposed communication system consists of a transceiver designed as a LED array, a water tank as the channel, and a camera receiver to capture and identify patterns of signals received at the receiving end. Furthermore, the performance of the communication system is improved by using encoding and decoding techniques to reduce errors and making the system adjust according to different channel conditions.


Other communication methods available for underwater communication are ultrasonic which is subject to noise from reflections, and radio waves which have limitations as the attenuation rate are higher. Visible Light Communication is a method of underwater communication that has gained interest as a research area to solve complications in other forms. In the proposed system, different channel conditions such as high turbidity conditions, water bubble disturbances, and obstacles are addressed as major underwater visible light communication problems. The communication model can detect the condition of the channel, and it is improved to adjust under the detected condition using image processing and machine learning techniques.


The proposed communication system consists of a transceiver designed as a LED array which consists of 36 points LED points with Green, Blue & Red LEDs in each point(108 LEDs altogether) which can be used to transmit signals in different wavelengths and transmit with more intensity when required depending on the channel conditions such as turbidity level, Air bubbles & other obstacles. A water tank with the size of 1m in length and height, a Width of 0.5m is used as the channel, and a camera receiver to capture and identify patterns of signals received at the receiving end. Feedbacks for the transmitter based on the channel conditions was indicated using the GUI.

𝗦𝘂𝗽𝗲𝗿𝘃𝗶𝘀𝗼𝗿𝘀
Dr.S.A.H.A Suraweera
Dr.G.M.R.I Godaliyadda
Dr.M.P.B. Ekanayaka

𝗧𝗲𝗮𝗺 𝗠𝗲𝗺𝗯𝗲𝗿𝘀
Name : Madushanka Munasinghe
E mail : [email protected]
LinkedIn :   / madushankamunasinghe  


Name : Disnaka Ranasinghe
E mail : [email protected]
LinkedIn :   / disnaka-ranasinghe-7145b5153  


Name : Ashani Seneviratne
E mail : [email protected]
LinkedIn :   / ashani-seneviratne-239b72177