Design of matched filter for signal‐to‐noise ratio maximization to improve data rate

Abstract

AbstractThe matched filter is an optimal linear filter for maximizing signal‐to‐noise ratio (SNR) in the presence of additive random noise. Learning‐based matched filter is the proposed work. The proposed model supports (i) maximizing the SNR and (ii) setting the threshold, it is also called an intelligent matched dual tire filter. This improves the SNR by about 20 dB and the corresponding data rate to over 400 MBit/s. The bit error rate for 20 dB is very lower than FEC (10−3) of up to 10−7. The accuracy of the proposed matched learning filter is proved to be above 95%. Indoor visible light communication is one of the applications, and it can also be used for image processing and outdoor communication such as satellite and radar communication. It is obvious for communication as it avoids signal degradation, lobing, and overfitting to achieve a better data rate. This requires funding support and infrastructure facility that can be used to deploy the proposed algorithm in the real‐time scenario in the future.