Intelligent Partial Pattern Mode Matching Receiver For Orbital Angular Momentum Systems

This work proposes an intelligent receiver for wireless Orbital Angular Momentum (OAM) systems. The proposed receiver consists of an optical filter followed by a neural network. The filter is designed to partially match a range of OAM modes such that the intensity of its output represents the projection of the received signal over the different modes (including fractional modes). The output of the filter is then passed as an input to a pre-trained shallow neural network to decide on the transmitted symbol. Simulations show that the proposed receiver outperforms both the conventional matched filters based receiver and the Fourier-based receiver.