A New Transceiver Design for Protograph LDPC-Coded LACO-OFDM VLC Systems with Deep Learning

As a spectral-efficient and transmission-reliable technique, channel coded layered asymmetrically clipped optical orthogonal frequency-division multiplexing (LACO-OFDM) has attracted considerable interest in visible light communications. In this letter, we investigate the protograph low-density parity-check (LDPC)-coded LACO-OFDM VLC systems. Specially, we design a novel type of interleaving scheme, referred to as “variable-node subcarrier matched mapping (VNSMM)” to optimize the performance of protograph-coded LACO-OFDM VLC systems. In such scenario, the conventional iterative detection framework cannot completely eliminate the inter-layer interference (ILI), which degrades the system performance. To further suppress the ILI at the receiver, we propose a deep learning (DL)-based detection framework, called “DeepLACO”, to simultaneously obtain the log-likelihood ratios (LLRs) used for the protograph decoders in all layers without the knowledge of channel state information (CSI). Analyses and simulations demonstrate that the protograph-coded LACO-OFDM exploiting the proposed interleaving scheme and detection framework can achieve excellent performance in VLC systems.