Non-line-of-sight optical camera communications based on CPWM and a convolutional neural network

Xiangyu Wan, Bangjiang Lin*, Zabih Ghassemlooy, Tianming Huang, Jiabin Luo, Yongqi Ding

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

1 Citation (Scopus)


Non-line-of-sight (NLOS) optical camera communications (OCC) exhibit greater link availability and mobility than line-of-sight links, which are more susceptible to blocking and shadowing. In this work, we propose an NLOS OCC system, where the data signal is mapped into color pulse width modulation (CPWM) symbols prior to transmission using a red-, green-, and blue light-emitting diode. A convolutional-neural-network-based receiver is used to demodulate the CPWM signal. Based on experimental results, the proposed scheme effectively mitigates the effects of diffuse reflection induced intersymbol interference, resulting in an increased data transmission rate to 7.2 kbps over a link span of more than 2 m, which is typical for indoor applications.
Original languageEnglish
Pages (from-to)7367-7372
JournalApplied Optics
Issue number28
Early online date21 Sept 2023
Publication statusPublished - 1 Oct 2023

Cite this