Sub-column pixel neural network scheme for modulation format shifting based optical camera communications.

In this Letter, a sub-column pixel neural network (SCPNN) scheme is proposed to label the pixel patterns in modulation format shifting based optical camera communication (MFS-OCC) systems. The MFS scheme, by identifying on-off keying (OOK) and Manchester signal formats, can additionally harvest one-bit information without sacrificing the pixel width per bit, resulting in a higher transmission rate for OCC without multi-level modulation. Compared with the conventional multi-threshold decision (MTD) scheme, the proposed SCPNN can effectively mitigate the impact of inter-symbol interference (ISI) and noise on signal decision by making full use of all pixels in one-bit period. The experimental results show that, in the case of the illuminance as low as 300 lux, the SCPNN scheme can attain a data rate of 8.16 Kbps with an average bit-error-rate (BER) of 2.2 × 10 below the HD-FEC limitation of 3.8 × 10. Meanwhile, the SCPNN scheme achieves remarkable BER performance improvement compared to both MFS-OCC with MTD scheme and conventional OOK-OCC system.