PAM4 rolling-shutter demodulation using a pixel-per-symbol labeling neural network for optical camera communications

Area of Science:

• Optical Camera Communication (OCC)
• Data Transmission
• Machine Learning

Background:

• Traditional optical camera communication (OCC) relies on binary intensity modulation, limiting data transmission rates.
• Higher data rates are crucial for advancing OCC applications, necessitating multi-level modulation formats.

Purpose of the Study:

• To introduce and demonstrate a novel rolling shutter 4-level pulse amplitude modulation (PAM4) demodulation scheme for OCC systems.
• To enhance data transmission speeds in OCC by utilizing a pixel-per-symbol labeling neural network (PPSL-NN).

Main Methods:

• Developed a pixel-per-symbol labeling neural network (PPSL-NN) tailored for rolling shutter OCC.
• Implemented a scheme to calculate and re-sample pixel-per-symbol (PPS) data for accurate PAM4 symbol labeling.
• Integrated PAM4 modulation within the OCC framework to enable multi-level signaling.

Main Results:

• Achieved a significant bit-rate distance product of 28.8 kbit/s • m per color.
• Demonstrated the efficient demodulation of high-speed PAM4 signals within the rolling shutter OCC pattern.
• Validated the effectiveness of the PPSL-NN in handling the complexities of the proposed scheme.

Conclusions:

• The proposed rolling shutter PAM4 demodulation scheme offers a viable method for increasing data rates in OCC systems.
• The PPSL-NN is effective in processing and demodulating multi-level signals in OCC, paving the way for faster optical wireless communication.
• This advancement represents a key step towards higher-performance optical camera communication.