Deep learning based transceiver design for multi-colored VLC systems.

Area of Science:

• Optical Communications
• Artificial Intelligence
• Signal Processing

Background:

• Visible Light Communication (VLC) systems offer high bandwidth potential but face challenges in multi-color modulation and performance.
• Designing efficient transceivers for multi-colored VLC under specific color and illuminance constraints is complex.
• Existing methods often struggle with optimal symbol recovery in complex optical channels.

Purpose of the Study:

• To develop a deep learning-based framework for designing multi-colored VLC systems.
• To jointly optimize the transmitter and receiver pair for enhanced symbol recovery.
• To meet stringent color and illuminance requirements in VLC systems.

Main Methods:

• Utilized an unsupervised deep learning technique, an autoencoder (AE), for end-to-end training.
• Modeled the VLC transceiver pair and the optical channel, including LED intensity control, within the AE framework.
• Employed a deep learning (DL) approach for multi-dimensional color modulation using RGB LED lamps.

Main Results:

• The jointly designed and optimized VLC transmitter and receiver demonstrated superior performance.
• The learned VLC system significantly outperformed existing techniques in average symbol error probability.
• Achieved efficient symbol recovery under multi-color modulation and specific optical channel conditions.

Conclusions:

• Deep learning techniques are highly viable for designing advanced optical communication systems like VLC.
• The proposed autoencoder-based framework enables efficient joint optimization of VLC transceivers.
• This approach paves the way for more robust and high-performance multi-colored VLC systems.