Push the Limit of Light-to-Camera Communication

Recent years have witnessed the conceptualization and development of light-to-camera communication (LCC). Various innovative modulation schemes are designed for throughput improvement. Nevertheless, the fundamental limits of LCC itself are still unknown from a theoretical perspective. In this paper, we present the LCC channel model to shed light on a question: how to improve the system performance under identical conditions of transmitting/receiving device and modulation scheme. We reveal the channel parameters of LCC that affect the formation of the signal layer and provide a design guideline for increasing the symbol rate. Moreover, we improve the demodulation performance by sampling judging points from the separated signal layer instead of detecting the widths of bands in capture images. To enhance the signal-to-noise ratio, we propose a short-time linear regression algorithm to filter out the noise from the signal layer. We build a hardware prototype with the simple ON-OFF keying to demonstrate the proposed channel model under various experiment settings. The results show that the system throughput reaches 8.2 kbps with a symbol rate of 25 kBd. We believe our techniques for symbol rate increasing and demodulation improvement can help other LCC systems improve their performance.