Luminance Inversion for Parallel Transmission Visible Light Communication Between LCD and Image Sensor Camera

Abstract

Luminance inversion (LI) is proposed to compensate nonlinear characteristics between a liquid crystal display (LCD) and an image sensor (IS) camera of the parallel transmission (PT) visible light communication (VLC). In the PT VLC system reported in <xref ref-type="bibr" rid="ref1" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">[1]</xref> , which incorporates <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">N</i> spatial frequencies with <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">P</i> -ary phase shift keying ( <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">P</i> -PSK) through <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">M</i> -point inverse discrete Fourier transform ( <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">M</i> -IDFT), each point consists of regions of pixels on an LCD and the <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">M</i> points transmit data spatially in parallel at one frame. Since the transfer characteristics between the LCD and IS camera is nonlinear, the received signal at the IS camera receiver is distorted. The effect is severe because the signal of the PT system is, in essence, the discrete multitone (DMT) as a baseband version of orthogonal frequency-division multiplexing (OFDM) signal, thus suffering from high peak-to-average ratio (PAPR). To linearize it between the LCD and IS camera, our approach is to calculate a complement of the digital data fed to the driver of the pixel and then to feed the pair of the digital data and its luminance complement to the driver during two successive frames. In the receiver, the difference of the signals received in the two successive frames is calculated for each of the <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">M</i> detected regions on the IS plane, and then they input to the following <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">M</i> -point DFT in the PT system. With this LI process, wider linear characteristics is obtained than without LI. Experimental results show that the number of bits per frame that the PT system with LI can accommodate without any error is 315 bits/frame and 100 bits/frame for 8- and 16-PSK, which are much larger than 12 bits/frame and 0 bits/frame for the 8- and 16-PSK PT system without LI, respectively.