Decoding scheme based on CNN for mobile optical camera communication
Ke Yu, Jing He, Zheng Huang
Abstract
A decoding scheme based on a convolution neural network (CNN) is proposed and experimentally demonstrated in mobile optical camera communication (OCC). The CNN can be used to extract features between bright and dark stripes in images effectively. Thus, it can alleviate the stripe distortion in the mobile environment and reduce bit error rates (BERs) by using the proposed decoding scheme based on CNN. A controllable lateral and vertical mobile platform is built to simulate the mobile scenarios with different moving speeds (40–80 cm/s). The experimental results show that, at the moving speed of 80 cm/s, the proposed scheme based on CNN can achieve the BERs of <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" display="inline"> <mml:mrow class="MJX-TeXAtom-ORD"> <mml:mn>3.8</mml:mn> </mml:mrow> <mml:mo>×</mml:mo> <mml:mrow class="MJX-TeXAtom-ORD"> <mml:msup> <mml:mrow class="MJX-TeXAtom-ORD"> <mml:mn>10</mml:mn> </mml:mrow> <mml:mrow class="MJX-TeXAtom-ORD"> <mml:mo>−</mml:mo> <mml:mn>5</mml:mn> </mml:mrow> </mml:msup> </mml:mrow> </mml:math> at the lateral case and <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" display="inline"> <mml:mrow class="MJX-TeXAtom-ORD"> <mml:mn>1</mml:mn> </mml:mrow> <mml:mo>×</mml:mo> <mml:mrow class="MJX-TeXAtom-ORD"> <mml:msup> <mml:mrow class="MJX-TeXAtom-ORD"> <mml:mn>10</mml:mn> </mml:mrow> <mml:mrow class="MJX-TeXAtom-ORD"> <mml:mo>−</mml:mo> <mml:mn>5</mml:mn> </mml:mrow> </mml:msup> </mml:mrow> </mml:math> at the vertical case in a mobile OCC system.