Deep-Learning-Based Security of Optical Wireless Communications for Intelligent Transportation Digital Twins Systems

After the Internet of Things technology is applied to the intelligent transportation system (ITS), the ITS is becoming more perfect. Optical wireless communications (OWC) in intelligent transportation is optimized by in-depth analysis in the present work. Working principles of visible light communications (VLC) in intelligent transportation are discussed systematically. Afterward, a scheme based on carrier-less amplitude/phase modulation is proposed. VLC is optimized by introducing an adaptive decision feedback equalizer (DFE). While the optical signal spectra are analyzed, a convolutional neural network is employed to extract features from the original images, avoiding data reconstruction. Simulation experiments reveal that by combining the digital domain linear equalizer with the modified DFE, the error vector magnitude can be increased by about 4 dB compared to that yielded by single linear equalizers. Taking a support vector machine as the binary classifier, a hyperplane is generated to classify the two sets of data. Consequently, the recognition accuracy for spectrum parameters reaches almost 100 percent, which is significantly better than other deep learning (DL) algorithms. To sum up, DL-based spectrum analysis can apply to OWC systems for intelligent transportation. The research is of great significance to improve the performance of optical wireless communication in intelligent transportation and meet higher traffic demand.