Channel Prediction in Vehicular-to-Everything (V2X) Networks Using Image Sensing and Computer Vision

With the rapid development of intelligent trans-portation systems (ITS), a growing number of vehicular ap-plications have emerged to provide an entirely new experience in our daily life. To satisfy such demanding quality of service (QoS), mmWave frequency bands are widely used in 5G and 6G wireless communications. To compensate for the severe path loss, the beamforming techniques realized by the ultra-massive multiple-input-multiple-output (UM-MIMO) antenna arrays are inevitable in mm Wave systems. To obtain an accurate channel information in a real-time manner, we propose channel prediction using sensing and computer vision (CP-SCV). Since CV technique can precisely observe surrounding environment, we can extract the positions, orientations, and materials of mobile devices and entities affecting signal propagation. From practical experiments, we show that CP-SCV achieves significant channel prediction accuracy improvement over the conventional schemes.