Intelligent reflecting surfaces aided millimetre wave blockage prediction for vehicular communication

• Main Authors: Woon, Fu Seong, Leow, Chee Yen
• Format: Conference or Workshop Item
• Published: 2022
• Subjects: TK Electrical engineering. Electronics Nuclear engineering
• Online Access: http://eprints.utm.my/id/eprint/98796/; http://dx.doi.org/10.1109/ISTT56288.2022.9966540

The ability of millimeter-wave (mmWave) to deliver gigabit throughput has led to its widespread adoption in Fifth Generation (5G) networks. However, mmWave links between Base Station (BS) and users can be easily obstructed by obstacles. In vehicular networks with dynamic environments and mobile users, the mmWave link blockage issue is even more pronounced. In order to preserve the mmWave link in the vehicular network, it is necessary to predict blockages. For blockage prediction, sensor information from Lidar, Radar, and cameras has been considered. Nonetheless, these non-radio frequency methods necessitate the use of additional equipment and signal processing, which raises the implementation cost and complexity. The existing literature also considers the use of BS and user's Radio Frequency (RF) signatures to predict blockage. However, users' mobility has not been taken into account. An Intelligent Reflecting Surface (IRS), on the other hand, has been viewed as a promising method for providing an alternate path by reflecting the mmWave signal between the BS and user in order to improve the reliability of vehicular networks. Therefore, this research investigates the IRS-assisted blockage prediction in order to determine the future link status in the vehicular environment with respect to user mobility. The proposed solution employs a number of active elements that are randomly distributed on the IRS to obtain the RF signatures. Furthermore, it utilises Machine Learning (ML) techniques to learn the pre-blockage wireless signatures, which can predict future blockages. The results indicate that the proposed method can predict blockages between a single IRS and a moving user with a greater than 98 percent accuracy up to one second before they occur.