Spectral-Based Proactive Blockage Detection for Sub-THz Communications.

Abstract

Human body blockage is recognized as a major challenge in sub-terahertz (sub-THz,100-300 GHz) cellular systems, resulting in significant performance degradation. To mitigate this issue, the ability to detect blockage events before their occurrence is required for effective blockage avoidance solutions such as multi-connectivity. Traditional time-domain approaches, often utilizing machine learning, have been proposed; however, they are limited in performance. In this work, the spectral domain is explored as a more efficient alternative, where the differences between blockage and non-blockage periods are wider and directly measurable. A simple threshold-based proactive blockage detection algorithm is proposed and evaluated. The algorithm was tested using experimentally measured blockage traces at 156 GHz. Key performance metrics, including blockage detection probability, mean time to blockage, and false alarm rate, were used to assess its effectiveness. It was found that the proposed method can detect blockage events at least 50 milliseconds before their occurrence, with a false alarm rate of less than one event per second, even in fading environments. The proposed spectral-domain-based algorithm is shown to outperform traditional time-domain approaches in the proactive detection of blockage events, making it a promising solution for future sub-THz and mmWave communication systems.