Zero-Shot Generalization for Blockage Localization in mmWave Communication

TL;DR

This paper presents a self-supervised deep learning method that predicts blockage locations in mmWave communication systems using RF data, achieving high accuracy and adaptability in dynamic environments.

Contribution

It introduces a novel self-supervised approach that labels RF data with LiDAR-derived blockage locations, enabling blockage prediction without retraining for different transmitter-receiver positions.

Findings

Achieves up to 74% accuracy in dynamic environments.

Demonstrates robustness and adaptability of the method.

Uses RF data alone for blockage localization.

Abstract

This paper introduces a novel method for predicting blockages in millimeter-wave (mmWave) communication systems towards enabling reliable connectivity. It employs a self-supervised learning approach to label radio frequency (RF) data with the locations of blockage-causing objects extracted from light detection and ranging (LiDAR) data, which is then used to train a deep learning model that predicts object`s location only using RF data. Then, the predicted location is utilized to predict blockages, enabling adaptability without retraining when transmitter-receiver positions change. Evaluations demonstrate up to 74% accuracy in predicting blockage locations in dynamic environments, showcasing the robustness of the proposed solution.