Machine Learning-Driven UAV-Enabled Sensor Networks and V2X Communication for Robust Vehicle Guidance in Underground Passages

Abstract

Machine learning-based UAV networks integrate advanced sensor data to enhance the accuracy of vehicle guidance systems in underground passages by combining information from multiple sources. Autonomous vehicles rely on uninterrupted communication channels and real-time environmental feedback, yet subterranean corridors pose challenges for GPS-based navigation and traditional wireless transmission protocols. Machine learning modeling coupled with robust channel estimation techniques helps overcome these limitations by facilitating detection of obstacles, improving path planning, and mitigating interference. V2X communication frameworks enable stable vehicle-to-infrastructure interactions with dynamically updated network topologies and edge-based computing strategies. Research efforts examine the effectiveness of machine learning algorithms in predicting network conditions and adjusting UAV trajectories to maintain link quality and vehicle control. This paper explores architectural designs, algorithmic foundations, and experimental evaluations of machine learning-driven UAV-enabled sensor networks and V2X communication strategies to provide reliable guidance for vehicles operating in underground environments. Results illustrate improvements in communication throughput, location precision, and navigation safety, thereby demonstrating the viability and impact of these integrated approaches. Focus is placed on architectural flexibility, scalability, and robustness against faulty links and environmental uncertainties. Insights gained offer guidelines for deploying machine learning-driven solutions that address critical communication, localization, and safety objectives in underground transportation infrastructures.