Learning-Based User Association for MmWave Vehicular Networks With Kernelized Contextual Bandits

2025-04-15

AI Summary

In mmWave vehicular networks, base station association requires real-time channel state information (CSI), yet frequent channel estimation incurs prohibitive overhead. To address this trade-off, we propose the Distributed Kernelized Contextual Upper Confidence Bound (DK-UCB) algorithm. DK-UCB predicts instantaneous transmission rates online using only historical vehicle context—such as position and velocity—without additional channel measurements. Its key contributions are: (1) a customized kernel function incorporating mmWave propagation characteristics, enabling nonlinear rate–context modeling in a reproducing kernel Hilbert space; and (2) a lightweight inter-vehicle exploration information sharing mechanism that enables distributed online learning with low communication overhead and accelerated global convergence. Experiments demonstrate that DK-UCB significantly reduces channel probing overhead, improves association accuracy and network throughput, and approaches the optimal long-term cumulative rate even under high-mobility conditions.

Technology Category

• Machine Learning: Kernel Methods
• Planning, Routing, and Scheduling: Scheduling under Uncertainty
• Search and Optimization: Distributed Search

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Experiences and lessons learnt from Web-based algorithms and system deployments
• Economics, Online Markets and Human Computation: Incentives in network design for Web infrastructures and ecosystems
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

Vehicles require timely channel conditions to determine the base station (BS) to communicate with, but it is costly to estimate the fast-fading mmWave channels frequently. Without additional channel estimations, the proposed Distributed Kernelized Upper Confidence Bound (DK-UCB) algorithm estimates the current instantaneous transmission rates utilizing past contexts, such as the vehicle's location and velocity, along with past instantaneous transmission rates. To capture the nonlinear mapping from a context to the instantaneous transmission rate, DK-UCB maps a context into the reproducing kernel Hilbert space (RKHS) where a linear mapping becomes observable. To improve estimation accuracy, we propose a novel kernel function in RKHS which incorporates the propagation characteristics of the mmWave signals. Moreover, DK-UCB encourages a vehicle to share necessary information when it has conducted significant explorations, which speeds up the learning process while maintaining affordable communication costs.

Problem

• Estimates mmWave channel rates without frequent costly measurements
• Uses kernelized contextual bandits for nonlinear rate-context mapping
• Enables efficient vehicle-BS association with shared exploration data

Innovation

• Uses DK-UCB algorithm for rate estimation
• Maps context to RKHS for linear mapping
• Incorporates mmWave propagation in kernel function