Dynamic Resource Allocation for LTE-Based Vehicle-to-Infrastructure Networks

• Published in: IEEE transactions on vehicular technology Vol. 68; no. 5; pp. 5017 - 5030
• Main Authors: Shi, Jianfeng, Yang, Zhaohui, Xu, Hao, Chen, Ming, Champagne, Benoit
• Format: Journal Article
• Language: English
• Published: New York IEEE 01.05.2019; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Algorithms; Computer simulation; Downlink; Dynamic resource allocation; Dynamic scheduling; Heuristic algorithms; LTE-based V2I networks; Mobile communication systems; multi-antenna transceiver; mutlivalue discrete particle swarm optimization (MDPSO); Optimization; Particle swarm optimization; Power consumption; Precoding; Resource allocation; Resource management; Roadsides; Subcarriers; Vehicle dynamics; Vehicle-to-infrastructure; Wireless communications
• ISSN: 0018-9545; 1939-9359
• DOI: 10.1109/TVT.2019.2903822

This paper studies the dynamic resource allocation (DRA) problem for LTE-based vehicle-to-infrastructure networks, where the goal is to minimize the total power consumption (TPC) in the downlink, subject to both power constraints and rate requirements. Under time-varying channel conditions, the TPC minimization takes the form of a discrete-time sequence of NP-hard combinational optimization problems. To solve these sequential problems, we propose a novel two-stage algorithm, named as DRA and precoding algorithm (DRA-Pre). In the first stage, the resource allocation problem (i.e., pairing of vehicle users to roadside units, and subcarrier allocation) is solved by applying the multi-value discrete particle swarm optimization method. This approach takes advantage of the channel correlation by exploiting the relationship between resource allocation solutions in adjacent time slots, which can improve the TPC performance. In the second stage, the precoding design problem is solved by a low-complexity algorithm, where the original problem is split into two subproblems, i.e., a rate max-min subproblem and a TPC minimization subproblem. Simulation results show that the proposed algorithm converges rapidly and significantly outperforms benchmark approaches in terms of TPC.