On Optimal Device-to-Device Resource Allocation for Minimizing End-to-End Delay in VANETs

In vehicular ad hoc networks (VANETs), the IEEE 802.11p is a popular and standardized protocol for communications among vehicles and infrastructure (e.g., roadside units). However, because of a limited communication range and the randomly access nature of the carrier sense multiple access with collision avoidance (CSMA/CA) mechanism, the end-to-end delay could be high due to both store-and-catch-up (SAC) delay when the network is temporally disconnected and the channel access contention delay. In this paper, we propose a new method based on fifth-generation device-to-device (D2D) technology to improve the delay performance of VANETs. The basic idea is that direct D2D-based communications among vehicles remove the contention delay and can support longer distance. Specifically, we design a hybrid system with both D2D- and IEEE 802.11p-based communications, where the D2D links are controlled by the cellular base stations (BSs) in the overlay scheme. Each vehicle periodically checks its packet lifetime and requests the BSs to establish D2D links, if needed. The optimal resource allocation problem at the BSs is to select optimal receiver vehicles to establish D2D links and assign proper channels for them so that the total delay is minimized. The problem is equivalent to a maximum weighted independent set problem with dependent weights (MWIS-DW), which is NP-hard. To calculate the weights, an analytical approach is developed to model the expected end-to-end delay. Furthermore, we propose a greedy-based algorithm to solve this problem and develop a theoretical performance lower bound for the algorithm. The effectiveness of the algorithm under various scenarios is evaluated through simulations.