Age of Information Modeling and Optimization for Fast Information Dissemination in Vehicular Social Networks

Autonomous vehicles (AVs) with advanced communication, computing, and control capabilities will provide not only a convenient means of transportation but also an emerging platform for real-time social communications and networking. Thereby, it is crucial to enable timely exchange of information over the dynamic cyber-physical-social system enabled by the AVs. In this paper, we consider age of information (AoI)-centric real-time information dissemination over vehicular social networks (VSNs), where the AoI captures the freshness of received data packets. We first propose a mathematical framework based on the mean-field theory (MFT) to analyze the network AoI (NAoI) of VSNs, namely AoI for the AV that lastly receives the information update in the network. The proposed analytical framework considers both the social features of vehicular networks, which are characterized using the scale-free network theory, and the underlying wireless communication process to evaluate the NAoI. Then, we further consider joint optimization of the information update rate at the source node and the transmit probabilities at the AVs for minimization of the average peak NAoI (PNAoI), i.e., the time average of peak values occurred in the evolution of the NAoI, which is solved via a novel parametric optimization scheme. Both analytical and simulation results show that compared with several baseline schemes, the proposed scheme can exploit the scale-free feature of the VSNs to significantly lower the average PNAoI by up to 96$\%$.

I. Introduction

Thanks to advances in mobile communications, mobile sensing, and autonomous driving technologies, the vehicular networks are evolving towards a smart cyber-physical-social system [1], [2]. It is expected that future autonomous vehicles (AVs) will enable not only reliable connected driving control [3], [4], but also versatile social communications such as information and content sharing among passengers and human drivers [5]. For the latter communication paradigm, social relations/ties among onboard humans can significantly impact the information dissemination process in vehicular networks. For example, AVs may prefer to communicate and exchange information with other AVs that share close social connections [6]. However, how to exploit existing social relations to improve the performance of vehicular networks has been a new research challenge.