انتخاب وسایل نقلیه سیار قابل اعتماد در شبکه های ۵G
Abstract
1- Introduction
2- Related works
3- System model and problem statements
4- Design of the MLCD scheme
5- Experiments and simulations
6- Conclusions
References
Abstract
Recently, the evolving of 5G networks is foreseen as a major driver of future mobile vehicular social networks (VSNs), which can provide a novel method of code disseminations. Based on this concept, vehicles can be used as code disseminators. That is, infrastructures of a smart city can be upgraded by receiving updated program codes that are disseminated by vehicles in the VSNs. Specifically, vehicles in the 5G network are hard to be managed. Under this domain, safety of program codes is a key challenge. Meanwhile, improving coverage of program codes is also challenging. However, arranging plenty of vehicles as code disseminators will incur large costs of the ground control station (GCS). Therefore, by utilizing machine learning methods, this paper proposes a “Machine Learning based Code Dissemination by Selecting Reliability Mobile Vehicles in 5G Networks” (MLCD) scheme to choose vehicles with higher reliable degree and coverage ratio as code disseminators to deliver code with lower costs. Firstly, reliable degrees of vehicles are calculated and selected to improve safety degree of code disseminations. Secondly, vehicles with higher coverage ratio are preferred to promise code coverage. Thirdly, machine learning methods are utilized to select vehicles with both higher coverage ratios and reliable degrees as code disseminators with limited costs. Compared to random-selection and coverage-only scheme respectively, the MLCD scheme can improve safety degree of code dissemination process by 83.6% and 18.86% in 5G networks, and can improve coverage ratio of updated information by 23.16%. Comprehensive performances of the proposed scheme can be improved by 80.56% and 17.25% respectively. Future works focus on improving code security in 5G networks by more advanced and suitable machine learning methods.
Introduction
Recently, the increasing demands and requirements of vehicular social networks (VSNs) poses challenges for the 5G networks [1,2]. The 5G mobile system is complex and is hard to be managed. Therefore, a hot research aspect where a VSNs systems’ reliability to enhance and promise a safety environment via the machine learning schemes (ML) becomes more prudent and significant. In the 5G mobile networks, because of the characteristics of large-number and mobility, the vehicles have been widely utilized in varies of situations in smart city. Large number of vehicles can act as mobile data mules to deliver data or to collect data from plenty of sensors that spread across virtually every smart city streets (garbage cans, streetlamps, bus stations, etc., with low costs [3]. Moreover, it is convenient for those mobile vehicle to transmit codes to hundreds of thousands of sensor derives in a smart city, especially in dynamic city networks, where there are plenty of temporarily deployed sensing devices that are not connected to the Internet. [4,5]. Researches above have developed and provided more convenient services to the social life.