Search

To operate intelligent vehicular applications such as automated driving, machine learning, artificial intelligence and other mechanisms are used to abstract from information what is commonly referred to as knowledge. Defined as a state of understanding obtained through experience and analysis of collected information, knowledge is promising for vehicular applications. However, it lacks a unified framework to be cooperatively created and shared to achieve its full potential. This paper investigates on the meaning and scope of knowledge applied to vehicular networks, and suggests a structure for vehicular knowledge description, storage and sharing. Through the example of passenger comfort-based rerouting, it exposes the potential benefits for network load and delay of such knowledge structuring.

Content caching on the edge of 5G networks is an emerging and critical feature to quench the thirst for content of future connected cars. However, the tight packaging of 5G cells, the finite storage capacity at the edge, and the need for content availability while driving motivate the need to develop smart edge caching strategies adapted to the mobility characteristics of connected cars. In this paper, we propose a scheme called RICH (RoadsIde CacHe), which optimally caches content at edge nodes where connected vehicles require it most. In particular, our scheme is designed to ensure in-order delivery of content chunks to end users. Unlike blind popularity decisions, the probabilistic caching used by RICH accounts for the user mobility information that the system can realistically acquire. Furthermore, we provide a complete system architecture and define the protocols through which the different system entities can interact. We assess the performance of our approach against state-of-the-art solutions, under realistic mobility datasets and system scenarios. Our RICH edge caching scheme improves significantly the content availability at the caches and reduces the required backhaul bandwidth, with beneficial effects for both the end users and the network operators.

802.11p based V2X communication uses stochastic medium access control, which cannot prevent broadcast packet collision, in particular during high channel load. Wireless congestion control has been designed to keep the channel load at an optimal point. However, vehicles' lack of precise and granular knowledge about true channel activity, in time and space, makes it impossible to fully avoid packet collisions. In this paper, we propose a machine learning approach using deep neural network for learning the vehicles' transmit patterns, and as such predicting future channel activity in space and time. We evaluate the performance of our proposal via simulation considering multiple safety-related V2X services involving heterogeneous transmit patterns. Our results show that predicting channel activity, and transmitting accordingly, reduces collisions and significantly improves communication performance.

Lagrangian formulation of kinematic wave provides a more accurate representation than the most commonly used Eulerian formulation. Furthermore, Lagrangian representation offers a flexibility to study certain traffic phenomena (e.g. capacity drop, traffic delay, trajectory). The natural resemblance of Lagrangian representation to road traffic data collection methods (probe vehicles) also renders it suitable for applications such as traffic motoring. This paper presents a multi-class Lagrangian representation for a traffic flow consisting of cars and two-wheelers. A numerical method for solving the continuity equation is presented. We compare the results obtained with Eulerian and Lagrangian formulations., Comment: 6 pages, 10 figures, unpublished paper