OCommVehits 2017 Abstracts


Short Papers
Paper Nr: 1
Title:

Digital map structure for an Equitable and Sustainable road use

Authors:

Jorge Marto Bandeira, Margarida Coelho, Bernhard Friedrich and Carlos Borrego

Abstract: The main goal of this research is the development of a dynamic map and modelling structure to support the implementation of equitable and sustainable advanced traffic management systems. The first tangible objective is to develop a dynamic map structure which will assimilate a library of forecasting traffic models and associated externalities. A key feature of this platform is the recognition that these impacts are spatiotemporally dynamic due to the heterogeneity of activity patterns of each link. Accordingly, dynamic weights on the full range of distributional effects for each impact will be assigned in order to develop a unique link-based eco-indicator. The second objective is to enhance the potential of new sources of traffic data to improve the networks efficiency. This will be done creating new methods for managing different sources of real-time information to determine the energetic and environmental network performance. Innovative functional relationships between microscale speed patterns based on floating car data and different levels of macroscopic traffic performance scenarios will be developed. Finally, it is intended to deliver an integrated optimization platform for determining efficient traffic management measures. Based on optimization algorithms and artificial intelligence techniques, advanced traffic management strategies ATMS (e.g.. optimal flow distributions, smart road pricing systems, optimum link speeds, eco-routing information) will be evaluated.

Paper Nr: 2
Title:

Towards Robust Vehicle Connectivity using Multipath TCP

Authors:

Basavaraj Tonshal, Oleg Gusikhin and Aziz Makkiya

Abstract: Internet connectivity with wireless communication has become a preferred form of communication in the modern era. Whether it is a smartphone using a cellular connection to transfer data, or a laptop or tablet in a living room, tied into a wireless router, everyday wireless device use is almost a fact of life for an average citizen. In most wireless environments, a user ties into a known source using an appropriately ranged form of communication. For example, a laptop or smart device such as tablet user might use Wi-Fi or Bluetooth to physical media for data transfer. A smartphone might use a longer range signal such as cellular, and if the laptop is using the smartphone as a mobile hotspot, then the combined data-provision system might use both local and long range wireless communication. With the proliferation of smartphones, consumers are used to ubiquitous connectivity. They demand same type of connectivity user experience in the vehicle. Internet connectivity in the vehicle enables critical features such as data analytics, over the air update and connected user experience such as traffic update and media streaming. Vehicular cellular data services pose somewhat of a problem under this paradigm, because the vehicle tends to move at speeds that quickly bring the vehicle into and out of range of various available networks. The vehicle can even easily travel into areas of no cellular coverage, and even when cellular coverage is available it may not be a preferred form of data communication because of cost constraints. It’s worth noting that computing in vehicle is highly distributed in nature, with various Electronic Control Modules (ECU’s) controlling a specific component such as, a seat electronic module sensing and actuating a seat module and powertrain control module senses and controls powertrain. These electronic component modules generate huge amount of data and they use a multi-master Control Area Network (CAN) to receive and send data between the modules. Collecting the data that is being transferred between modules along with the diagnostic and internal module data enables for supporting diagnostic and prognostic analysis of vehicle. To support data analytics, ability to transport data efficiently at an optimized cost is highly critical. Vehicles have multiple network interfaces for cloud connectivity, making them multi-homed endpoints. Internet network connection fidelity and effective utilization of multiple interfaces are important enablers of automotive connectivity to address the above described use cases. Mobile data traffic has grown 4000 times in the last 10 years and going forward with proliferation of connected cars and autonomous vehicles, the rate of growth in mobile data without a doubt can only increase. Another important statistics that emerges is the offloading of cellular data on to Wi-Fi, 51% of mobile data was offloaded on to the Wi-Fi in 2015. This trend of offloading cellular data is going to go up. Hence, integrating unlicensed Wi-Fi and licensed cellular becomes a key enabler for cellular offloading to work seamlessly in vehicles. TCP at the transport layer does not support simultaneous usage of multiple network interfaces for data transfer. Vertical handover, maintaining of a single TCP connection across many interfaces, has been attempted before by several methods such as Mobile IP and Steam Control Transmission Protocol (SCTP). Mobile IP (MIPv6 for IPv6) provides mobility by addressing handover at the networking layer, and is less effective than transport layer handover. This is because TCP’s congestion control becomes ineffective using Mobile IP and this is problem when binding with cellular interface as it tends to be more dynamic. SCTP does not offer compatibility with TCP and middle boxes simply discard the packets. Dedicated Short Range Communication (DSRC) is another communication protocol used in vehicles. DSRC is the most reliable high speed vehicle based communication technology. This leads to addressing multi-homing at the transports layer. It’s worth noting that 90% of the applications todays use TCP, so addressing effective utilization of multiple interfaces at the transport layer using Multipath TCP provides the best solution As DSRC supports conventional TCP-IP connections, Multipath TCP can conceivably be adopted to provide roaming between cellular, Wi-Fi and as well as DSRC [3]. Multipath TCP ensures smooth handovers between DSRC, Wi-Fi and cellular hotspots, which is not possible with regular TCP connection. This presentation discusses an innovative In-vehicle telematics platform utilizing Multipath TCP that we developed to address car connectivity issues. Our platform can effectively use all the available in-vehicle network interfaces like Wi-Fi, cellular, Bluetooth and DSRC, to effectively schedule data transfer over available network interfaces based on application networking intent. The architecture also provides solution to the critical problems like intermittent network availability and power consumption. The modular architecture of the platform allows easy implementation and usage.