Section: New Results

Smart City and ITS

Participants : Indra Ngurah, Christophe Couturier, Rodrigo Silva, Frédéric Weis, Jean-Marie Bonnin [contact] .

In the last years, we contributed to the specification of the hybrid (ITS-G5 + Cellular) communication architecture of the French field operation test project SCOOP@F. The proposed solution relies on the MobileIP family of standards and the ISO/ETSI ITS Station architecture we contributed to standardize at IETF and ISO. On this topic our contribution mainly focussed on bringing concepts from the state of the art to real equipments. For the last year of the SCOOP@F part 2 project, we took part to the performance evaluation process by providing a test and validation platform for IP mobility protocols (MobileIP, NEMO) and IPsec cyphering. This platform allows us to identify the performance limits of current implementation of mobility and security protocols. Moreover it spotted implementation incompatibilities between the open source implementations of theses protocols (namely UMIP and StrongSwan) and helped the industrial partners of the project to identify associated risks.

InDiD is the logical follow up of SCOOP@F part 2. This 3.5 years long European project (mid 2019-2023) aims at testing ITS applications on a large scale national deployment of connected vehicles and infrastructure. This version of the project specifically complex use cases (so called day 1.5) and urban application. For the beginning of this project, we proposed several innovative use cases. Our "Backward cartography update" scenario has been selected as a priority candidate for implementation. In line with the collaborative approaches of EASE, we propose to use vehicles' observations to inform other vehicles and/or a cartography server about differences between the digital map and the reality.

We also want to explore the benefits of new capabilities of upcoming communication technologies to enrich the interactions between vehicle and smartphones or wearable devices. We defined an architecture for both localisation and communication with vulnerable users (workers in road and construction works). Short range communications between dangers (maneuvering construction vehicles) and workers rely on the advertisement feature of Bluetooth Low Energy (BLE). This connectionless communication mode enables for easy direct communication between any node in the neighborhood. It is inspired from the ITS-Station communication standard and we aim to integrate our work into future versions of the standards. Another contribution in this project aims at enhancing the localisation precision in harsh conditions. Recent version of radio communication standards (eg. Bluetooth 5.1 or 802.11ax) now integrate intrinsic real time localisation primitives giving information such as Angle of Arrival (AoA), Angle of Departure (AoD) or distance evaluation based on Time of Flight (ToF) measurements. We started to study how to merge this information with other localisation evidence sources and how to structure a collaborative framework to share it with other objects in the environment. This early works opens the doors to many other works in the future.

The development of innovative applications for smart cities has also been made possible by the rise of Internet of Things and especially the deployment of numerous low energy devices. The collection of the huge amount of data produced by all these piece of hardware become a challenge for the communication networks. In smart cities, the mobility of vehicles can be used to collect data produced by connected objects and to deliver them to several applications which are delay tolerant. The Vehicular Delay Tolerant Networks (VDTN) can be utilized for such services. We designed DC4LED (Data Collection for Low Energy Devices): a hierarchical VDTN routing which takes advantage of the specific mobility patterns of the various type of vehicles. It provides a low-cost delivery service for applications that need to gather data generated from the field. The idea is to propose a simple routing scheme where cars, taxis, and buses route data hierarchically in a store-carry-forward mechanism to any of the available Internet Point-of-Presences in the city. We compare using simulation tools the performance of DC4LED routing with two legacy VDTN routing schemes which represent the extreme ends of VDTN routing spectrum: First-contact and Epidemic routing. It show that DC4LED has much lower network overhead in comparison with the two legacy routing schemes, which is advantageous for its implementation scalability. The DC4LED also maintains comparable data delivery probability and latency to Epidemic routing.

The situational viewing and surveillance in cities is one such category of applications which can benefit from various networking solutions available to transport images or data from installed sensor cameras. We explore how our DC4LED mechanism can be used to for a city-wide image and data collection service. We study the networking performance in terms of increasing image sizes that can be transported with respect to varying vehicular density in city. We focus mainly on two technologies for sensors to vehicles communications: ZigBee and ITS-G5. We show that, surprisingly such very simple mechanism could meet the requirements of multiple services.