Section: New Results

Smart cities architecture

Participants : Valeria Loscri, Nathalie Mitton, Riccardo Petrolo, Nicola Zema.

Smart City represents one of the most promising and prominent Internet of Things (IoT) applications. In the last few years, indeed, smart city concept has played an important role in academic and industry fields, with the development and deployment of various middleware platforms. However, this expansion has followed distinct approaches creating, therefore, a fragmented scenario, in which different IoT ecosystems are not able to communicate between them. To fill this gap, there is a need to re-visit the smart city IoT semantic and offer a global common approach. In order to allow cities to share data across systems and coordinate processes across domains, it is essential to break these silos. A way to achieve the purpose is sensor virtualization, discovery and data restitution. This last year, the FUN team has lead several investigations in this direction.

We have looked at the heterogeneity of devices and network technologies under a different perspective by not perceiving it as a limitation but as a potential to increase the connectivity in a smart city [1] . We propose a new generation of network nodes, called stem nodes, based on the innovative idea of 'stemness', which pushes forward the well-known self-configuration and self-management concepts towards the idea of node mutation and evolution. We also deployed prototypes that demonstrate the stem-node architecture and basic operations in different hardware platforms of common communication devices (an Alix-based router, a laptop and a smartphone)

In [7] , we illustrate semantic interoperability solutions for IoT systems. Based on these solutions, we describe how the FP7 VITAL project aims to bridge numerous silo IoT deployments in smart cities through repurposing and reusing sensors and data streams across multiple applications without carelessly compromising citizens' security and privacy. This approach holds the promise of increasing the Return-On-Investment (ROI), which is associated with the usually costly smart city infrastructures, through expanding the number and scope of potential applications.

To this purpose, [21] browses the semantic annotation of the sensors in the cloud, and innovative services can be implemented and considered by bridging Clouds and Internet of Things. Things-like semantic will be considered to perform the aggregation of heterogeneous resources by defining the Clouds of Things paradigm. We survey the smart city vision, providing information on the main requirements and highlighting the benefits of integrating different IoT ecosystems within the cloud under this new CoT vision. This paper also discusses relevant challenges in this research area.

Going further, we also presented [22] a first implementation of this federation: a federation of FIT IoT-LAB within OpenIoT. OpenIoT is a middleware that enables the collection of data streams from multiple heterogeneous geographically dispersed data sources, as well as their semantic unification and streaming with a cloud infrastructure. Future Internet of Things IoT-LAB (FIT IoT-LAB) provides a very large scale infrastructure facility suitable for testing small wireless sensor devices and heterogeneous communicating objects. The integration proposed represents a way to reduce the gap existing in the IoT fragmentation, and, moreover, allows users to develop smart city applications by interacting directly with sensors at different layers. We illustrate it trough a basic temperature monitoring application to show its efficiency.

So, once all city network and infrastructure are set at the same level thanks to the above mentioned approaches, they can go further and offer additional services. An example of them is navigation[11] as also described in "Localization" section. Another example is to make use of the urban bikes [19] . Indeed, besides the growing enthusiast provoked by bicycles in smart and green cities and the benefit for health they bring, there still exists some reluctance in using bikes because of safety, road state, weather, etc. To counter-balance these feelings, there is a need to better understand bicycle users habits, path, road utilization rate in order to improve the bicycle path quality. In this perspective,in this paper, we propose to deploy a set of mobile sensors on bicycles to gather this different data and to exploit them to make the bike easier and make people want to ride bicycles more often. Such a network will also be useful for several entities like city authorities for road maintenance and deployment, doctors and environment authorities, etc. Based on such a framework,we propose a first basis model that helps to dimension the network infrastructure and the kind of data to be real time gathered from bikes. More specifically, we present a theoretical model that computes the quantity of data a bike will be able to send along a travel and the quantity of data a base station should be able to absorb.We have based our study on real data to provide first numerical results and be able to draw some preliminary conclusions and open new research directions.