SMILeS – Smart, Mobile, Intelligent and Large Scale Sensing

Project Duration: 
07/2015 - 06/2018
Project Code: 
NORTE-01-0145-FEDER-000020
Funding: 
1,275,000€ (240,000€ for CRACS)
Funding Entity: 
CCDRN
PI
CRACS Member: 
Luís Lopes
Research team
from CRACS: 
Fernando Silva
from CRACS: 
Manuel Eduardo Correia
from CRACS: 
Ricardo Rocha

Transportation systems are a fundamental piece of society development. They provide mobility for people and goods and are a major factor of economical competitiveness. While there have been significant developments in technology, there are major challenges still to be met in what concerns the integrated management of such systems. Our reliance on fossil fuels, for example, introduces many environmental problems related to the emission of carbon dioxide and other noxious gases with immediate, local impact on human health and more globally in the ecosystems. Nowadays, a major shift in the energy paradigm is taking shape, with the increasing importance of renewable sources and the use of hybrid or electric cars. The infrastructure required to support such a paradigm change is a considerable technological challenge as it relies on the integration with other core resources like electrical power networks. Another important problem is that of resource efficiency and planning, to minimize waste, time and environmental footprint, and maximize profit. Security, also, is of critical importance, especially for governments, where an intelligent transportation infrastructure is a fundamental asset though care must be taken to protect the privacy of citizens.

For many decades the development of Intelligent Transport Systems has been conditioned by the availability of adequate technology. Today, however, vehicles have integrated hardware and software that enables their use for seamless sensing in a transportation infrastructure. Vehicles now use wireless communication and have a substantial computational power, with current models featuring hundreds of small sensors connected to several computing nodes, running Real‐Time Operating Systems and capable of significant online processing and supporting infotainment. Intravehicular wireless sensor networks are also becoming common to minimize cabling complexity. Thus, vehicles can be used as ideal sensing probes in a Transportation System, with the major advantage that there is no need to build a specific sensing infrastructure and is highly adaptive. Preliminary experiments have been done to infer weather/road/traffic conditions using position and in‐vehicle sensors. Moreover, if organized in mobile networks, reductions of this collected data could be exchanged between cars, or uploaded to a more centralized infrastructure, to provide for a highly adaptive management system and more adequate feedback/actuation. Environmental important applications would also be possible under this scenario, like real‐time measurements of pollution levels. In‐vehicle video cameras could also be used to record real‐time footage of accidents or other unusual events that together with location can be used for planning assistance or tracking unfolding events.

The infrastructure can also be used for sensing at the cost of instrumenting it from scratch and being less dynamical. Inductive loops have been used to count the number of cars per time and detect speed excesses. MAC detection, collision avoidance systems, dynamic traffic lights, plate recognition, automatic tolls are common examples that require support from a base infrastructure. An important improvement in environmental impact would be making road lights activated by movement sensors.

This is possible with new LED luminaires and would significantly reduce energy consumption in most roads during the night and diminish light pollution a problem that has an enormous impact on wildlife and that raises increasing concerns on human health.

From a research point of view, there is still much to do to produce systems capable of the level of integration such as the one just described. The problem is more complex since it requires a multidisciplinary approach, with expertise from several areas of Computer Science and Engineering, e.g., low level protocols for communication, embedded systems, sensors and wireless sensor networks, middleware, distributed computing and storage, data dissemination and aggregation, data mining and security and privacy, to name a few. Together, the INESC‐TEC units collaborating in this proposal possess a considerable know‐how in the aforementioned areas while having a tradition of internal collaboration and sharing of knowledge. For this reason we are in a position to make a standing contribution in this area by significantly advancing the state‐of‐the‐art and provide a fundamental science package to drive innovation and competitiveness in the local industry, especially that whose business is automobile components and systems.