innovacity context & background

INNOVACITY project context and background

Cities present a clear trend to concentrate most of the world’s population in the years coming. The forecast is that about 65% of world population will live in urban spaces by 2040 [1] [2] [3]. City numbers are impressive: 30 cities will have more than 10 million inhabitants and largest cities will consume 85% of world’s energy and produce 80% of the planet’s waste [4].

A Smart City (SC), by definition, addresses the urban problems looking for citizen quality of life improvement, promoting sustainability and engaging citizens through transparent government decisions [5].

Smart city projects need to innovate. Innovation must be present due to the complexity, multidimensional characteristic, size and multidisciplinary nature of the problems involved. Most of current legacy smart city solutions do not present the set of required capabilities or are not capable to cope with the ever-increasing functionality complexity required by huge, efficiently managed and multipurpose urban spaces [3] [6].

In general terms, innovation can be achieved in smart city projects by using more dynamic and intelligent networks supporting new context-adaptable and intelligent application and systems.

Intelligent, dynamic and context-aware characteristics for smart city project deployment is a current research gap [7] [8] [9] [10]. Current research is extensively focusing on enforcing these characteristics for application level, communication infrastructure and physical components [11] [12] [13]. As an example, at communication level, most current networks supporting smart city projects have been deployed as static solutions that barely adapt their routes under failure. Concurrently, smart city application systems have not yet widely adopted an intelligent approach capable of dealing with actual smart city context that has a huge number of users with multiple and variable requirements.

As discussed in [9] and [3], at communication level, smart city projects require the transfer of huge volume of data captured by heterogeneous sensors over large distributed areas. In this scenario, the communication infrastructure must cope with heterogeneous communications requirements that are difficult to realize with a single or not dynamically configurable network infrastructure. Another important issue at the communication level is routing and traffic patterns variability. Routes must be defined between data source and destinations and the network infrastructure must adjust itself to the variability of communication resources demanded. Due to the variety of objectives and requirements involved in these communications such as QoS (Quality of Service) [14], QoE (Quality of Experience) [15] and SLA (Service Level Agreement) compliance, a flexible and programmable network infrastructure is the best possible approach [7].

New paradigms like artificial intelligence and software-defined networks are a potential solution for deploying innovative smart city projects as discussed in [16] [17] [2] and [3]. Artificial intelligence using machine learning techniques applied to smart city has a tremendous application potential by exploring, for instance, a cognitive processing and management approach for networks, IoT and data [16] [3].

The effective contribution that machine learning brings to smart city application and systems comes from the fact that [3]:

● A cognitive approach allows the computation of solutions for highly complex problems with multiple requirements and objectives.

● Cognitive solutions have the capability to extract "knowledge" and "learn" from huge and dynamic volume of data and this is an actual requirement to allow more “intelligent” and easier to use applications.

● Cognitive-based application and systems have the capability to substitute, at least in part, some human tasks (like management tasks), which turn out to be difficult to realize and subject to errors when a large and sometimes unrelated volume of information is involved.

Software-defined networking (SDN) allows, in summary, the creation and deployment of programmable networks and systems [17]. In technical terms, SDN uses a logically centralized controller to program network equipment using a well know interface and protocol like Openflow. As a result, we can control network state transitions by monitoring network operational parameters and program any modification of its operational behavior in terms of packet handling and manipulation.

These operational principle and facility can be applied to smart city application and communication levels [3] [18].