WP.2 "Energy-Efficient and Location-Aware Computing"
Task T2.1 aims at developing a methodology and a new family of algorithms permitting client code running on the devices processes in fixed and/or mobile (i.e., vehicular) networks to undergo adaptation. In particular, task T2.1 is built on the theory of Diffusion Adaptation (DA) and proposes novel fixed/mobile device-compliant self-learning algorithms that allow to meet constraints posed by limited processing, memory, energy and communication resources, hence supporting the green paradigm of the project. Moreover, T2.1 investigates a distributed learning algorithm for enabling adaptation facilities within a mobile devices network. Assessed results derived from this task will enable WP4 as far as the Connected Vehicles (CV) application in T4.1 is concerned.
Task T2.2 foresees the identification of contextualization mechanisms that would permit applications in execution on devices to exchange and retrieve information from the FC/FN architecture according to a distributed, energy-efficient and context-aware approach. As such, the main goal of T2.2 is the design of efficient information-centric scalable protocols for retrieving distributed data able to exploit both physical and logical spatially-contextualized information. In particular, the research activity will embrace the development and validation of novel self-learning, fully distributed algorithms for the "intelligent" fusion of data.
Task T2.3 has as the following main objectives: i) to peruse an optimal energy-saving distributed management of the FN/FC resources; ii) to enable suitable data-transfer optimization mechanisms able to exploit wireless heterogeneous networks (HetNets ) for energy saving and performance improvement, while taking into account the constraints of fixed/mobile devices; and iii) to allow a fully adaptive allocation of the communication and computing resources of the GAUChO architecture able to cope with the unpredictable context variations of the FC units workload. In particular, in T2.3 a Context-Oriented Programming approach is investigated in order to extend language-level abstractions meant to express context-dependent behavioral variations and their run-time activation, as well as design the adaptation logic to the FC units in an organic and disciplined way. Moreover, by taking into account both the complexity of the scenarios considered within the GAUChO project and the need for an effective computing and communication resources management, T2.3 investigates suitable methodologies to dynamically reconfigure data flows and communication links according to resource availability so as to take full advantage of the envisioned context awareness.