ARAGORN · Adaptive Reconfigurable Access and Generic interfaces for Optimisation in Radio Networks

The primary aim of the ARAGORN project is to research and develop a Cognitive Resource Manager (CRM) that aims to ensure that efficient use is made of both node-local and shared resources in a collaborative wireless system. These include, for example, local energy consumption and shared use of available bandwidth. In order to achieve this, ARAGORN will develop a range of standardised interfaces, using which the CRM will be able both to obtain information from and to update the configuration of each layer of the protocol stack, including at the application layer. Given this, the CRM will seek to optimise cross-layer and inter-node performance using multi-dimensional optimisation algorithms provided by partners from the machine learning and artificial intelligence communities. The proposed approaches have been selected because of the need to synthesise a relatively low-dimensional response from the many potential dimensions that could affect the optimal, all in the presence of incomplete and out-of-date information.In contrast to other cognitive radio research, the ARAGORN project does not focus on dynamic spectrum access networks but aims to add real cognition to mobile devices, enabling them to take reasonable decisions autonomously.The main expected outcomes of the project are a solid basis of theoretical and architectural work for next generation cognitive radios and networks and a working prototype implementation. The prototype will apply the techniques developed and prove their practical feasibility working in one of the ISM frequency-bands. As key components of ARAGORN, exploitable interfaces and the Cognitive Resource Manager Framework will be implemented.The high quality consortium possesses a unique combination of expertise from across the wireless, software and AI communities, and forms a well balanced team of academic and industrial partners, some of whom already participate in pertinent standardisation work.