BAE Systems/Rolls-Royce/Qinetiq/Universities of Oxford, Southampton

ARGUS is a collaborative research programme between UK industry and academia addressing data fusion — how to integrate different sources of data so that the resulting information is better than any one source. The project was initiated to meet the future perceived need to develop distributed information processing technologies that can deal with the endemic uncertainty in the world. Its goal was to close the gap between the science of data fusion and the engineering needs of UK industry, where data fusion and signal processing are enabling technologies.

This posed technical and organisational challenges. Could the domain of Bayesian statistical inference that can deal with the challenges of modelling uncertain data be married with emerging agent-based techniques for controlling large data systems? If so, could the resulting technology be applied and exploited in different application domains? To address these challenges a multi-disciplinary team was built. The programme has five partners, two academic and three industrial. The academics are the Pattern Analysis and Machine Learning group at Oxford and the Intelligence, Agents, Multimedia group at Southampton. The success of ARGUS has been to demonstrate how by forging a strong team ethic, partnership between academia and industry can lead to the rapid development and exploitation of state-of-the-art technology. It has broken down the barriers between the two groups, leading to highly novel multi-disciplinary research.

The success is demonstrated by the large number of papers produced by the project (21 in internationally leading journals and conferences), and the success of the industrial partners in exploiting these technologies. For example, Qinetiq is developing a wide-area surveillance system to provide better security and monitoring of the apron side of an airport facility. Its demonstration uses the underpinning ARGUS technology to show how a next-generation system can be developed to be both flexible and robust, based upon the theoretically well-founded data fusion algorithmic components proposed within the programme.

For more information go to