Research helping to make this a smarter world
A monitor on Professor Weiru Liu’s desk shows an image of Spiri, a programmable, autonomous flying robot equipped with sensors, GPS, cameras, Wi-Fi and cloud support.
Director of Research at the Knowledge and Data Engineering (KDE) Cluster within the School of Electronics, Electrical Engineering and Computer Science, Weiru joined Queen’s in 2004 and was offered a Chair in 2010. She obtained her PhD from the Department of Artificial Intelligence in 1995, one of the world-leading departments in AI, at the University of Edinburgh.
Her research is focused on the next generation of autonomous systems which could be deployed in hospitals, in manufacturing, schools, care homes and elsewhere.
She points to Spiri. ‘An autonomous system such as Spiri must possess intelligent capabilities including being able to constantly perceive its environment, correlate heterogeneous information that’s often uncertain and inconsistent, reason in real-time, make decisions and take actions if necessary.’
She gives an example. ‘Imagine a situation where multiple sensors are used to retrieve signals from an underwater target, its location to be determined. Uncertain sensor data fusion is to locate the target precisely and eliminate noise and error from sensors or other sources.
‘Let’s further imagine a large, intelligent autonomous system with many clusters of sensors, each cluster monitoring one aspect of the actual world. Correctly managing and correlating such dispersed information from all these sources is crucial for determining what is actually happening and what are the right decisions to be made.
‘Now if we extend the concept of sensors beyond normal physical devices and consider other digital media forms, such as emails and tweets, then autonomous systems can be developed for all kinds of real-worldAutonomous systems can be developed for all kinds of real-world scenarios. scenarios.’
At KDE, one research theme is to develop an intelligent system to detect fraudulent insurance claims. There is a proposed £1m Invest NI project in collaboration with the insurance giant Allstate, with which a Knowledge Transfer Partnership has been formed. ‘If we can catch an extra ten per cent of fraud by utilising information from all sources, that will mean savings of tens of millions of pounds.’
Since 2001, KDE research has informed more than 20 projects involving major UK and EU funding bodies including EPSRC, TSB, The Royal Society, the Royal Academy of Engineering, EU FP7 and Invest NI. Current projects include being part of the ESRC-funded Northern Ireland Administrative Data Research Centre and an EPSRC-funded project under the programme of autonomous intelligence systems. Partners include BAE Systems, Sellafield, the National Nuclear Laboratory and the UK Space Agency.
‘We are world leading in several key areas. The first is Knowledge and Information Fusion under Uncertainty. The second is Data Extraction and Integration. We’re doing a tremendous amount of work on extracting data from deep web, from social media and unstructured texts, making sense of big, dispersed, unstructured data.’
And she is enthusiastic about the development of what is known as the Smart City. ‘This is about trying to improve the quality of life of citizens, improving the services they get – integrating transport, energy, water and so on – and providing a better-connected future. Data Science and Autonomous Systems are the key research areas KDE has been strengthening and they’re the underpinning technologies for Smart Cities.
‘Big cities like Dublin, London and Manchester have started experimenting in certain areas, such as public transport and tourist information. People are using mobiles and apps to get information quickly. In five or ten years' time, this will be a household concept. The impact will be huge.’