Researchers at NUI Maynooth want to build a new kind of intelligence into complex engineering systems. It involves making each component of a system smart enough to respond to what others are doing, as a way of keeping everything running smoothly. A football team is the inspiration for the new research initiative, writes Dick Ahlstrom
High-performance engineering systems already involve hardware and software linked to mechanical components. In simple terms, the computer provides the intelligence to keep the machine running as planned.
Prof Douglas Leith wants to go one better, building complex systems that perform more like a football team. The idea is to produce a collection of separate subsystems that can respond dynamically to what others in the group are doing.
Prof Leith is the director of the new Hamilton Institute, which was established at NUI Maynooth with the support of funding from Science Foundation Ireland. "The Hamilton Institute did not exist before I arrived and the SFI funding is core funding for the institute," he says.
He comes to Maynooth from the University of Strathclyde, where he received Royal Society funding for his systems work. He heard about what SFI was doing and the opportunities for principal investigators and decided to apply. "It is a big move; I have a young family," he says about coming to the Republic. "It was a great opportunity. It is a chance to grow some critical mass" in this specialised systems area.
His decision was somewhat easier following a meeting with a Dubliner at a workshop in the Norwegian town of Trondheim, in 1997, at which complex systems were being discussed. He and Dr Bob Shorten found they were working on similar problems, but from different angles.
This led to the two setting up a multidisciplinary training network involving top European researchers. The Hamilton Institute grew out of the activity, he says.
Its main task is to research "the science of designing and building systems", he says. In particular, he wants to research the way subsystems interact with one another.
Examples include a car or a wind turbine, which blend a mix of subsystems - in areas such as aerodynamics, electrical, electronic and mechanical engineering - and a human interface. A traditional way to control complex systems is to use software centrally to control all of the elements, but this poses a problem, says Prof Leith.
"One of the bottlenecks in industry is there is a trend to having a higher software content." He wants to look into ways of distributing intelligence to subsystems rather than relying only on central control.
The simplest analogy is that of a football team, says Dr Shorten, who joined the institute as a senior researcher. It involves 11 players who interact in a dynamic way to achieve one aim: to score goals.
"Each of these separate subsystems is able to handle the ball," he says. "They can pass to other players; there has to be an amount of interaction between them. They have to be able to communicate and these systems have to be robust."
The team has to be able to adapt if a player is injured or sent off, changing the original game plan. There is also a "predictive" element as, say, a winger tries to reach a striker with a cross.
"We are interested in designing and building systems that have these characteristics," says Shorten.
Software development will be a key issue, but communications between subsystems is also crucial, says Leith. This means the institute will be truly interdisciplinary, involving experts in computer science, engineering, mathematics and physics.
The institute is recruiting now and also bringing in expertise from outside Ireland. Prof Bill Leithead, of the University of Strathclyde, is joining as a senior researcher and Prof Leith has established links with Prof Kumpati Narendra of Yale, an expert in systems science. "We want to keep a flow of ideas and have a strong visitor programme," says Leith.
The research team will be "working with real problems" in industry, but it will be more than a research aid for companies. As well as pursuing fundamental research on these systems, it will also carry out applied research into strategic problems associated with complex systems.