Yufan, who won the University Program Award with his ‘All-in-one self-adaptive computing platform for a smart city’ was one of just 14 award winners out of the 165 qualified competition entries from 35 countries.
The work was partially supported by the Embedded and Intelligent Systems Laboratory in our School of Computer Science and Electronic Engineering, as well as the Engineering and Physical Sciences Research Council’s National Centre for Nuclear Robotics and EDGE projects.
His project involved deep neural networks which are a key technique in modern artificial intelligence (AI). It is a popular application used in smart city technologies, where sensors provide information about traffic flow and people movements to provide benefits such as avoiding congestion. However, this technology places heavy demands on the embedded computers which deliver the system.
Yufan’s project addressed this problem by designing a flexible video processing framework which is both faster and more energy efficient.
“I am hoping my proposed solution will be able to open a new research direction that could help resolve the challenges between the workloads of AI tasks and constraints of hardware resources,” explained Yufan. “This is not only a challenge for smart cities technologies, but also highly relevant for many more application domains, such as satellites, spacecraft, robots and self-driving cars. I hope that my design can be applied to more scenarios to solve more real-world problems, not just a simulation or theory.”
Professor Klaus McDonald-Maier, Director of Research and Head of the Embedded and Intelligent Systems Laboratory, said: “Having encouraged Yufan to participate in this highly competitive challenge, makes it especially rewarding to see that this work has win this award. Dr Xiaojun Zhai and I, as his supervisors, were delighted to hear this outstanding news and are we are very proud of supervising such an excellent student.”