There can be a lot of data captured and tracked in facilities like hospitals and factories to help keep the important work that’s going on continue smoothly.
But if that data isn’t being transmitted in a timely fashion, in context, and with care, the data doesn’t get turned into useful decision making; it just becomes noise.
For instance, a smart factory might have sensors tracking humidity, temperature, and output. Having the sensors in place is the first step, but there are challenges to think about like when and where the data should be delivered and how often it should be refreshed so that it can be useful to the factory manager, providing a complete picture of what is going on and what actions need to be taken.
Ning Lu is a researcher working at the intersection of machine learning and networking, helping to ensure future communications networks are serving their operators well and allowing places like hospitals and factories to run more efficiently.
Lu is an Associate Professor in the Department of Electrical and Computer Engineering.
“We need a better design for routing and scheduling this information, and we can use advanced machine learning to do it, “he says. “For example, with the machine learning, we can optimize the network as a whole so the network of sensors can monitor what's going on and help make decisions and make changes to adapt to the environment automatically and intelligently.”
Lu and his research team are developing AI models that learn from multiple data sources, while keeping the raw data private. Their work is also enabling wireless networks to process data as they are transmitted in order to speed up computation, as well as redefining network performance metrics to assess learning efficiency rather than just speed. These innovations will make mobile, cloud, and edge networks more adaptive and efficient in industries such as telecommunications, smart cities, and health care.
“I'm one of the earliest researchers in this exciting field, and previously worked on allowing cars to talk to each other,” he says. “We did a lot of research improving the network protocol design and, during my PhD, we submitted a proposal of our protocol design to a global innovation challenge hosted by a vehicle parts vendor supplier in France. We made it to the final and won second place.”
Whether on the road or in the workplace, there are broad and growing applications for Lu’s work. That’s why his Tier 2 Canada Research Chair in Future Communication Networks position was recently renewed, allowing him to further build on the last five years of research in this field.
“What I'm doing for the second term is making the best use of machine learning and trying to implement machine learning at the network edge, which is where the device is connected to the Internet,” he says. “We want to make the edge devices more intelligent so the network can adapt to demand variability and the traffic load so they can do self optimization, and also so they can work together to train a machine learning model.”
Lu also intends to continue seeking graduate students to support the research of his lab, the Connected Intelligence Research Lab (CIRL), and to attract outside support for his research. His past work has been supported by NSERC Discovery and Research Tools and Instrument grants, as well as from the Canada Foundation for Innovation.
He’s currently pursuing other government projects including one related to improving Canadian rural 5G connectivity. Learn more about Ning Lu’s work and publications on his faculty profile page.
