Using AI and machine learning, Professor Karim Seddik is engineering energy-efficient ways to keep us all connected
Our world runs on wireless networks. Millions of people and devices require strong connections to communicate phone calls and fire alarms.
But how do networks prevent calls from dropping on the airwaves? The secret is communications engineers: the brains behind the algorithms that keep our world connected.
Karim Seddik, professor of electronics and communications engineering and associate dean for graduate studies and research in the School of Sciences and Engineering, researches how we can apply AI and machine learning to keep calls online while managing interference between the millions of signals networks handle every day.
Staying Connected
Wireless communications work like this: When you make a call, send a message or use mobile data, your device sends a signal in the form of a radio wave to a base station, which processes this data and routes the signal back to its intended destination. With millions of devices connected to mobile networks, this process requires significant amounts of energy, and base stations can get overloaded by the high demand.
“One of the major challenges of mobile operators worldwide, not just in Egypt, is energy consumption,” Seddik explained. “The energy bill for mobile operators is one of the most expensive parts of their operation, so energy efficiency is key.”
The challenge is that managing this energy use is not yet automated. Mobile operators have to track and monitor thousands of cells across a network, with varying levels of demand. Traditionally, communications engineers have relied on classical mathematical models to find an optimal solution. But this has proven difficult with the growing scale of modern mobile networks.
To address this issue, Seddik researches how to make base stations energy-efficient by using machine learning to manage these cells. He and his team, including AUC students, worked with Vodafone Egypt to apply machine learning algorithms on real network data from New Cairo to reduce energy consumption and effectively manage network congestion. “Using AI and machine learning, we can have automated, real-time, data-driven tools that are built into the network. We have tested some of these solutions, with machine learning at the core of our algorithm, to make sure that the quality of service and user experience are not negatively affected,” Seddik said.
“Communications, it's everywhere. You cannot live without your cell phone or access to the internet”
Their research created an algorithmic tool to manage a network’s energy load. For example, if there is a smaller number of users on a certain cell, the tool could choose to offload those users to adjacent cells and turn off the current one to save energy. This load-balancing could determine whether your phone has service in crowded areas.
Networks need to support multiple types of traffic, from enhanced broadband to devices communicating without human initiation. As a result, communications engineers use multiple access methods to prevent signal interference on increasingly high-traffic wireless networks. This is crucial to keeping people connected and ensuring reliable communications for critical systems, such as emergency scenarios.
Seddik is investigating an access method called Non-Orthogonal Multiple Access (NOMA) that could save resources, including time and bandwidth for networks. NOMA allows multiple users to send signals simultaneously with controlled interference. Seddik’s research involves programming intelligent reflecting surface algorithms to reduce signal interference and leverage the reflected signal to enhance NOMA performance.
“Using AI and machine learning, we can have automated, real-time, data-driven tools that are built into the network. We have tested some of these solutions, with machine learning at the core of our algorithm, to make sure that the quality of service and user experience are not negatively affected”
“Communications, it's everywhere. You cannot live without your cell phone or access to the internet,” Seddik said.
Energy Efficient Future
If you want to watch the future of communications, Seddik says to keep an eye on virtual reality and online gaming. These fields require large amounts of data to be transmitted without delay, and the next step is the expansion of AI-native 6G networks as the new backbone of communications.
“Right now, we are very satisfied with megabytes per second, but in the future, we will be talking about gigabytes per second — huge data rates,” Seddik said.
Open radio access networks (O-RANs) are another potential evolution in the business of telecommunications. Mobile networks rely on vendors for network equipment, which is often sold as “black boxes” that limit operator access to network data that could optimize energy use. O-RANs allow operators to build a network combining hardware and software from different vendors — like a Lego set — to improve network performance and energy efficiency. “This could open up new access for startups to specialize in different equipment, diversifying the market for new ideas and investment,” said Seddik.
Even with these changes in his field, what remains important for Seddik is continuing to build the next generation of Egyptian engineers. “I’m always proud of my students,” Seddik said. “When I see that some of them right now are professors abroad and researchers in top companies in the U.S., Europe and everywhere, I feel that I have done something, and this is what keeps me going.”