Intelligent NextG Cellular Networks and Applications Laboratory

Welcome to the Intelligent NextG Cellular Networks and Applications Lab. This prospective thematic data science lab intends to focus on conducting fundamental and applied research into machine learning for NextG cellular networks and applications, supported over mobile, edge, and cloud computing resources. On the side of NextG cellular network and compute management, the lab will study learning approaches to disaggregation of network processes, their orchestration onto distributed compute resources, and intelligent configuration and control of the cellular radio access network at different layers and timescales. On the side of networked applications, the lab will research learning approaches to enabling mobile, edge and cloud computing supported applications connected via low latency and high throughput NextG cellular networks. These will include cloud-supported robotics, augmented reality/virtual reality, gaming and visualization, real-time applications. Altogether, this effort will bring together areas such as statistical learning theory, resource management and control of distributed systems via reinforcement learning, and federated and meta learning across a variety of tasks and agents in support of emerging real world applications with broad societal impacts. New design paradigms, collected data sets, and case studies will be disseminated broadly. Projects and findings will be leverage in the creation of educational material and infused in the classroom through course development.

Thematic Area of Data Science

The increased softwarization and disaggregation of cellular and compute systems, together with the networked sensing and control applications that utilize them, strongly suggest that data science will play an increasingly important role in the management, control and analysis of wileless infrastructures. This trend constitutes a significant shift from current technology, and it poses important and timely engineering challenges. Data science is at the forefront of this revolution, with pressing needs at every level. To address such challenges, the lab will bring together areas such as statistical learning, resource management and control of distributed systems via reinforcement learning, and federated and meta learning across a variety of tasks and agents in support of emerging real-world applications with broad societal impacts.

In addition, access to high throughput and low latency wireless communication and computation resources will facilitate a variety of sense–communicate–compute–control loops supporting a plethora of data-driven applications. Areas such as cloud-robotics, smart power grid management, autonomous transportation and remote healthcare will fundamentally depend on data science methods that are enabled by this merger of communication and computing. The lab has the potential to become a catalyst for collaborative research efforts aimed at bringing this broad vision to reality.

A Multifaceted Effort

The building of a multi-disciplinary team is a key step in being able to pivot and ready wireless infrastructures for the traffic of tomorrow. Data science permeates this effort at many levels. Machine learning has been heralded as a key technology to design the physical layer of NextG systems, which are increasingly sophisticated, with multiple antennas, coexistence, intelligent surfaces, and multiple bands. As mentioned above, higher layers of NextG cellular networks also need to be flexible, with a built-in ability to adapt to changing operational conditions. To achieve this goal, the lab will research learning approaches to disaggregation of network processes and their orchestration onto distributed compute resources. Algorithmic development and the intelligent control of cellular radio resources at different layers and timescales will be key aspects of our envisioned effort.

We look forward to working together in establishing robust and resilient communication systems attuned to the digital landscape of tomorrow through theoretical and applied data science.