The NSF recently awarded Rutgers WINLAB a planning grant, entitled STAIRWAI to COSMOS: Sensor-enabled Testbed for Advancing Innovative Research in Wireless+AI, whose/ goal is to  enhance the existing COSMOS testbed with tools for using AI and supporting AI applications.

  • If you wish to get email updates, please subscribe to our Google Group. You can do so via a Google Apps account or by sending an email to stairwai-announce+subscribe@winlab.rutgers.edu. This will be a low-traffic list.
  • This website/blog will have more content, including interviews with key stakeholders, publication announcements, and other information. If you're interested, please hit subscribe!

The STAIRWAI vision: This is a 2-year planning process with the eventual goal of proposing a full center on Wireless and AI. We envision two main branches of work which this new testbed could catalyze:

  • AI for Wireless (AI4W): developing AI methods for managing wireless communications and networks.
  • Wireless for AI (W4AI): developing AI applications which are “wireless aware”.

To do this we will identify the key testbed resources needed to support these two branches:

  1. Easier collection and use of datasets gathered from COSMOS, including benchmark datasets for better reproducibility.
  2. A framework for deploying and evaluating “wireless aware” AI applications, including real-time measurement execution and feedback
  3. Support research on accelerating AI for real-time wireless systems.

Our vision for the testbed is a “living lab” model for evaluating AI system prototypes. It will serve as a community resource for research on AI4W and W4AI. During the planning phase we want to get research community input to prioritize enhancements to COSMOS.

Planned Activities: we are planning a number of activities during this process that you may be interested in:

  • A workshop on AI4W: scheduled for January 30, 2026 at WINLAB, this workshop will identify the key functionalities for an AI-enabled COSMOS to accelerate research on AI4W.
  • A workshop on W4AI: tentatively scheduled for May/June 2026, this workshop will identify what measurements and cyberinfrastructure needed to develop wireless-native AI applications. 
  • A prototyping/demo event: in 2027 to showcase the future directions which could be enabled by the STAIRWAI project.

Who we are: this project is led by Anand D. Sarwate with a team of other researchers associated with Rutgers WINLAB: co-PIs Jorge Ortiz, Bo Yuan, Ivan Seskar, and Yingying Chen as well as senior personnel Narayan Mandayam, Waheed Bajwa, Zhao Zhang, and Hang Liu.

STAIRWAI: Building an AI-Enabled Testbed for Wireless+AI