About
Suman Banerjee is a professor of Computer Sciences at UW–Madison since Dec 2003. He was a visiting researcher at Intel Research and University of Cambridge (2003), a visiting professor of Sorbonne University (2018), and a Santander Chair of Excellence Professor to Universidad Carlos III Madrid and IMDEA Networks (2019).
He received the inaugural ACM SIGMOBILE Rockstar Award for early career achievements in mobile computing and wireless networking. He served as the Chair of ACM SIGMOBILE from 2013–2017 and has co-founded multiple startups, including StratusWorX, acquired by Ericsson in 2020.
He is the founding director of the WiNGS Laboratory at UW-Madison, focused on wireless and mobile networking systems.
⭐ Recent Highlights
- 2026🎓 Launched a new CS Entrepreneurship Seminar Series featuring founders, investors, and industry leaders — inaugural session with Epic founder Judy Faulkner MS'67. Read more →
- 2021–26💰 Co-PI of the $20M NSF AI Institute — Athena for Edge Computing Leveraging Next Generation Networks.
- 2025–29💰 Co-PI, NSF NeTS award on unifying last-mile Low-Power Wide Area Network technologies.
- 2026📰 MEDUSA featured in UW–Madison News: contactless radar vital-sign monitoring from the lab to the NICU.
Teaching
- CS 407: Foundations of Mobile Systems & ApplicationsFall 2025, 2024, 2023…
- CS 640: Introduction to Computer NetworksSpring 2026, 2025, 2024…
- CS/ECE 707: Mobile and Wireless NetworkingFall 2022, 2020, 2019…
- CS 740: Advanced Computer NetworksFall 2004, 2005
- Foundations of Wireless Networking Protocols — Universidad Carlos III MadridSpring 2019
Research
EasyVizAR: Collaborative AR for First Responders
A collaborative augmented reality system that supports first responders in challenging indoor environments. EasyVizAR provides real-time situational awareness through shared AR headsets, helping teams navigate and coordinate during emergencies. A companion study explores the design space of optical see-through AR head-mounted displays tailored to first responder needs in the field.
Sustainable Spectrum Crowdsensing
Large-scale, energy-efficient crowdsourced spectrum monitoring. Distributed low-cost sensors collaboratively map radio spectrum usage with an adaptive uplink compression scheme (FlexSpec) that minimizes overhead while preserving measurement fidelity.
On-Device AI: Benchmarking & Efficient Inference
A series of projects pushing large AI models onto mobile and battery-powered hardware. PalmBench provides a comprehensive benchmark of compressed LLMs on mobile platforms. NANOMIND goes further with a hardware-software co-design inference framework that breaks large multimodal models into modular components and maps each to its optimal accelerator (NPU, GPU, DSP), enabling full LMM inference on a compact battery-powered device.
Edge Computing in the Extreme
A series of projects pushing computation to the very edge of the network. ParaDrop enables lightweight multi-tenancy at the home gateway. EdgeEye delivers real-time intelligent video analytics as an edge service. The Roaming Edge explores edge computing for nomadic and mobile users. Together these projects define a vision for computation at the network's extreme edge.
LoRa Networking: Interference Cancellation & Cloud Radio Access
Two complementary advances in LoRa networking. Concurrent Interference Cancellation decodes multi-packet collisions in LoRa using successive interference cancellation, dramatically improving network capacity. Cloud-LoRa extends this by enabling cloud radio access LoRa networks via reinforcement learning-based bandwidth-adaptive compression.
WiNEST: City-Scale 5G Wireless Testbed
NSF-funded wireless research infrastructure deployed across Madison, WI, enabling city-scale experimentation with next-generation technologies in real-world settings.
Wireless Interference Management
A series of projects tackling interference in wireless networks. Airshark detects non-WiFi interferers — Bluetooth, microwave ovens, game controllers, ZigBee — using only commodity WiFi hardware. PIE maps WiFi-to-WiFi interference across enterprise WLANs in real time. CENTAUR realizes the full potential of centralized WLANs using a hybrid data path, winning the best paper award at ACM MobiCom 2009. FLUID improves throughput in enterprise WLANs through flexible channelization (Best Paper Nominee, MobiCom 2011). Diagnosing Wireless Packet Losses separates collisions from weak signal in 802.11 networks.
WiScape
Monitors wide-area cellular network performance via client assistance. Data collected over 2+ years covering 60+ sq. miles in and around Madison and across the US Midwest.
WiRover
High-bandwidth Internet for moving vehicles via network diversity and aggregation. Grand Prize, Wisconsin Governor's Business Plan Competition 2011.
Value-Aware Wireless Systems
Two projects that make wireless systems "value-aware" — understanding the relative importance of different bits to the application. Apex improves media streaming by making the PHY layer aware of bit importance, enabling graceful quality degradation. Scalable WiFi Media Delivery extends this to multicast, using adaptive broadcasts to deliver media to many clients simultaneously at scale.