AmI (Ambient Intelligence) Lab Vision
Led by Prof. Ke Sun, the AmI Lab was established in CSE of the EECS Department at the University of Michigan, Ann Arbor in January 2025.
We envision a future where our environments intuitively adapt in real-time to augment human abilities through Ambient Intelligence (AmI). In this future, devices, sensors, and processors are seamlessly embedded into everyday objects and spaces, creating intelligent systems that proactively adjust to individual needs. Guided by this vision, the AmI Lab at UMich CSE focuses on developing intelligent, cost-effective, deployable, human-centric, and trustworthy Mobile, Wearable, and IoT (MWIoT) systems. Our research goals include:
- Enable novel applications through advanced sensing technologies;
- Advance computational sensing techniques to enhance the capabilities and performance of MWIoT systems;
- Address system bottlenecks in MWIoT ecosystems, ensuring efficiency and reliability.
Lab News
Samsung START Project supports our lab's research on multi-modal lifelong memory augmentation systems.
August 01, 2025
Prof. Ke Sun (PI), Prof. Anhong Guo (Co-PI), and Prof. Kang G.Shin (Co-PI) have been awarded a Samsung START project to develop a multi-modal lifelong memory augmentation system that can capture, store, and retrieve human experiences in real-time.
More...
AmI lab warmly welcomes the first two Ph.D. students, Linzhen Zhu and Leonard Zhang, to join the lab!
August 01, 2025
Please check our team member page for more information about Linzhen and Leonard.
More...
AmI lab launches a new website!
August 01, 2025
UltraPoser has been accepted at ACM UIST 2025. Congratulations to the Yadong and Shuning!
July 15, 2025
UltraPoser is a multi-modal full-body motion capture system using commdity wearable devices only to integrate ultrasonic sensing with inertial measurements for improved fidelity, broader coverage and increased reliability.
More...
EveGuard has been accepted at IEEE S&P (Oakland) 2025. Congratulations to the Jung-woo!
May 01, 2025
EveGuard is a software-driven defense framework that creates adversarial audio, protecting voice privacy from side channels (including mmWave radars, light sensors, and accelerometers) without compromising human perception.
More...