Existing wearables such as fitness trackers and smart watches rely on very simple processor chips due to their highly constrained battery life, which in turn limits the applications supported on wearables.
With advances in computer architecture that can run highly sophisticated software on tiny wearables, and the advent of flexible sensors, the next generation of smart wearables promises exciting applications.
With this vision, Professor Peh Li Shiuan from NUS School of Computing, is leading an interdisciplinary research team with other NUS faculty that architects ultra-low-power chips and uses these to power flexible sensors for next-generation wearable applications.
Prof Peh is known for her work in developing on-chip networks, where small chip islands are networked to enable them to run computation in parallel at low power. On-chip networks are now used commercially in data centre servers, running aggressive workloads such as artificial intelligence (AI). At NUS, her group has been designing on-chip networks for various ultra-low-power wearable chips, offering a high-performance computing platform for running sophisticated software on tiny wearables.
Armed with such wearable computing platforms, her group envisions enabling several next-generation applications that these platforms enable. For instance, pH Watch is the first demonstration of a reusable sweat sensor. It pairs the heart rate sensor (pulse oximeter) used in today’s wearables, with a sensor that measures the pH of and generates outputs, using aggressive software processing that benefits from the ultra-low-power wearable processor chips her group has designed.
The EMPOWER app uses AI to analyse data collected from wearable devices such as smart watches to help users monitor their health-related habits and deliver timely in-app notifications or digital ‘nudges’
The Laser fault Attack Benchmark Suite (LABs) is a software tool developed that helps computer chip designers create more secure computer chips against hardware attacks
Stitch is a novel processing chip that is low-powered and high-performing across many application domains, enabling the wearable device to perform effectively and independently
Upadhyay, M., Juneja, R., Wang, B., Zhou, J., Wong, W. F., & Peh, L. S. (2022, July). REACT: a heterogeneous reconfigurable neural network accelerator with software-configurable NoCs for training and inference on wearables. In Proceedings of the 59th ACM/IEEE Design Automation Conference (pp. 1291-1296).
Balaji, A. N., & Peh, L. S. (2021, May). AI-on-skin: Enabling On-body AI Inference for Wearable Artificial Skin Interfaces. In Extended Abstracts of the 2021 CHI Conference on Human Factors in Computing Systems (pp. 1-7).
Wang, B., Karunarathne, M., Kulkarni, A., Mitra, T., & Peh, L. S. (2019, November). Hycube: A 0.9 v 26.4 mops/mw, 290 pj/op, power efficient accelerator for iot applications. In 2019 IEEE Asian Solid-State Circuits Conference (A-SSCC) (pp. 133-136). IEEE.
Balaji, A. N., Yuan, C., Wang, B., Peh, L. S., & Shao, H. (2019, June). pH watch-leveraging pulse oximeters in existing wearables for reusable, real-time monitoring of pH in sweat. In Proceedings of the 17th annual international conference on mobile systems, applications, and services (pp. 262-274).
Karunaratne, M., Mohite, A. K., Mitra, T., & Peh, L. S. (2017, June). Hycube: A cgra with reconfigurable single-cycle multi-hop interconnect. In Proceedings of the 54th Annual Design Automation Conference 2017 (pp. 1-6).
More Researchers