Low power architectures for always-on sensing in IOT devices
Bldg. 320-109
About the talk: Contextual-awareness is a set of techniques by which a smartphone or smartwatch can become “aware” of the “context” of its user, such as the user’s activities, location, and routine. This context can be used to support services and apps such as health monitoring. One of the challenges in mobile battery operated devices is the battery energy spent in extracting the contextual-awareness from data available from the sensors. This is in turn limits the duration over which the apps can run without draining the battery. Ideally consumers want “always-on” contextual-aware apps that run in the background 24/7. An example of this is parents keeping track of their children’s location, environment or activities to ensure they are safe. In this talk we discuss architectures reducing average battery current for always-on apps.
About the speaker: Rajeev Jain received his B. Tech in EE from IIT Delhi in 1978 and Ph.D. in EE from K.U. Leuven in 1985. He leads ML R&D for SoC design at Qualcomm Inc. He has done research in CAD and IC design at Siemens, IMEC, UC Berkeley, UCLA (where he is currently Professor Emeritus). In 1999 he was elected IEEE Fellow for his contributions to Computer-Aided Design Tools for Signal Processing Circuits. At Qualcomm he also led the research in low-power Machine Learning architectures for sensing for which he and his team of co-inventors were awarded a US patent.