MICL Seminar

Energy Management for Battery-Powered Embedded Systems

Sarma Vrudhula

The requirement of portability places severe restrictions on size and weight, which in turn limits the amount of energy that is continuously available to
maintain system operability. As a result, efficient energy utilization has become a important issue in the design of
battery-powered embedded computing systems.

We first present a new analytical model of a battery, and show how it can be used to estimate its "lifetime" given a time varying load profile. Predictions based on the model are compared with estimates obtained from low level simulation as well as direct measurements using a pocket computer. Next we describe how the model can be used to formulate a cost function for optimizing the lifetime of the battery. The optimization problem is to schedule a collection of tasks so to maximize the residual life of the battery subject to preserving ordering relations (if any) among the tasks and satisfying a delay budget. Voltage and clock frequency selection and insertion of idle periods are included. Finally, we show this work can be used to optimize the operational life of a distributed sensor network.
Sarma Vrudhula is a Professor in the Electrical and Computer Engineering department at the University of Arizona, in Tucson, AZ., and the Director of the NSF Center for Low Power Electronics.

His research and teaching interests are in VLSI design automation and CAD. Project he is currently working on include, stochastic methods for performance analysis and optimization of nano-scale circuits, system level energy management strategies, and analysis of systems with stochastic parameters

Sponsored by

Micron Technology Foundation