e3 - A Framework for the Automated Generation of...

Info iconThis preview shows pages 1–2. Sign up to view the full content.

View Full Document Right Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: A Framework for the Automated Generation of Power-Efficient Classifiers for Embedded Sensor Nodes Ari Y. Benbasat and Joseph A. Paradiso The Media Laboratory Massachusetts Institute of Technology { ayb,joep } @media.mit.edu Abstract This paper presents a framework for power-efficient de- tection in embedded sensor systems. State detection is struc- tured as a decision tree classifier that dynamically orders the activation and adjusts the sampling rate of the sensors (termed groggy wakeup), such that only the data necessary to determine the system state is collected at any given time. This classifier can be tuned to trade-off accuracy and power in a structured, parameterized fashion. An embedded instan- tiation of these classifiers, including real-time sensor control, is described. An application based on a wearable gait monitor provides quantitative support for this framework. The decision tree classifiers achieved roughly identical detection accuracies to those obtained using support vector machines while drawing three times less power. Both simulation and real-time opera- tion of the classifiers demonstrate that our multi-tiered clas- sifier determines states as accurately as a single-trigger (bi- nary) wakeup system while drawing as little as half as much power and with only a negligible increase in latency. Categories and Subject Descriptors I.5.2 [ Pattern Recognition ]: Design Methodology; C.3 [ Computer Systems Organization ]: Special-Purpose and Application-Based Systems General Terms Algorithms, Design, Performance Keywords tiered wakeup, power-efficient detection, dynamic power management, wearable sensors 1 Introduction Embedded sensor nodes are currently being used in a wide array of applications. These include, but are certainly not limited to, detecting degenerative diseases [3], monitor- ing remote regions [29], and ensuring the safety of house- Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, to republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. SenSys’07, November 6–9, 2007, Sydney, Australia. Copyright 2007 ACM 1-59593-763-6/07/0011 ...$5.00 bound elders [15]. Such systems are part of a new class of sensor-driven applications, leveraging the decrease in both price and size of components to allow rich, multimodal data streams to be captured by very compact systems. However, as sensors nodes increase in functionality, they require more frequent activation and therefore more frequent replacement or recharging of batteries. This creates an in- creasing gap between the capabilities of a device and its lifes- pan under normal use. Thus, current applications of embed- ded sensor systems are mostly limited to prototype and ex- perimental usage or very simple implementations. The mostperimental usage or very simple implementations....
View Full Document

Page1 / 14

e3 - A Framework for the Automated Generation of...

This preview shows document pages 1 - 2. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online