Trading off Prediction Accuracy and Power Consumption for Context-Aware Wearable Computing

Authors:
Andreas Krause;Matthias Ihmig;Edward Rankin;Derek Leong;Smriti Gupta;Daniel Siewiorek;Asim Smailagic;Michael Deisher;Uttam Sengupta
Affiliations:
School of Computer Science, Carnegie Mellon University, Pittsburgh;Dept. of Electrical Engineering and Information Science, Technische Universität München, Germany;Electrical and Computer Engineering Department, Carnegie Mellon University, Pittsburgh;Electrical and Computer Engineering Department, Carnegie Mellon University, Pittsburgh;Electrical and Computer Engineering Department, Carnegie Mellon University, Pittsburgh;School of Computer Science, Carnegie Mellon University, Pittsburgh;School of Computer Science, Carnegie Mellon University, Pittsburgh;Intel, Hillsboro, OR;4Intel, Hillsboro, OR
Venue:
ISWC '05 Proceedings of the Ninth IEEE International Symposium on Wearable Computers
Year:
2005

Citing 0
Cited 18

Power and accuracy trade-offs in sound-based context recognition systems

Pervasive and Mobile Computing
MEDIC: Medical embedded device for individualized care

Artificial Intelligence in Medicine
Functionality-power-packaging considerations in context aware wearable systems

Personal and Ubiquitous Computing - Special Issue: Selected Papers of the ARCS06 Conference
On-body activity recognition in a dynamic sensor network

Proceedings of the ICST 2nd international conference on Body area networks
Classifying wheelchair propulsion patterns with a wrist mounted accelerometer

BodyNets '08 Proceedings of the ICST 3rd international conference on Body area networks
A framework of energy efficient mobile sensing for automatic user state recognition

Proceedings of the 7th international conference on Mobile systems, applications, and services
Using mobile phones to determine transportation modes

ACM Transactions on Sensor Networks (TOSN)
Markov-optimal sensing policy for user state estimation in mobile devices

Proceedings of the 9th ACM/IEEE International Conference on Information Processing in Sensor Networks
Power-accuracy tradeoffs in human activity transition detection

Proceedings of the Conference on Design, Automation and Test in Europe
A hybrid context-aware wearable system with evolutionary optimization and selective inference of dynamic bayesian networks

HAIS'11 Proceedings of the 6th international conference on Hybrid artificial intelligent systems - Volume Part I
Balancing energy, latency and accuracy for mobile sensor data classification

Proceedings of the 9th ACM Conference on Embedded Networked Sensor Systems
Semantic management of multiple contexts in a pervasive computing framework

Expert Systems with Applications: An International Journal
Network-Level Power-Performance Trade-Off in Wearable Activity Recognition: A Dynamic Sensor Selection Approach

ACM Transactions on Embedded Computing Systems (TECS)
SymPhoney: a coordinated sensing flow execution engine for concurrent mobile sensing applications

Proceedings of the 10th ACM Conference on Embedded Network Sensor Systems
An event-driven energy efficient framework for wearable health-monitoring system

AMT'12 Proceedings of the 8th international conference on Active Media Technology
Resource-aware architectures for adaptive particle filter based visual target tracking

ACM Transactions on Design Automation of Electronic Systems (TODAES)
A system for visualizing human behavior based on car metaphors

Proceedings of the 4th Augmented Human International Conference
Activity recognition for creatures of habit

Personal and Ubiquitous Computing

Quantified Score

Hi-index	0.00

Visualization

Abstract

Context-aware mobile computing requires wearable sensors to acquire information about the user. Continuous sensing rapidly depletes the wearable system's energy, which is a critically constrained resource. In this paper, we analyze the trade-off between power consumption and prediction accuracy of context classifiers working on dual-axis accelerometer data collected from the eWatch sensing and notification platform. We improve power consumption techniques by providing competitive classification performance even in the low frequency region of 1-10 Hz and for the highly erratic wrist based sensing location. Furthermore, we propose and analyze a collection of selective sampling strategies in order to reduce the number of required sensor readings and the computation cycles even further. Our results indicate that optimized sampling schemes can increase the deployment lifetime of a wearable computing platform by a factor of four without a significant loss in prediction accuracy.