Upload
demetris-trihinas
View
158
Download
2
Embed Size (px)
Citation preview
Low-Cost Adaptive Monitoring Techniques for the Internet of Things
The AdaM Framework
The Challenges…Challenge 1 - Taming data volume and data velocity with limited processing and network capabilities
Challenge 2 - IoT devices are usually battery-powered which means intense processing leads to less battery-life
Data Volume Network Bandwidth
Battery Life
The Problems…
Research QuestionCan we reduce the volume of IoT generated data while at the same
time reduce on device processing to preserve battery life?What if:
• A framework could predict metric stream evolution and dynamically
adapt the rate at which metrics are collected…
• Filter out metric values when consecutive values do not differ and adapt
to ensure accuracy defined by users is always met…
𝑠𝑖 𝑠𝑖+1𝑇 𝑖+1
Metric Stream dynamically sampledMetric Stream with
The AdaM Framework
Raspberry Pi
Arduino
Beacons
ProcessingUnit
DisseminationUnit
AdaptiveSampling
Knowledge Base
AdaM
IoT Devicemetrics
adjust filter range
AdaptiveFiltering
API
Activity Trackers
SensingUnit
adjust sampling rate
filtered metrics
• Reduces on device processing, energy consumption and allocated bandwidth
• Reduces volume and velocity of data generated in streaming networks
• Achieves a balance between efficiency and accuracy
• AdaM dynamically adapts the monitoring intensity of IoT devices
based on metric stream evolution and variability
AdaM: an Adaptive Monitoring Framework for Sampling and Filtering on IoT Devices. Trihinas, D.; Pallis, G.; and Dikaiakos, M. D. In 2015 IEEE International Conference on Big Data, (IEEE BigData 2015), pages 717--726, 2015.
AdaM in action!
Steps Heartrate
Calories
94% accuracy!
96% accuracy!
91% accuracy!
AdaM in action!
6x less processing! 4x less network traffic!
4x less energy usage!
+ AdaM
3 - 5 days
7 - 8 days