Spotlight: Personal Natural Resource Consumption Profiler

1
Spotlight Personal Natural Resource Consumption
Profiler

• Younghun Kim, Zainul Charbiwala, Akhilesh
  Singhania, Thomas Schmid, Mani B. Srivastava
• Networked and Embedded Systems Laboratory
• Department of Electrical Engineering
• University of California, Los Angeles
• kimyh, zainul, akhi, thomas.schmid,
  mbs_at_ucla.edu

2
Outline

• Problem Definition
• Introduction of SPOTLIGHT
• System Architecture
• Prototype Implementation
• Evaluation
• Conclusion

3
Problem Definition

• Severe natural disasters due to global climate
  changes
• An urge to optimize resource consumption due to
  natural resource crisis.
• Household level and monthly basis resource
  consumption measurement ? Very vague

4
Motivation

• Individual wants to have fine grained information
  of the resource consumption.
• Study shows individuals are willing to conserve
  energy.
• Measurement of energy, water, and other resources
  at person level.

5
Introduction of SPOTLIGHT

• Resource measuring system using sensor network
  technology.
• Estimating an individuals personal resource
  consumption profiler.
• Use commercially available components
• Currently focusing on electrical energy.

6
Resource Monitoring Subsystem (1/3)

• Examples of available technologies Power and gas
  meters for apartments.
• Problem Information not at per individual level.

7
Resource Monitoring Subsystem (2/3)

• Challenges
• Scalability
• Accuracy
• Non-obstructive due to aesthetic, privacy and
  security concerns.

8
Resource Monitoring Subsystem (3/3)

• Aims to profile energy consumption of each
  appliance based on the status of an appliance and
  total consumption.

9
Activity Monitoring Subsystem

• Identify who and what activities happened in the
  area of interest.
• Existing activity recognition approaches but
  issue of mass deployment due to complexity and
  cost reason still persists.
• Aims to fuse information from both activity and
  resource monitoring to improve accuracy.

10
Resource Management Subsystem

• Fuse data provided by resource monitoring and
  activity monitoring subsystems
• Provide persuasive resource usage report
• Employ Smart Resource Controller to improve
  resource saving.

11
(No Transcript)
12
Prototype Implementation (1/2)

• Monitor the power consumption in a small
  household.
• Participating appliances Lighting, coffee
  machine and TV.
• Regards physical proximity and the appliance
  status as an activity indicator.

13
Prototype Implementation (2/2)

• MicaZ
• Used to for sensing proximity of the user to the
  appliance
• Communicating with SPOTLIGHT server
• Connected to the serial interface of power meter

14
3. Identify user proximity and individual
electrical energy consumption
15
Deployment Setting
16
Evaluation of RSSI based Activity Region

• Define the a rough activity region
• Traverse along the boundaries and the system
  averages RSSI for initial calibration.
• Issues tokens, an indication of current activity.

17
Comparison With Ground Truth

• Installed cameras monitoring the activity
  regions.
• Provided information of who was present in the
  activity region.

18
Comparison With Ground Truth
19
Energy Consumption Report
20
Energy Consumption Report
21
Future Work

• Measuring water consumption of a household.
• Using accelerometer connected to the water pipe.
• Vibration of the pipe is a good indicator of how
  much water flows through the pipe

22
Conclusion

• Providing information on resource consumption is
  meaningful for inducing resource-conserving
  behavioral changes
• Such a system needs to fuse different embedded
  sensing system, machine learning technique and
  algorithm