AAAI Publications, Workshops at the Twenty-Fourth AAAI Conference on Artificial Intelligence

Font Size: 
Appliance Recognition and Unattended Appliance Detection for Energy Conservation
Shih-Chiang Lee, Gu-Yuan Lin, Wan-Rong Jih, Jane Yung-Jen Hsu

Last modified: 2010-07-07


Providing energy conservation services becomes a hot research topic because more and more people attach importance to environmental protection. This research proposes a framework that consists of four process models: appliance recognition, activity-appliances model, unattended appliances detection, and energy conservation service. Appliance recognition model can recognizes the operating states of appliances from raw sensing data of electric power. An activity-appliances model has been built to associate activities with appliances according to the data of Open Mind Common Sense Project. Using the relationship between activities can help to detect unattended appliances, which are consuming electric power but not take part in the resident’s activities. After obtain information of appliance operating states and unattended appliances, residents can receive energy conservation services for notifying the energy consumption information. Finally, the experimental results show that dynamic Baysian network approach can achieve higher than 92% accuracy for appliance recognition. Data of activity-appliances model shows most appliances are strong activity-related.


appliance recognition; energy conservation service; common sense; machine learning

Full Text: PDF