Power Consuming Activity Recognition in Home Environment

Liu, Xiaodong; Liu, Qi

Abstract

This work proposed an activity recognition model which focus on the power con-suming activity in home environment, to help residents modify their behavior. We set the IoT system with lower number of sensors. The key data for identifying activity comes from widely used smart sockets. It first took residents’ acceptability into consideration to set the IoT system, then used a seamless indoor position system to get residents’ position to help recognize the undergoing activities. Based on ontology, it made use of domain knowledge in daily activity and built an activity ontology. The system took real home situation into consideration and make full use of both electric and electronic appliances’ data into the context awareness. The knowledge helps improve the performance of the data-driven method. The experiment shows the system can recognize the common activities with a high accuracy and have a good applica-bility to real home scenario.

Citation

Liu, X., & Liu, Q. (2017, June). Power Consuming Activity Recognition in Home Environment. Presented at the 3rd International Conference on Cloud Computing and Security (ICCCS 2017)

• Presentation Conference Type: Conference Paper (published)
• Conference Name: the 3rd International Conference on Cloud Computing and Security (ICCCS 2017)
• Start Date: Jun 16, 2017
• End Date: Jun 18, 2017
• Acceptance Date: May 22, 2017
• Online Publication Date: Oct 31, 2017
• Publication Date: 2017
• Deposit Date: Nov 15, 2017
• Publicly Available Date: Nov 16, 2017
• Electronic ISSN: 1611-3349
• Publisher: BMC
• Pages: 361-372
• Series Title: Lecture Notes in Computer Science
• Series Number: 10602
• Series ISSN: 0302-9743
• Book Title: Cloud Computing and Security. ICCCS 2017
• ISBN: 9783319685045; 9783319685052
• DOI: https://doi.org/10.1007/978-3-319-68505-2_31
• Keywords: Activity recognition, Ontology, Second-order HMM,
• Public URL: http://researchrepository.napier.ac.uk/Output/1010653
• Contract Date: Nov 15, 2017