An innovative edge-based Internet of Energy solution for promoting energy saving in buildings. (March 2022)

• Record Type: Journal Article
• Authors: Alsalemi, Abdullah; Himeur, Yassine; Bensaali, Faycal; Amira, Abbes
• Is Part Of: Sustainable cities and society. Volume 78(2022)
• Publication Date: 2022-03
• Subjects: Internet of Energy (IoE) -- Edge computing -- Anomaly detection -- Appliance identification -- Recommendations -- Energy consumption; Sustainable urban development -- Periodicals; Sustainable buildings -- Periodicals; Urban ecology (Sociology) -- Periodicals; 307.76
• DOI: 10.1016/j.scs.2021.103571
• Languages: English
• ISSNs: 2210-6707

Abstract

Abstract: Due to the ubiquity and maturity of Artificial Intelligence (AI), it became an essential tool in the development real-time Internet of energy (IoE) solutions. Also, since cloud platforms are not being the first implementation choice due to their bandwidth and latency issues that limit data transmission capacity, Edge Computing is becoming a popular alternative in both IoE and IoT applications. To that end, a novel Edge IoE system is proposed, namely M2SP-EdgeIoE, which enables various potential IoE use scenarios for domestic energy efficiency applications. Specifically, it incorporates (i) the sensing unit used as input stage for collecting data; (ii) the energy disaggregation unit, in which non-intrusive load monitoring algorithms are implemented to identify appliances using their disaggregated power traces; (iii) the anomaly detection platform in which energy consumption abnormalities are detected using multimodal data and AI algorithms; and (iv) the recommender system unit that helps in generating tailored recommendations for improving energy consumption behavior. Promising performance has been obtained under a number of scenarios, for example, an accuracy of 98.49% and 95% have been reached for appliance identification and anomaly detection, respectively. In conclusion, the M2SP-EdgeIoE provides a blueprint for the future of Edge IoE solutions for improved global energy efficiency. Highlights: Edge Computing is becoming prominent for Internet of Energy. Novel Edge IoE system is proposed enabling IoE scenarios for energy efficiency. The system comprises a sensing unit along with an edge AI classification system. Classification accuracy of ∼ 95% for appliance identification and anomaly detection.