A Lightweight MBConv-Based Model for Real-Time Load Identification
Non-Intrusive Load Monitoring (NILM) disaggregates household power consumption into appliance-level data, offering economic and privacy-preserving solution for energy management. However, many SOTA models such as transformers require high computational power and large datasets, posing practical and privacy challenges. Inspired by lightweight CNN advancements, this paper proposes a real-time MBConv-based model that reduces parameters while maintaining strong performance.
Authors
- Zibin Pan
- Caijun Zhong (ORCID: https://orcid.org/0000-0002-9960-7800)
- Zhipeng Lin (ORCID: https://orcid.org/0000-0001-5941-2163)
- Pengfei Lu (ORCID: https://orcid.org/0000-0001-6541-6216)
- Jing Jin (ORCID: https://orcid.org/0009-0004-6153-1558)
- Shuwen Zhang (ORCID: https://orcid.org/0000-0003-3555-6677)
- Guolong Liu (ORCID: https://orcid.org/0000-0002-3726-4216)
- C. J. Dai
Institutions
- Nanyang Technological University (SG)
- Chinese University of Hong Kong, Shenzhen (CN)
- Shanghai Electric (China) (CN)
Publication Details
- Published
- 2024-11-29
- DOI
- https://doi.org/10.1109/ei264398.2024.10991737
- Primary Topic
- Elevator Systems and Control
- Type
- article
- Field-Weighted Citation Impact
- 0.00