Occupant activities and clothes detection based on semi-supervised learning for occupant-centric thermal control

被引:0
|
作者
Jung, Seunghoon [1 ]
Jeoung, Jaewon [1 ]
Kong, Minjin [1 ]
Hong, Taehoon [1 ]
机构
[1] Yonsei Univ, Dept Architecture & Architectural Engn, Seoul, South Korea
基金
新加坡国家研究基金会;
关键词
Occupant-centric control; Thermal comfort; Real-time monitoring; Multi-task learning; Semi-supervised learning;
D O I
10.1016/j.buildenv.2024.112178
中图分类号
TU [建筑科学];
学科分类号
0813 ;
摘要
Real-time monitoring of metabolic rate (MET) and clothing insulation (CLO) is essential to ensure effective occupant-centric control (OCC) for thermal comfort. This study aims to develop multi-task model using semisupervised learning to enhance occupant activities and clothes detection performance by utilizing both labeled and unlabeled data. The convolutional neural network-based model and training approach with pseudo labels to update all parameters comprehensively were proposed. The developed model is validated by conducting comparative analysis with state-of-the-art models and applying it in a real-world environment. The results demonstrate that the developed model, employing semi-supervised learning and the dual-phase training method (DPTM), achieves superior performance in activity and clothes detection outperforming previous studies with a 15.8 % higher mean Average Precision (mAP) for activity detection and a 25 % improvement for clothes detection. The findings highlight the potential of this multi-task model using semi-supervised learning to automate data collection improving the accuracy of estimating occupant thermal comfort. This approach can dynamically optimize indoor environments tailored to individual needs within the OCC framework, enhancing thermal comfort and energy efficiency through precise monitoring of occupant information.
引用
收藏
页数:12
相关论文
共 50 条
  • [41] Developing building-specific, occupant-centric thermal comfort models: A methodological approach
    Upasani, Nitant
    Guerra-Santin, Olivia
    Mohammadi, Masi
    JOURNAL OF BUILDING ENGINEERING, 2024, 95
  • [42] Investigating the influence of uncertainty on office building energy simulation through occupant-centric control and thermal comfort integration
    Kang, Eunho
    Lee, Hyomun
    Yoon, Jongho
    Cho, Heejin
    Chaichana, Chatchawan
    Kim, Dongsu
    ENERGY AND BUILDINGS, 2024, 322
  • [43] An occupant-centric adaptive facade based on real-time and contactless glare and thermal discomfort estimation using deep learning algorithm
    Wang, Yuxiao
    Han, Yunsong
    Wu, Yuran
    Korkina, Elena
    Zhou, Zhibo
    Gagarin, Vladimir
    BUILDING AND ENVIRONMENT, 2022, 214
  • [44] Bio-signals based occupant-centric lighting control for cognitive performance, visual fatigue and energy consumption
    Jung, Dahyun
    Jung, Seunghoon
    An, Jongbaek
    Hong, Taehoon
    BUILDING AND ENVIRONMENT, 2025, 269
  • [45] Machine learning-based analysis of occupant-centric aspects: Critical elements in the energy consumption of residential buildings
    Alsharif, Rashed
    Arashpour, Mehrdad
    Golafshani, Emadaldin Mohammad
    Hosseini, M. Reza
    Chang, Victor
    Zhou, Jenny
    JOURNAL OF BUILDING ENGINEERING, 2022, 46
  • [46] Temperature-preference learning with neural networks for occupant-centric building indoor climate controls
    Peng, Yuzhen
    Nagy, Zoltan
    Schluter, Arno
    BUILDING AND ENVIRONMENT, 2019, 154 : 296 - 308
  • [47] Occupant-centric control of transparent dynamic facades through an integrated co-simulation framework
    Giovannini, Luigi
    Baracani, Manuela
    Favoino, Fabio
    Serra, Valentina
    PROCEEDINGS OF BUILDING SIMULATION 2021: 17TH CONFERENCE OF IBPSA, 2022, 17 : 2679 - 2686
  • [48] Enhancing occupant-centric ventilation control in airport terminals: A predictive optimization framework integrating agent-based simulation
    Tang, Hao
    Yu, Juan
    Geng, Yang
    Liu, Xue
    Huang, Zujian
    Yang, Yuren
    Wang, Zhe
    Chen, Ying
    Lin, Borong
    BUILDING AND ENVIRONMENT, 2025, 276
  • [49] Current practices and infrastructure for open data based research on occupant-centric design and operation of buildings
    Kjaergaard, Mikkel B.
    Ardakanian, Omid
    Carlucci, Salvatore
    Dong, Bing
    Firth, Steven K.
    Gao, Nan
    Huebner, Gesche Margarethe
    Mahdavi, Ardeshir
    Rahaman, Mohammad Saiedur
    Salim, Flora D.
    Sangogboye, Fisayo Caleb
    Schwee, Jens Hjort
    Wolosiuk, Dawid
    Zhu, Yimin
    BUILDING AND ENVIRONMENT, 2020, 177
  • [50] HVAC energy savings, thermal comfort and air quality for occupant-centric control through a side-by-side experimental study
    Kong, Meng
    Dong, Bing
    Zhang, Rongpeng
    O'Neill, Zheng
    APPLIED ENERGY, 2022, 306