Vision AI-based Occupant Metabolic Rate and Clothing Insulation Estimation for Building Cooling Load and Thermal Comfort Simulation Analysis

Jin-Ju Lee; Eui-Ryeoung Jeong; Cham-Sol Kim; Wonseok Oh

doi:10.22696/jkiaebs.20260018

All Issue

2026 Vol.20, Issue 3 Preview Page Next Page

Research Article

Vision AI-based Occupant Metabolic Rate and Clothing Insulation Estimation for Building Cooling Load and Thermal Comfort Simulation Analysis 비전 AI기반 재실자 대사량‧착의량 추정을 통한 건물 냉방 부하 및 열쾌적성 시뮬레이션 분석

30 June 2026. pp. 206-222

PDF XML

Abstract

This study proposes a video-based artificial intelligence pipeline that automatically estimates the metabolic rate (MET) and clothing insulation (CLO) of building occupants, and applies the derived dynamic behavior data to an eQUEST building energy simulation. The MET estimation algorithm uses YOLOv11n-Pose to classify three activity states (sitting, standing, and walking) through joint-angle analysis and ankle-motion FFT, achieving an average error of 24.9% (R = 0.44) in single-occupant validation. The CLO estimation algorithm combines YOLOv7-Pose keypoint extraction, DeepFashion2-based garment category matching, and YCrCb skin-ratio correction for sleeve and trouser length, achieving an average error of 11.5% (R = 0.83). Applied to a real classroom environment (Room 209, N13 Building, Hanbat National University; 31 occupants), the algorithms yielded a group-averaged MET of 1.08 and CLO of 0.565. eQUEST simulations showed that applying the measured dynamic MET (1.08) increased the peak space cooling load by 6.93% (from 3,581 W to 3,829 W) relative to the conventional static assumption (MET = 1.0). Thermal comfort analysis using the CBE Thermal Comfort Tool (ASHRAE 55) indicated that Case 4, which integrates both dynamic MET and CLO, achieved PMV = −0.14 and PPD = 5%, the most neutral result among the four cases. The corresponding SET* of 24.6℃ indicating a shift toward thermal neutrality at the same set-point. However, the validation was limited to a single classroom on a single observation day, and the group-averaged estimates may be subject to bias from occlusion-induced detection loss; future work will extend the framework to multi-season and multi-density scenarios across various space types.

Keywords

Occupant behavior

Metabolic rate (MET)

Clothing insulation (CLO)

Building energy simulation

Thermal comfort

키워드

재실자 행동패턴

대사량

착의량

건물 에너지 시뮬레이션

열쾌적성

References

Ainsworth, B.E., Haskell, W.L., Herrmann, S.D., Meckes, N., Bassett, D.R., Tudor-Locke, C., Greer, J.L., Vezina, J., Whitt-Glover, M.C., Leon, A.S. (2011). Compendium of Physical Activities: A Second Update of Codes and MET Values. Medicine & Science in Sports & Exercise, 43(8), 1575-1581.

10.1249/MSS.0b013e31821ece12

Cao, Z., Hidalgo, G., Simon, T., Wei, S.-E., Sheikh, Y. (2021). OpenPose: Realtime Multi-Person 2D Pose Estimation Using Part Affinity Fields. IEEE Transactions on Pattern Analysis and Machine Intelligence, 43, 172-186.

10.1109/TPAMI.2019.2929257

Kim, Y., Yoo, B., Yun, J.Y., Lee, H., Park, S., Moon, J.W., Choi, E.J. (2025). Evaluation of thermal control based on spatial thermal comfort with reconstructed environmental data. Journal of Building Engineering, 115, 114512.

10.1016/j.jobe.2025.114512

Kim, Y., Choi, J. (2024). Multimodal Deep Learning Model for Human Activity Recognition Using Body Landmarks from MediaPipe Extraction and IMU. The Journal of KING Computing, 20(3), 12-21.

Kim, S.-G., Ahn, J.-S., Kim, W.-H. (2017). Implementation of fall-down detection algorithm based on Image Processing. Journal of Satellite, Information and Communications, 12(2), 56-60.

Na, J.-W., Park, D.-Y., Kim, K.-S. (2015) Comparison of Building Envelope Simulation Results between EnergyPlus and eQuest for a Small Sized Office Building. Journal of Architectural Institute of Korea, 31(6), 175-183.

10.5659/JAIK_PD.2015.31.6.175

O’Brien, W., Gunay, H.B. (2014). The contextual factors contributing to occupants’ adaptive comfort behaviors in offices. Building and Environment, 78.

10.1016/j.buildenv.2014.03.024

Park, D., Yoon, K.C., Kim, K.S. (2013) A Comparative Study on Heating Energy Consumption of Multi-Family Apartment using EnergyPlus and eQUEST. Journal of Korean Solar Energy Society, 33(1), 48-56.

10.7836/kses.2013.33.1.048

Wang, Z., Wang, J., Chen, Y. (2020). Vision-Based Human Activity Recognition: A Survey. IEEE Access, 8, 129-143.

Yang, C.C., Hsu, Y.L. (2010). A Review of Accelerometry-Based Wearable Motion Detectors for Physical Activity Monitoring. Sensors, 10(8), 7772-7788.

10.3390/s10080777222163626PMC3231187

Ge, Y., Zhang, R., Wu, L., Wang, X., Tang, X., Luo, P. (2019). DeepFashion2: A Versatile Benchmark for Detection, Pose Estimation, Segmentation and Re-Identification of Clothing Images. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).

10.1109/CVPR.2019.00548

Yang, W., Luo, P., Lin, W., Tang, X. (2016). A large-scale fashion dataset for attribute prediction. Proceedings of the IEEE International Conference on Computer Vision (ICCV).

Konak, O., Wischmann, A., van de Water, R., Arnrich, B. (2023). A Real-time Human Pose Estimation Approach for Optimal Sensor Placement in Sensor-based Human Activity Recognition. arXiv preprint arXiv:2307.02906.

10.1145/3615834.3615848

Pfister, T., Charles, J., Zisserman, A. (2015). Flowing ConvNets for Human Pose Estimation in Videos. arXiv preprint arXiv:1506.02897.

10.1109/ICCV.2015.222

Pishchulin, L., Andriluka, M., Schiele, B. (2014). Fine-grained Activity Recognition with Holistic and Pose Based Features. arXiv preprint arXiv:1406.1881.

10.1007/978-3-319-11752-2_56

Wang, C.Y., Bochkovskiy, A., Liao, H.Y.M. (2022). YOLOv7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors. arXiv preprint arXiv:2207.02696.

10.1109/CVPR52729.2023.00721

Hirsch J.J., Associates (2009). eQUEST Modeling Procedures Quick Reference Guide. Dec 2009 Edition, Camarillo, CA: James J. Hirsch & Associates.

ASHRAE. (2020). ANSI/ASHRAE Standard 55-2017 Addendum d: Thermal Environmental Conditions for Human Occupancy, ASHRAE.

ASHRAE. (2022). Energy Standard for Sites and Buildings Except Low-Rise Residential Buildings, I-P Edition, Peachtree Corners, GA: ASHRAE.

Jocher, G., Chaurasia, A., Qiu, J. (2023). YOLOv8: Ultralytics YOLO for Real-Time Object Detection. Ultralytics. Available at: https://docs.ultralytics.com [Accessed on 05/01/2025].

Information

Publisher :Korean Institute of Architectural Sustainable Environment and Building Systems
Publisher(Ko) :한국건축친환경설비학회
Journal Title :Journal of Korean Institute of Architectural Sustainable Environment and Building Systems
Journal Title(Ko) :한국건축친환경설비학회논문집
Volume : 20
No :3
Pages :206-222
Received Date : 2026-04-23
Revised Date : 2026-05-16
Accepted Date : 2026-05-22
DOI :https://doi.org/10.22696/jkiaebs.20260018

Journal of Korean Institute of Architectural Sustainable Environment and Building Systems ISSN:1976-6483(Print) 2586-0666(Online) 한국건축친환경설비학회논문집

All Issue