Journal · 2025
Home activity recognition using infrequently-monitored HEMS Data
Abstract
This paper proposes a method for estimating household activities based only on the cumulative power consumption data obtained from the HEMS home distribution board every 30 min. The proposed method predicts the activity of each 30 min timeslot from the eight activity labels; household-level waking-up, household-level going-to-bed, room-level waking-up, room-level going-to-bed, cooking, laundry, dishwashing, and bathing. For the prediction, we first identify the branch circuit that is strongly correlated with each activity label and detect the turn-on/off of home appliances on the circuit to detect those activities. We also incorporate machine learning for estimating the other activities based on the circuit’s time series of power consumption. Furthermore, to cope with the difference among households, we apply transfer learning to the constructed model. In collaboration with a Japanese home builder, we conducted an experiment on five households using their HEMS data. In parallel, we obtained verifiable activity labels as our ground truth by the installation of specialized sensors in the respective homes. Under a ±30 min tolerance (i.e. allowing a prediction in the immediately preceding or following half-hour slot), our model achieved an average F1 score of 0.689 across all activities. We also confirmed that transfer learning improved the F1 score of each activity recognition and achieved an average improvement of 0.260 in household-level waking-up, household-level going-to-bed, room-level waking-up, room-level going-to-bed, and bathing activities.
Research Note
In recent years, HEMS (Home Energy Management System), which monitors the amount of electricity and gas used in the home and "automatically controls" home appliances, has been widespread. The government has set the goal of installing HEMS in all households by 2030 in its "Green Policy Framework," therefore, HEMS is expected to spread and be utilized more widely than ever before.
If the activity of residents can be identified from HEMS data, it can be used for various applications such as feedback on energy consumption to residents, profiling of consumers based on their energy use trends, and short- and long-term prediction of energy demand.
In addition to demand forecasting, it will enable low-cost applications of big data such as customer profiling and target marketing, estimation of household behavior trends, and remote healthcare, including monitoring of the elderly.
Although various activity recognition methods based on power consumption data have been studied, all previous studies assumed relatively high temporal granularity of power data.
In this study, we propose a method for estimating in-home behavior based only on the cumulative power consumption of each branch circuit every 30 minutes, which is obtained from the HEMS distribution board.
In this method, features that can be recognized at a low granularity are specially designed for each activity, and each activity is estimated using transition learning.
Published Papers
- 田中 福治, 石津 紘太朗, 水本 旭洋, 山口 弘純, 東野 輝夫, "低粒度な分岐回路電力データを用いた家庭内行動認識手法", マルチメディア,分散, 協調とモバイル (DICOMO 2021) シンポジウム論文集, pp.1391-1399,情報処理学会, 2021 年 7 月 , 最優秀論文賞
- 田中 福治, 水本 旭洋, 山口 弘純, 東野 輝夫, “複雑な建築図面における部屋のセマンティクス情報の抽出,” モバイルコンピューティングと新社会システム (MBL-101) , pp1-8, 2021年11月, 優秀論文賞.
- 田中 福治, 水本 旭洋, 山口 弘純, "HEMS電力データを用いた家庭内行動認識手法の実家庭における評価", マルチメディア,分散, 協調とモバイル (DICOMO 2022) シンポジウム論文集, pp.389-399,情報処理学会, 2022 年 7 月
- 田中 福治, 水本 旭洋, 山口 弘純, "低粒度な分岐回路電力データを用いた家庭内行動認識手法",情報処理学会論文誌, 64巻, 4号, 2023 年
- 田中 福治, 水本 旭洋, 山口 弘純, “時系列データを扱うニューラルネットワークによる低粒度HEMSデータからの行動認識の精度検証,” マルチメディア,分散, 協調とモバイル (DICOMO 2023) シンポジウム論文集, pp.93-99,情報処理学会, 2023年7月.
- Fukuharu Tanaka, Teruhiro Mizumoto, Hirozumi Yamaguchi, “Recognition of House Structures from Complicated Electrical Plan Images,” Information, Vol.15, No.3, pp. 147, March, 2024.
- Fukuharu Tanaka, Teruhiro Mizumoto, Hirozumi Yamaguchi, “Recognizing Home Activity from Coarse Branch Circuit Energy Usage Data” in Proceedings of the 13th International Workshop on the Reliability of Intelligent Environments (WoRIE 2024), pp. 6-15, June, 2024 (Workshop).
