ABSTRACT
Knowing indoor occupancy states is crucial for energy optimization in buildings. While neural networks can effectively be used to detect occupancy based on carbon dioxide measurements, their application is impeded by the need for sufficient labeled training data. In this study, we analyze the prediction performance of three different transfer learning (TL) methods leveraging target room data jointly with data from other rooms. The methods include (1) pretraining and fine-tuning, (2) layer freezing, and (3) domain-adversarial learning. Using data from five real-world rooms and one simulated room, including multiple room types, we provide the most extensive evaluation of TL in the field of occupancy prediction from environmental variables to date. This work’s contribution further includes the architecture and hyperparameters of a deep CNN-LSTM model for CO2-based occupancy detection. Our results indicate that TL effectively reduces the required amount of target room data. Moreover, while previous literature was focused on pretraining with related real-world data, we show that similar performance can be achieved by the more practical approach of leveraging simulated data.
- Krzysztof Arendt, Aslak Johansen, Bo Nørregaard Jørgensen, Mikkel Baun Kjærgaard, Claudio Giovanni Mattera, Fisayo Caleb Sangogboye, Jens Hjort Schwee, and Christian T. Veje. 2018. Room-level occupant counts, airflow and CO 2 data from an office building. First Workshop on Data Acquisition To Analysis (DATA) (2018), 13–14.Google ScholarDigital Library
- Irvan B. Arief-Ang, Margaret Hamilton, and Flora D. Salim. 2018. A Scalable Room Occupancy Prediction with Transferable Time Series Decomposition of CO2 Sensor Data. ACM Transactions on Sensor Networks 14, 3-4 (2018), 1–28.Google ScholarDigital Library
- Irvan B. Arief-Ang, Flora D. Salim, and Margaret Hamilton. 2017. DA-HOC: Semi-Supervised Domain Adaptation for Room Occupancy Prediction using CO2 Sensor Data. Proceedings of the 4th ACM International Conference on Systems for Energy-Efficient Built Environments (BuildSys) (2017), 1–10.Google ScholarDigital Library
- Farzan Banihashemi and Manuel Weber. 2023. “Non-intrusive indoor environmental data in two double-occupied offices”. Mendeley Data (2023). https://doi.org/10.17632/v7kw9ycccp.1Google ScholarCross Ref
- Farzan Banihashemi, Manuel Weber, Fatma Deghim, Chujun Zong, and Werner Lang. 2023. Occupancy modeling on non-intrusive indoor environmental data through machine learning. SSRN (2023). https://doi.org/10.2139/ssrn.4477794Google ScholarCross Ref
- Vincent Becker and Wilhelm Kleiminger. 2018. Exploring zero-training algorithms for occupancy detection based on smart meter measurements. Computer Science - Research and Development 33, 1 (2018), 25–36.Google ScholarCross Ref
- Jonathan Brooks and Prabir Barooah. 2014. Energy-efficient control of under-actuated HVAC zones in buildings. 2014 American Control Conference (2014), 424–429.Google ScholarCross Ref
- Davide Calì, Peter Matthes, Kristian Huchtemann, Rita Streblow, and Dirk Müller. 2015. CO2 based occupancy detection algorithm: Experimental analysis and validation for office and residential buildings. Building and Environment 86 (2015), 39–49.Google ScholarCross Ref
- Luis M. Candanedo and Véronique Feldheim. 2016. Accurate occupancy detection of an office room from light, temperature, humidity and CO2 measurements using statistical learning models. Energy and Buildings 112 (2016), 28–39.Google ScholarCross Ref
- Luis M. Candanedo, Véronique Feldheim, and Dominique Deramaix. 2017. A methodology based on Hidden Markov Models for occupancy detection and a case study in a low energy residential building. Energy and Buildings 148 (2017), 327–341.Google ScholarCross Ref
- Dong Chen, Sean Barker, Adarsh Subbaswamy, David Irwin, and Prashant Shenoy. 2013. Non-Intrusive Occupancy Monitoring Using Smart Meters. Proceedings of the 5th ACM Workshop on Embedded Systems For Energy-Efficient Buildings (BuildSys) (2013), 1–8.Google ScholarDigital Library
- Zhenghua Chen, Rui Zhao, Qingchang Zhu, Mustafa K. Masood, Yeng Chai Soh, and Kezhi Mao. 2017. Building Occupancy Estimation with Environmental Sensors via CDBLSTM. IEEE Transactions on Industrial Electronics 64, 12 (2017), 9549–9559.Google ScholarCross Ref
- Jacob Cohen. 1960. A Coefficient of Agreement for Nominal Scales. Educational and Psychological Measurement 20, 1 (1960), 37–46.Google ScholarCross Ref
- Jawher Dridi, Manar Amayri, and Nizar Bouguila. 2022. Transfer learning for estimating occupancy and recognizing activities in smart buildings. Building and Environment 217 (2022), 109057.Google ScholarCross Ref
- Afrooz Ebadat, Giulio Bottegal, Damiano Varagnolo, Bo Wahlberg, and Karl H. Johansson. 2013. Estimation of Building Occupancy Levels through Environmental Signals Deconvolution. Proceedings of the 5th ACM Workshop on Embedded Systems For Energy-Efficient Buildings (BuildSys) (2013), 1–8.Google ScholarDigital Library
- Hassan Ismail Fawaz, Germain Forestier, Jonathan Weber, Lhassane Idoumghar, and Pierre-Alain Muller. 2018. Transfer learning for time series classification. IEEE International Conference on Big Data (2018), 1367–1376.Google Scholar
- Yaroslav Ganin, Evgeniya Ustinova, Hana Ajakan, Pascal Germain, Hugo Larochelle, François Laviolette, Mario Marchand, and Victor Lempitsky. 2016. Domain-Adversarial Training of Neural Networks. Journal of Machine Learning Research 17 (2016), 1–35.Google ScholarDigital Library
- Hwataik Han, Kyung-jin Jang, Changho Han, and Junyong Lee. 2013. Occupancy estimation based on CO2 concentration using dynamic neural network model. Proc. 34th Air Infiltration and Ventilation Centre Conference (AIVC) (2013).Google Scholar
- IEA. 2022. World Energy Outlook 2022. https://www.iea.org/reports/world-energy-outlook-2022Google Scholar
- Chaoyang Jiang, Zhenghua Chen, Rong Su, Mustafa Khalid Masood, and Yeng Chai Soh. 2020. Bayesian filtering for building occupancy estimation from carbon dioxide concentration. Energy and Buildings 206 (2020), 109566.Google ScholarCross Ref
- Ming Jin, Nikolaos Bekiaris-Liberis, Kevin Weekly, Costas J. Spanos, and Alexandre M. Bayen. 2018. Occupancy Detection via Environmental Sensing. IEEE Transactions on Automation Science and Engineering 15, 2 (2018), 443–455.Google ScholarCross Ref
- Jonathan Brooks, Siddharth Goyal, Rahul Subramany, Yashen Lin, Timothy Middelkoop, Laura Arpan, Luca Carloni, and Prabir Barooah. 2014. An experimental investigation of occupancy-based energy-efficient control of commercial building indoor climate. IEEE Conference on Decision and Control 53 (2014), 5680–5685.Google ScholarCross Ref
- Mohamad Khalil, Stephen McGough, Zoya Pourmirza, Mehdi Pazhoohesh, and Sara Walker. 2021. Transfer Learning Approach for Occupancy Prediction in Smart Buildings. 12th International Renewable Engineering Conference (IREC) (2021), 1–6.Google Scholar
- Pawalai Kraipeerapun and Somkid Amornsamankul. 2017. Room Occupancy Detection using Modified Stacking. Proceedings of the 9th International Conference on Machine Learning and Computing (ICMLC) (2017), 162–166.Google ScholarDigital Library
- William O’Brien, Andreas Wagner, Marcel Schweiker, Ardeshir Mahdavi, Julia Day, Mikkel Baun Kjærgaard, Salvatore Carlucci, Bing Dong, Farhang Tahmasebi, Da Yan, Tianzhen Hong, H. Burak Gunay, Zoltan Nagy, Clayton Miller, and Christiane Berger. 2020. Introducing IEA EBC annex 79: Key challenges and opportunities in the field of occupant-centric building design and operation. Building and Environment 178 (2020), 106738.Google ScholarCross Ref
- Sinno Jialin Pan and Qiang Yang. 2010. A Survey on Transfer Learning. IEEE Transactions on Knowledge and Data Engineering 22, 10 (2010), 1345–1359.Google ScholarDigital Library
- Mikko Rinta-Homi, Naser Hossein Motlagh, Agustin Zuniga, Huber Flores, and Petteri Nurmi. 2021. How Low Can You Go? Performance Trade-Offs in Low-Resolution Thermal Sensors for Occupancy Detection: A Systematic Evaluation. Proc. ACM Interact. Mob. Wearable Ubiquitous Technol. 5, 3 (2021).Google ScholarDigital Library
- Yarden Rotem, Nathaniel Shimoni, Lior Rokach, and Bracha Shapira. 2022. Transfer Learning for Time Series Classification Using Synthetic Data Generation. Cyber Security, Cryptology, and Machine Learning (CSCML): 6th International Symposium (2022), 232–246.Google Scholar
- Luis Rueda, Kodjo Agbossou, Alben Cardenas, Nilson Henao, and Sousso Kelouwani. 2020. A comprehensive review of approaches to building occupancy detection. Building and Environment 180 (2020), 106966.Google ScholarCross Ref
- Aya Nabil Sayed, Yassine Himeur, and Faycal Bensaali. 2022. Deep and transfer learning for building occupancy detection: A review and comparative analysis. Engineering Applications of Artificial Intelligence 115 (2022), 105254.Google ScholarDigital Library
- Davor Stjelja, Juha Jokisalo, and Risto Kosonen. 2022. Scalable Room Occupancy Prediction with Deep Transfer Learning Using Indoor Climate Sensor. Energies 15, 6 (2022).Google Scholar
- Ryan Turner, David Eriksson, Michael McCourt, Juha Kiili, Eero Laaksonen, Zhen Xu, and Isabelle Guyon. 2021. Bayesian Optimization is Superior to Random Search for Machine Learning Hyperparameter Tuning: Analysis of the Black-Box Optimization Challenge 2020. Proc. NeurIPS 2020, PMLR 133 (2021), 3–26.Google Scholar
- Manuel Weber, Maximilian Auch, Christoph Doblander, Peter Mandl, and Hans-Arno Jacobsen. 2021. Transfer Learning With Time Series Data: A Systematic Mapping Study. IEEE Access 9 (2021), 165409–165432.Google ScholarCross Ref
- Manuel Weber, Christoph Doblander, and Peter Mandl. 2020. Detecting Building Occupancy with Synthetic Environmental Data. Proceedings of the 7th ACM International Conference on Systems for Energy-Efficient Buildings, Cities, and Transportation (BuildSys) (2020), 324–325.Google ScholarDigital Library
- Manuel Weber, Christoph Doblander, and Peter Mandl. 2020. Towards the Detection of Building Occupancy with Synthetic Environmental Data: arXiv:2010.04209.Google Scholar
- Yixuan Wei, Shu Wang, Longzhe Jin, Yifei Xu, and Tianqi Ding. 2022. Indoor occupancy estimation from carbon dioxide concentration using parameter estimation algorithms. Building Services Engineering Research and Technology 43, 4 (2022), 419–438.Google ScholarCross Ref
- Jörg Wicker, Nicolas Krauter, Bettina Derstorff, Christof Stönner, Efstratios Bourtsoukidis, Thomas Klüpfel, Jonathan Williams, and Stefan Kramer. 2015. Cinema Data Mining: The Smell of Fear. Proc. 21th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (2015), 1295–1304.Google ScholarDigital Library
- Dominic Wörner, Thomas von Bomhard, and Felix Wortmann. 2013. Occupancy-Based Heating Control for Residential Buildings Using Environmental Sensors. Proceedings of the 5th ACM Workshop on Embedded Systems For Energy-Efficient Buildings (BuildSys) (2013), 1–2.Google ScholarDigital Library
- Tianyu Zhang and Omid Ardakanian. 2019. A domain adaptation technique for fine-grained occupancy estimation in commercial buildings. Proc. 2019 Internet of Things Design and Implementation (IoTDI) (2019), 148–159.Google ScholarDigital Library
- M. S. Zuraimi, A. Pantazaras, K. A. Chaturvedi, J. J. Yang, K. W. Tham, and S. E. Lee. 2017. Predicting occupancy counts using physical and statistical Co2-based modeling methodologies. Building and Environment 123 (2017), 517–528.Google ScholarCross Ref
Index Terms
- Overcoming Data Scarcity through Transfer Learning in CO2-Based Building Occupancy Detection
Recommendations
ABODE-Net: An Attention-based Deep Learning Model for Non-intrusive Building Occupancy Detection Using Smart Meter Data
Smart Computing and CommunicationAbstractOccupancy information is useful for efficient energy management in the building sector. The massive high-resolution electrical power consumption data collected by smart meters in the advanced metering infrastructure (AMI) network make it possible ...
Deep and transfer learning for building occupancy detection: A review and comparative analysis
AbstractThe building internet of things (BIoT) is quite a promising concept for curtailing energy consumption, reducing costs, and promoting building transformation. Besides, integrating artificial intelligence (AI) into the BIoT is essential ...
Highlights- The first review of deep and transfer learning for building occupancy detection.
Deep Transfer Learning For Abnormality Detection
ICCSE'19: Proceedings of the 4th International Conference on Crowd Science and EngineeringDeep learning has proven to be effective in learning scenarios with massive training data. However, in many real applications (i.e., abnormality detection), there is a lack of sufficient data to achieve a good deep learning model. Considering the fact ...
Comments