The unprecedented rise in global temperatures, extreme weather events, and the depletion of natural resources underscore the urgent need for sustainable practices. Energy-saving solutions are pivotal in mitigating these challenges by reducing carbon emissions, lessening our reliance on finite energy sources, and ultimately contributing to a more resilient and environmentally conscious future. Thus, this paper presents a novel approach to enhancing energy efficiency within residential environments by integrating a Digital Twin (DT) on the Home-Assistant platform. Home-Assistant provides a user-centric approach to the DT technology, allowing homeowners to establish and oversee virtual replicas of their living spaces by incorporating a variety of Internet of Things (IoT) devices and sensors. Leveraging the capabilities of this open-source home automation platform, we have developed a sophisticated system for providing real-time energy consumption data, personalized energy-saving recommendations, and a data-driven occupancy detection mechanism. This system was rigorously tested in a controlled laboratory environment with multiple users, simulating diverse household scenarios. The DT technology enabled the creation of accurate virtual representations of users' physical environment, facilitating the optimization of energy-intensive devices and systems. Our data-driven occupancy detection approach utilized Machine Learning (ML) algorithms to intelligently determine room occupancy, allowing for precise energy management based on real-time usage patterns. The occupancy detection approach has proven highly effective, with a testing accuracy rate of 95.12% and f1-scores of 94.55%. Additionally, a mini-pilot study evaluated the recommender system and found an outstanding 80% favorable user reaction, demonstrating its efficiency in giving energy-saving advice. Moreover, the DT comprehensively represented the user's state of presence, their engagement with the connected appliances, and the ambient conditions of the lab. These results demonstrate significant potential for reducing energy waste and cost while maintaining user comfort. This research contributes to the growing field of smart home technology by showcasing the practical implementation of DT and data-driven strategies to promote sustainable and efficient energy practices in everyday living spaces.

中文翻译：

---

提高建筑可持续性：用于能源效率和占用监控的数字孪生方法


全球气温空前上升、极端天气事件和自然资源枯竭凸显了对可持续实践的迫切需求。节能解决方案通过减少碳排放、减少我们对有限能源的依赖并最终为更具弹性和环保意识的未来做出贡献，在缓解这些挑战方面发挥着关键作用。因此，本文提出了一种通过在 Home-Assistant 平台上集成数字孪生 （DT） 来提高住宅环境能源效率的新方法。Home-Assistant 提供了一种以用户为中心的 DT 技术方法，允许房主通过整合各种物联网 （IoT） 设备和传感器来建立和监督其生活空间的虚拟副本。利用这个开源家庭自动化平台的功能，我们开发了一个复杂的系统，用于提供实时能耗数据、个性化的节能建议和数据驱动的占用检测机制。该系统在多用户受控实验室环境中进行了严格测试，模拟了不同的家庭场景。DT 技术能够创建用户物理环境的准确虚拟表示，从而促进能源密集型设备和系统的优化。我们的数据驱动型占用检测方法利用机器学习 （ML） 算法来智能地确定房间占用率，从而根据实时使用模式进行精确的能源管理。占用检测方法已被证明非常有效，测试准确率为 95.12%，f1 分数为 94.55%。 此外，一项小型试点研究评估了推荐系统，发现用户反应良好，达到 80%，证明了它在提供节能建议方面的效率。此外，DT 全面表示用户的状态、他们与互联设备的互动以及实验室的环境条件。这些结果表明，在保持用户舒适度的同时，减少能源浪费和成本具有巨大潜力。本研究通过展示 DT 和数据驱动策略的实际实施，在日常生活空间中促进可持续和高效的能源实践，为不断增长的智能家居技术领域做出贡献。