Scalable Cyber-Physical-Human Systems for Intelligent and Interactive Buildings

建筑行业约占加拿大终端能源消耗总量的30%，而加拿大人90%的时间都在室内度过。在这种背景下，为建筑物配备人类交互智能系统，即网络-物理-人类系统(CPHSS)，作为一种提高建筑节能性能和室内环境质量(IEQ)的经济有效的解决方案，受到了人们的关注。拟议研究计划的长期愿景是通过实现智能和互动建筑来缓解气候变化并提高生活质量。为了实现这一愿景，我的方法是追求领域知识和机器学习(ML)的无缝集成，为智能和交互建筑开发有效、可靠和可扩展的基于AI的CPHSS。在接下来的五年里，拟议研究计划的总体目标是最终确定并证明我的开创性方法在为智能和交互建筑开发基于人工智能的CPHSS方面的有效性。我最初的重点将是写字楼运营。相应的短期目标是：(1)开发一个知识建模框架，能够将现有的建筑领域知识编码到CPHSS中，并促进知识知情的ML；(2)开发一个因果推理框架，允许调查与居住者的需求、行为及其相应的决策过程有关的新的因果知识；(3)利用目标1和目标2中描述的框架，开发和演示用于智能和交互建筑的CPHS，配备(I)可扩展的自适应最优控制器和(Ii)具有有效反馈/激励计划的高级用户界面。拟议的研究将彻底改变建筑自动化行业，使智能和交互建筑成为商品，对加拿大的能源行业产生深远影响。此外，这项研究将有助于提高加拿大的生活质量，改善IEQ。这些目标的成功完成将是CPHS研究的一个突破，特别是在建筑自动化和管理方面。拟议的研究还将有助于对高素质人员进行跨学科培训，他们将获得两个领域的知识和技能：(1)建筑能源系统和居住者；(2)机器学习、最优控制和人机交互。HQP获得的独特专业知识将使他们在连接这两个领域方面发挥重要作用，并通过将现代人工智能解决方案应用于建筑，为提高建筑能效和室内环境质量做出贡献。

The building sector accounts for about 30% of Canada's total end-use energy consumption, while Canadians spend 90% of their time indoors. In this context, equipping buildings with human interactive intelligent systems, i.e., cyber-physical-human systems (CPHSs), has received attention as a cost-effective solution to improve building energy performance and indoor environmental quality (IEQ). The long-term vision of the proposed research program is to mitigate climate change and advance the quality of life through the realization of intelligent and interactive buildings. Towards this vision, my approach is to pursue seamless integration of domain knowledge and machine learning (ML) to develop effective, reliable, and scalable AI-based CPHSs for intelligent and interactive buildings. Over the next five years, the overall objective of the proposed research program is to conclusively determine and demonstrate the efficacy of my pioneering approach that integrates domain knowledge and ML in developing AI-based CPHSs for intelligent and interactive buildings. My focus will initially be on office building operations. The corresponding short-term objectives are to: (1)Develop a knowledge modelling framework that enables encoding of existing building domain knowledge into CPHSs and facilitates knowledge-informed ML; (2)Develop a causal inference framework that allows investigating new causal knowledge related to occupants' needs, behaviours, and their corresponding decision-making processes; (3)Develop and demonstrate a CPHS for intelligent and interactive buildings, equipped with (i) a scalable adaptive optimal controller and (ii) an advanced user interface with an effective feedback/incentive program, by utilizing the frameworks described in Objective 1 and 2. The proposed research will revolutionize the building automation industry so that intelligent and interactive buildings become commodities resulting in a profound impact on the energy sectors in Canada. In addition, the research will contribute to advancing the quality of life in Canada with improved IEQ. The successful completion of the objectives will be a breakthrough in CPHS research, in general, and building automation and management, in particular. The proposed research will also contribute to the interdisciplinary training of highly qualified personnel who will gain knowledge and skills in two domains: (i) building energy system and occupant and (ii) machine learning, optimal control, and human-machine interaction. The unique expertise HQP acquires will allow them to take a significant role in bridging these two domains and contribute to improving building energy performance and indoor environmental quality by applying modern AI solutions to buildings.