PFI:BIC Humans-in-the-Loop: A Critical Link to Enable Smart Control of Building Infrastructure in a Complex Service System

加州大学默塞德分校的这一创新合作伙伴关系：建设创新能力（PFI：BIC）项目为学术界和行业的合作伙伴关系提供支持，重点是开发基于人群的温度控制系统，以有效和高效地管理建筑能源费用和建筑居住者的舒适度，使大学创造的技术适应真实的商业应用。建筑物是复杂服务系统的重要背景。美国人90%的时间都在建筑物内度过。2011年，建筑物占美国能源消耗的近40%，其中75%是电能。建筑市场的能源支出是巨大的：2011年，超过4310亿美元用于能源。建筑物的气候变化足迹相应较大，而根据居住者舒适度调查的测量，建筑物的能源性能较差。建筑专业人士的适度目标“80%的居住者应该对热，空气质量，声学和照明环境感到满意”在实践中几乎从未实现过。 虽然降低能源消耗是一项战略性国策，但舒适性也不容忽视，必须考虑舒适性与能源使用之间的关系。ThermoVote是一种平台技术，用于收集和分析大量建筑用户的热舒适数据，并实时控制建筑供暖、通风和空调（HVAC）系统。它在复杂的服务系统中提供了一个关键环节，以实现对建筑基础设施的智能控制。该项目使用ThermoVote作为新建筑环境中基于人群的环境控制系统，以强化平台技术，将其与新建筑系统集成，并开发服务场景和相关组件，以大规模实施ThermoVote方法。从工程的角度来看，考虑到大量的参与者的复杂系统的设计很少被研究。从认知的角度来看，复杂的分布式认知系统很少被研究。从跨学科的角度来看，工程系统的设计很少被深入了解大量人员和技术的相互作用所产生的复杂的多反馈回路。该项目有几个关键的创新：使用真实的建筑物来展示系统的好处;调查个人和群体的行为，通过可调节的自主性来节省能源;以及使用多标准优化的新颖算法（例如，能源消耗和偏好）。还有一些悬而未决的问题：我们如何设计出由大群体控制的有效学习系统？ 一个系统应该如何描述其不断变化的行为？ 间接或直接控制的最佳方案是什么？这个项目由加州大学默塞德分校工程学院电气工程和计算机科学组的教师领导，加州大学默塞德分校管理和认知科学系的教师参与。 主要的工业合作伙伴是惠普实验室（位于加州帕洛阿尔托），它是美国一家大型企业惠普的研究部门。

This Partnerships for Innovation:Building Innovation Capacity (PFI:BIC) project from the University of California-Merced provides support for an academic-industry partnership to focus on the development of a crowd-based temperature control system to manage both building energy expenses and building occupant comfort effectively and efficiently, adapting university-created technology to real business applications. Buildings represent an important context for complex service systems. Americans spend 90% of their time inside buildings. Buildings accounted for near 40% of U.S. energy consumption in 2011, 75% of which was electrical energy. The energy expenditure in the building market is huge: in 2011, more than $431B was spent on energy. The climate-change footprint of buildings is correspondingly large, and the energy performance of buildings is poor, as measured by occupant comfort surveys. The modest goal of building professionals "that 80% of occupants should be satisfied with the thermal, air quality, acoustic, and lighting environments" is almost never met in practice. Though reducing energy consumption is a strategic national policy issue, comfort should not be overlooked, and the relationship between comfort and energy use must be taken into account. This project relies on a crowd-based control system for gathering occupant comfort data and managing building systems.ThermoVote is a platform technology for gathering and analyzing thermal comfort data from large numbers of building users and for controlling building heating, ventilation, and air conditioning (HVAC) systems in real-time. It provides a critical link in a complex service system to enable smart control of building infrastructure. This project uses ThermoVote as a crowd-based environment control system in new building environments to harden the platform technology, integrate it with new building systems, and develop service scenarios and related components to implement the ThermoVote approach on a large scale. From an engineering perspective, design of complex systems given large numbers of participants has rarely been studied. From a cognitive perspective, complex, distributed cognitive systems have rarely been studied. From an interdisciplinary perspective, design of engineering systems has rarely been done given deep knowledge of complex multiple feedback loops created by interactions of large numbers of people and technologies. There are several key innovations to the project: use of real buildings to show system benefits; investigation of both individual and group behavior to save energy by adjustable autonomy; and novel algorithms using multiple criteria optimization (e.g., energy consumption and preferences). And there are a number of open questions: How can we design effective learning systems that are controlled by large groups? How should a system portray its changing behavior? What are the best schemes for indirect or direct control? Many questions lie at the intersection of systems and people.This project is led by faculty in the Electrical Engineering and Computer Science group of the School of Engineering at the University of California, Merced, with participation by faculty in Management and Cognitive Science at the University of California, Merced. The primary industrial partner is HP Labs (Palo Alto, California), the research arm of Hewlett Packard, a large U.S. business.