PFI:BIC - A Cost-effective Accurate and Resilient Indoor Positioning System

“创新伙伴关系：建设创新能力”项目旨在开发一种成本效益高、准确、有弹性和智能的室内定位服务，用于建筑环境。定位系统彻底改变了我们与周围世界的互动方式。户外移动设备利用全球定位系统（GPS）等技术来提供各种各样的基于位置的服务。同样，室内定位系统将提供新的服务，为住宅和商业建筑环境中的人类提供巨大的社会和商业价值。室内定位服务可以被企业用来跟踪和管理资产。建筑物管理系统可以使用室内位置信息为建筑物管理人员、居住者和第一响应者提供服务，例如有效的应急响应、室内导航和周边保护。此外，室内定位服务将能够实施重要服务，例如在灾害情景中协调人员（例如，自然或人为（公共枪击）灾害和盲人导航服务）。不幸的是，基于卫星的方法，如GPS，不能在室内工作，因为卫星信号很弱，不能穿透建筑物的外墙。与现有方法不同，拟议的智能服务将在中断方面实现高精度和鲁棒性，同时保持较低的安装和维护成本。此外，用户将能够使用他们的移动设备（例如，智能手机、平板电脑、智能手表），而无需携带/佩戴额外的设备。该项目将开发并结合超声波、可见光和基于无线局域网（WLAN）的定位技术，以及基于射频（RF）的磁签名、人体移动模型和建筑信息模型（bim），用于定位、跟踪和可视化。几种独立定位技术的联合使用不仅会大大提高定位的准确性，而且还会增加必要的冗余，以承受除一个定位服务外的所有定位服务的中断，并且性能损失是有限的。即使在所有定位技术都不可用的情况下，移动模型与BIM一起，将能够在更粗的粒度水平上提供室内定位。反过来，冗余可以用于在不中断业务的情况下对任何子系统进行维护和定期系统校准。所提出的方法令人印象深刻的特点是，所有这些特性都可以以较低的安装和维护成本实现，因为该系统可以搭载建筑物现有的音频、照明和射频通信功能。该定位算法的独特之处在于其模块化和可扩展性。来自不同传感器的信息将被无缝地整合在一起，使算法能够在其中一个子组件间歇性故障的情况下工作。包括移动模型，以及加速度计，陀螺仪和罗盘数据，在今天的大多数智能手机上，将允许实现细粒度跟踪，这将提供平滑的轨迹，而不是位置序列。在本方案中，多传感器定位与BIM协同作用。通过提供精确定位测距源（如光、超声、Wi-Fi天线）和精确的拓扑信息来开发高保真测距模型，BIM将有助于降低安装和维护成本。此外，BIM提供的语义信息将有助于检测不可行的轨迹。另一方面，基于同步定位和映射（SLAM）的技术可以帮助改进bim并保持更新。动态信息可以通过向建筑管理者提供有关交通模式和占用情况的有用信息来增强bim。重要的是，智能服务的设计需要以人为中心，并考虑到每个利益相关者，即所有者和设施管理团队，服务开发人员，智能服务应用程序接口（API）的用户，他们将为特定设施或更普遍地为许多设施开发增值服务，当然，最终用户，设施的居住者和访客，他们将自己使用智能服务。为了了解这些不同利益相关者群体的需求和愿望，项目将通过开展一系列焦点小组直接让他们参与进来。参与式设计是一种成熟的技术，在这种技术中，设计团队直接与涉众一起设计工件或服务。涉众还将参与软件服务的正式测试，从安装到维护，到应用程序设计和应用程序使用。在项目开始时，合作伙伴包括牵头机构：宾夕法尼亚州匹兹堡的卡内基梅隆大学（电气与计算机工程系、土木与环境工程系和人机交互研究所），主要合作伙伴：博世匹兹堡RTC（宾夕法尼亚州匹兹堡，大企业）和体育与展览局（宾夕法尼亚州匹兹堡，大企业）。

This Partnerships for Innovation: Building Innovation Capacity project aims at developing a cost effective, accurate, resilient and smart indoor localization service to be used in built environments. Positioning systems have revolutionized how we interact with the world around us. Outdoor mobile devices make use of technologies like Global Positioning System (GPS) to deliver a wide variety of location-based services. Similarly, indoor positioning systems will enable delivery of new services that provide tremendous social and commercial value to humans in residential and commercial built environments. Indoor location services can be used by enterprises to track and manage assets. Building management systems can use indoor location information to enable services for building managers and occupants and first responders, such as effective emergency response, indoor navigation, and perimeter protection. Furthermore, indoor location services will enable implementation of important services such as coordination of people in a disaster scenario (e.g., natural or man-made (public shootings) disasters and navigation services for the blind). Unfortunately, satellite-based approaches, such as GPS, do not work indoors due to weak satellite signals that do not penetrate through building facades. Unlike existing methods, the proposed smart service will achieve high accuracy and robustness with respect to disruptions, while maintaining low installation and maintenance costs. In addition, users will be able to use their mobile device(s), (e.g., smartphone, tablets, smart watches), without the need to carry/wear additional equipment. The project will develop and combine ultrasound, visible light and Wireless Local Area Network (WLAN)-based positioning techniques with Radio Frequency (RF)-based, magnetic signatures, human ambulation models and building information models (BIMs) for localization, tracking and visualization. The combined use of several independent positioning techniquse not only will dramatically increase the accuracy of positioning over any single technique, but it will add the necessary redundancy to withstand disruption of all but one positioning service, with provably bounded loss of performance. Even in the case of unavailability of all positioning techniques, ambulation models, together with BIM, will be able to provide indoor positioning at a coarser level of granularity. In turn, redundancy can be used to perform maintenance and periodic system calibration on any subsystem without service interruption. The impressive feature of the proposed methodology is that all these properties can be achieved at low installation and maintenance costs, as the system can piggyback on a building's existing audio, lighting, and RF communication capabilities. One unique property of the proposed positioning algorithm will be its modularity and extensibility. Information coming from different sensors will be incorporated seamlessly, allowing the algorithm to work under intermittent failure of one of its subcomponents. The inclusion of ambulation models, together with accelerometer, gyroscope and compass data available on the majority of today's smartphones, will allow the achievement of fine-grain tracking, which will provide smooth trajectories in place of sequence of locations. In the proposed scheme, Multi-sensor localization and BIM play a synergetic role. BIM will contribute to decreasing installation and maintenance costs, by providing precise positioning of the sources of ranging (e.g., light, ultrasound, Wi-Fi antennas) and accurate topological information to develop high fidelity ranging models. Additionally, the semantic information provided by BIM will help with detecting infeasible trajectories. On the other hand, Simultaneous Localization and Mapping (SLAM)-based techniques can help refine BIMs and keep them updated. Dynamic information can enhance BIMs by providing useful information to building managers about traffic patterns and occupancy. Importantly, the design of the smart service needs to be human-centered and to take into account each of the stakeholders, i.e., owner and facilities management team, the service developers, the users of the smart service application program interface (API), who will develop value-added services customized for a particular facility or more generally for many facilities, and, of course, the end-users, the occupants and visitors of the facility, who will use the smart services themselves. To understand the needs and wants of such distinct groups of stakeholders, the project will directly involve them by conducting a series of focus groups. Participatory design is an established technique where a design team works directly with stakeholders to design an artifact or service. Stakeholders will also be engaged in the formal testing of the software service, from installation to maintenance, to application design and to application usage. At the inception of the project, partners include the lead institution: Carnegie Mellon University, (Departments of Electrical and Computer Engineering, Civil and Environmental Engineering, and the Human-Computer Interaction Institute) Pittsburgh, PA, with primary partners: Bosch RTC Pittsburgh (Pittsburgh, PA, large business) and Sports and Exhibition Authority (Pittsburgh, PA, large business).