Cyber-Physical Systems/IoT


此項計畫將研究與設計一個開放式嵌入式軟體系統架構,以提供一個健全且具有彈性的智慧整合感控系統(Cyber-Physical Systems, CPS)及服務的系統架構基礎。我們提出的中介軟體雛形將會整合計畫成員在智慧整合感控系統(CPS)工作流程管理與開放式即時概念的研究結果,並且將之具體化為可立即應用之形式。此一創新的中介軟體稱為分散式即時工作流程架構 (DiReWF)。 DiReWF把廣泛用於商業流程自動化的工作流程的概念應用到智慧整合感控系統(CPS)流程自動化,提供一個以工作流程模式為基礎的智慧整合感控系統(CPS)整合架構,以整異質應用元件、應用元件間的工作流程、以及定義和整合元件的工具,並保證系統的安全性、可靠性、服務品質以及柔韌性。DiReWF可減少設計、實作和維護智慧整合感控系統(CPS)服務所需的專業人力與物力,特別是自動化和家居生活輔助服務。此外,藉由管理點對點資源和服務品質(QoS)的機制和有效率的演算法,DiReWF還可以支援即時系統概念,此概念正是有嚴格時間限制的服務所需要的。最後,藉由提供實體數位系統(CPS)、應用元件、設備模型、服務模型、使用者模型和使用者行為模型的擴充型函式庫,DiReWF可支援模擬和評估廣泛的智慧整合感控系統(CPS)和使用情境。


Thanks to steady technological advances over the years in sensors, sensor networks and information systems, embedded and real-time devices and systems, mobile communication and computing, automation and robotics, and so on, we now witness the emergence of an ever broader spectrum of cyber-physical systems, as well as cyber-physical services that rely on such systems for execution and delivery.  Examples cyber-physical services include

  • Services for the elderly: Many consumer and assistive cyber-physical services for the purposes of enhancing quality of life and self-reliance of elderly or functionally limited individuals are enabled by personal and home automation and assistive devices and systems (e.g., [1-7]) developed by the Academia Sinica thematic project SISARL (Sensor Information Systems for Active Retirees and Assisted Living) [8] and other projects (e.g., [9-11]) on technologies for independent living. The demands for such services will surely accelerate as the global population ages in the coming decades.
  • Tele-assistance services: Services such as those described in [12 – 14] allow individuals who need close professional supervision and fully integrated health and medical care services to be monitored and supervised by their care providers while living at home with their families.
  • User-centric Automation Services: Services such as those described in [iNuC] allow individuals who can benefit the automation service to shorten their workflow, enhance the security and quality of their work. An intelligent nurse cart is one example of such services. Existing nurse carts provide fundamental services for the nurses. However, the heavy workload and complex medication administration safety workflow creates heavy work stress to the nurses, which may lead to medication errors in return. Intelligent nurse carts assist the nurse the manage their work schedule, monitor their operations, control the medication drawers according to the patient ID and medication administration schedule, and automatically upload their work record to the Hospital Information System (HIS). As a result, it reduces their work stress to avoid medication errors.
  • Personal services: An ever more diverse spectrum of cyber-physical personal services, ranging from safety and security monitor and alert, to group coordination and travel guides, to fashion advices at homes and in stores, etc. are made possible by wearable sensors, low-cost RFID readers and tags, indoor and outdoor location systems, mobile devices and web and web 2.0 technologies.

Hereafter, we refer to services illustrated by these examples as users-centric cyber-physical services (CPS), or simply as user-centric services, omitting the adjective “cyber-physical”  where this aspect requires no clarification or reminder. We use the acronym CPS to mean cyber-physical systems or cyber-physical services interchangeable.

According to its numerous definitions (e.g., [15-17]), a cyber-physical system tightly integrate computational and physically processes. In a CPS, “embedded computers and networks monitor and control physical processes, usually with feedback loops where physical processes affect computations and vice versa.” It is fair for one to observe that defined as such, cyber-physical systems and services are not new. They have been providing services vital to us for years by coordinating and regulating our electric power grids, monitoring and controlling air and vehicular traffic, managing strategy weapon systems, and so on. These classical CPS are typically machine-centric, however, according to the classification of devices and services based on their interactions with human users [18]. Machine-centric systems and services are for non-discretionary use: Their users (e.g., pilots and drivers) have essentially no choice but to use them. Consequently, they can demand sophisticated (and often expensive) operating environments and infrastructures, as well as intense user training needed to keep “human factors” and automation surprises in check. In contrast, user-centric services are typically for discretionary use [18]. They must be affordable and easy to use. It is not practical to require their users more than minimal training, if any training at all. In this project, we will hence focus on developing the enabling technology for user-centric cyber-physic systems.



The research focus of this project is on an architectural framework for integration of sensors, devices and other system and application components of user-centric cyber-physical systems and services. The proposed research is motivated by the need to overcome the following three challenges in building and delivering these systems and services.

First, as tools for improving quality of life of their users, helping their users stay well and live independently, or improving the quality and reducing the cost of care, 21-century services must be flexible. By a service being flexible, we mean that it is configurable, customizable and can adapt: It can be easily configured to work with a variety of sensors and control devices, rely on different support infrastructures and operate in different environments. It can be easily customized to suit its user. The purpose of many 21-century services (e.g., services for the elderly and tele-assistance services) is to compensate for the user’s skills and weaknesses. Such services may be in use for years, even decades, and must be able to adapt to changes in its user’s needs. Even today, all but the simplest cyber-physical systems and applications are handcrafted. Handcrafted systems, and services they can support, are often difficult and costly to configure, customize and be made to adapt.

The second challenge is how to make sure that CPS-based services and their users work as coherent symbiotic systems [19] in a safe and sound way. By a symbiotic system, we mean an entity consisting of a service, accesses to the service, the user or users, and user-service interactions. By it being safe and sound, we mean that it never does any harm and all unavoidable errors are either recoverable or tolerable. This problem arises from the fact that full automation is often not economically feasible and sometimes even undesirable. The user often may need to perform mission-critical functions. There have been extensive works on user models, formal verification methods, and runtime monitoring techniques, etc. for the purpose of keeping the user(s) safe and sound as parts of the symbiotic system changes. A typical assumption is that the device (or service as in our case) is used as intended. This assumption is usually valid for machine-centric devices with their well trained users, but is not valid for user-centric devices and services. Ensuring safe and sound operations is made more challenging by the fact that not only the users may be untrained but also their skills vary widely among the user population and for an individual user over time.

The third challenge arises from the need for appropriate real-time systems abstractions. This is one of the major challenges of the cyber-physical system research identified by the real-time systems community ([16, 17]). It is of particular importance for mobile systems and services. Like classical cyber-physical services, many user-centric services have critical timing requirements. Classical services typically operate in closed environments; run on private, dedicated computers and networks; and many use custom made sensors and controllers solely. Traditional real-time techniques for ensuring predictable responses, coherent time order and consistent system states are applicable only in closed environments. Many factors, including design, development and maintenance costs, force user-centric century services to be open. How to ensure predictable and timely responses for services delivered over open systems is challenging indeed.

News related to CPS systems: