at University of California, Irvine

Research

The following graph illustrates the main research directions in our project. The main goal is to provide a cyber-physical design framework, in which one can explore and evaluate the aspects of the cyber-physcial design space. We break down the main goal to five smaller research activities, which are shortly outlined below. You may also follow the links in the graph to learn more about the research activities.

We address the challenges of Cyber-Physical Systems Design in the following Sub-Projects.

  • Component Based Design of CPS
    The design of Cyber-physical systems is a complex, multi-disciplinary task. In this work we present a framework for component-based CPS design in which the components may originate from the physical and the cyber domain. In this framework we can investigate for instance performance, control quality, cyber-physical energy consumption, and security.
    A detailed use case is the inverted pendulum example.
  • Real-time scheduling with shared resources
    Real-time scheduling in CPSs is challenged by two major factors that are not considered in classic scheduling approaches: 1) transient overloads originating from fluctuating data and event flows or from faults, and 2) real-world physical resources for which access cannot easily be rolled back or preempted. To manage processing time and resource access of control applications running on shared processors we designed and implemented a hierarchical scheduler that applies PID control techniques to provide temporal isolation between the applications. The approach was simulated in RTSim and implemented on VxWorks.
  • Adding awareness for Faults and Robustness into the design process
    CPSs in practice are exposed to a range of anomalies and faults in all parts of the system. It is our goal to add fault models and robustness requirements as inputs of our framework and to propose system configurations that can cope with those fault models. The goal here is not only to find one specific solution but to develop design flows that integrate the fault and QoS models into the decision making, and application mapping. Our results include an extended study on control and sensor redundancy strategies for the Falling Ball Example  and sensor and timing faults in systems such as smart buildings.
  • Digital Mockup Devices
    "Digital Mockups" is a joint UC Riverside/Irvine research project developing real-time observable/controllable executable models of physical systems, such as the human lung.