Probabilistic Real-Time Data Access with Deadline and Interval Constraints

In this chapter, we address the issues of real-time data access in the context of manufacturing control applications using a combination of simulations and measurements based on the University of Michigan Open-Architecture Controller (UMOAC) testbed. Figure 1 shows the basic control loop of a typical real-time application. The controller takes the input (e.g., the desired position and velocity) from the user and the feedback (e.g., the actual position and velocity) from the controlled process, and computes a new control command to send to the controlled process. An open-architecture controller consists of a set of cooperative tasks, e.g., sensor-reading tasks and control-law tasks, and needs a database management system (DBMS) to support concurrent data access and to provide well-defined interfaces between different software modules [14, 23]. Typically these tasks are subject to a range of timing constraints, which require the DBMS to provide timing guarantees, sometimes under complex operating conditions.