Toward Automatic Synthesis of Schedulable Real-Time Controllers

Abstract

This paper presents an automatic approach to synthesizing schedulable timing constraints for real-time control systems. Given the performance speci cations and schedulability constraints of a real-time control system, the approach derives task-level timing constraints which can guarantee these requirements. The control performance is speci ed in terms of control output responses such as steady state error, maximum overshoot, settling time, and rise time, and the task-level timing constraints include task periods and deadlines. The approach consists of two components with a clean interface. The rst component translates the performance speci cations into a set of system-level timing constraints such as loop processing periods and input-to-output latency, via control theoretic modeling and optimization. The second then derives task-level timing constraints from the intermediate system-level timing constraints optimizing the system schedulability using period calibration method 4, 17, 9]. Our approach contributes to both the control and real-time areas: (1) it allows control engineers to take into consideration the e ect of task scheduling at the early stage of system design; and (2) it makes it possible to streamline the design of real-time control systems, since timing constraints are derived in an automatic manner.

Fundamentals of RTOS-Based Digital Controller Implementation

“…For example, the control system shown in Fig. 1 may include a task that samples the controlled system, a task that generates the reference signal, a task that perform the computation of the control algorithm, and an output task that communicates with the actuator, which results in the task graph [19] of Fig. 2.…”

“…The performance of a single-input single-output control system [19] can be characterized by the following features of its response to a step input: maximum overshoot, the rise time, the settling time in the transient state, and the steady-state error. The maximum overshoot M peak is defined as the difference between the reference input and the maximum value of the response curve.…”

“…The mathematical model devised in [19] expresses M peak , T rise , T set , and E ss as functions of sampling period and latency. Solving these functions results in values for the period and latency that satisfy the system performance specifications.…”

Handling execution overruns in hard real-time control systems

“…Finally, in [15], Ryu and Hong described a design methodology to synthesizing schedulable timing constraints for real-time control systems, however no runtime mechanisms are described to handle online variations of tasks' execution times.…”

Sampling Period Dependent RST Controller Used in Control/Scheduling Co-Design

“…the resource constraints are obtained by solving an optimization problem. In (Ryu and Hong, 1998) control law performances depend on the periods as well as on input/output latencies and an heuristic algorithm computes the optimal periods. In (Palopoli et al, 2002) an optimization problem is solved to find optimal task periods and feedback gains.…”