Run-Time Models for Measurement and Control Systems and Their Support in Ptolemy II
Jie Liu, Stan Jefferson, and Edward A. Lee,
UC Berkeley/Agilent Technologies, Palo Alto, CA
liuj@eecs.berkeley.edu
Measurement and control systems are intrinsically distributed and real-time, as they consist sensor and actuator nodes that interact with the physical world directly. Embedded software in the computational nodes is responsible for reacting to sensor data and produce actuation within time constraints. This report discusses run-time computation models for this kind of real-time embedded software, from both the data-flow and the control-flow perspectives. In general, data-flow centric models are natural for describing measurement and control algorithms and are easy to be used in distributed systems, but they lack mechanisms to control the execution order of components to fulfill time constrains. Control-flow centric models are good at handling reactiveness and real-time issues but are hard to distribute. Most useful run-time models have to support both data-flow and control-flow, and make trade-offs when necessary.
In this report, we discuss how this is done in Ptolemy II, a software framework for embedded software. Four models, the synchronous dataflow (SDF) model, , the finite state machine (FSM) model, the real-time process (RTP) model, and the time-synced discrete event (TSDE) model are studied with respect to the data-flow and the control-flow perspectives. Ptolemy II, which is usually used as a multi-model design environment, is extended to support various run-time models. A goal of this extension is to study the properties of run-time models and their composition. Several measurement and control systems are prototyped to illustrate the run-time support in Ptolemy II.