Modular Code Generation from Hybrid Automata based on Data Dependency

  • Authors:
  • Jesung Kim;Insup Lee

  • Affiliations:
  • -;-

  • Venue:
  • RTAS '03 Proceedings of the The 9th IEEE Real-Time and Embedded Technology and Applications Symposium
  • Year:
  • 2003

Quantified Score

Hi-index 0.00

Visualization

Abstract

Model-based automatic code generation is a processof converting abstract models into concrete implementationsin the form of a program written in a high-levelprogramming language. The process consists of twosteps, first translating the primitives of the model into(approximately) equivalent implementations, and thenscheduling the implementations of primitives accordingto the data dependency inherent in the model. Whenthe model is based on hybrid automata that combinecontinuous dynamics with a finite state machine, thedata dependency must be viewed in two aspects: continuousand discrete. Continuous data dependency ispresent between mathematical equations modeling time-continuousbehavior of the system. On the other hand,discrete data dependency is present between guardedtransitions that instantaneously change the continuousbehavior of the system. While discrete data dependencyhas been studied in the context of code generation frommodeling languages with synchronous semantics (e.g.,ESTEREL), there has been no prior work that addressesboth kinds of dependency in a single framework. In thispaper, we propose a code generation framework for hybridautomata which deals with continuous and discretedata dependency. We also propose techniques for generatingmodular code that retains modularity of the originalmodel. The framework has been implemented basedon the hybrid system modeling language CHARON, andexperimented with Sony's robot platform AIBO.