Automata Based Interfaces for Control and Scheduling

We propose the use of formal languages of infinite words over the alphabet of task identifiers as an interface between control designs and software implementations. We argue that this approach is more flexible than the classical real-time scheduling framework based on periodic tasks, and allows composition of interfaces by language-theoretic operations. We show that finite automata over infinite words offer analyzable representation and can capture many interesting interface specifications such as exponential stability of switched linear systems. Disciplines Computer Sciences Comments From the 10th International Workshop, HSCC 2007, Pisa, Italy, April 3-5, 2007. This conference paper is available at ScholarlyCommons: http://repository.upenn.edu/cis_papers/503 Automata Based Interfaces for Control and Scheduling Gera Weiss and Rajeev Alur University of Pennsylvania {gera,alur}@seas.upenn.edu Abstract. We propose the use of formal languages of infinite words over the alphabet of task identifiers as an interface between control designs and software implementations. We argue that this approach is more flexible than the classical real-time scheduling framework based on periodic tasks, and allows composition of interfaces by language-theoretic operations. We show that finite automata over infinite words offer analyzable representation and can capture many interesting interface specifications such as exponential stability of switched linear systems. We propose the use of formal languages of infinite words over the alphabet of task identifiers as an interface between control designs and software implementations. We argue that this approach is more flexible than the classical real-time scheduling framework based on periodic tasks, and allows composition of interfaces by language-theoretic operations. We show that finite automata over infinite words offer analyzable representation and can capture many interesting interface specifications such as exponential stability of switched linear systems.