Visible to the public Biblio

Filters: Author is Alvarez, Mario M.  [Clear All Filters]
2017-03-20
Malecha, Gregory, Ricketts, Daniel, Alvarez, Mario M., Lerner, Sorin.  2016.  Towards foundational verification of cyber-physical systems. :1–5.

The safety-critical aspects of cyber-physical systems motivate the need for rigorous analysis of these systems. In the literature this work is often done using idealized models of systems where the analysis can be carried out using high-level reasoning techniques such as Lyapunov functions and model checking. In this paper we present VERIDRONE, a foundational framework for reasoning about cyber-physical systems at all levels from high-level models to C code that implements the system. VERIDRONE is a library within the Coq proof assistant enabling us to build on its foundational implementation, its interactive development environments, and its wealth of libraries capturing interesting theories ranging from real numbers and differential equations to verified compilers and floating point numbers. These features make proof assistants in general, and Coq in particular, a powerful platform for unifying foundational results about safety-critical systems and ensuring interesting properties at all levels of the stack.

Malecha, Gregory, Ricketts, Daniel, Alvarez, Mario M., Lerner, Sorin.  2016.  Towards foundational verification of cyber-physical systems. :1–5.

The safety-critical aspects of cyber-physical systems motivate the need for rigorous analysis of these systems. In the literature this work is often done using idealized models of systems where the analysis can be carried out using high-level reasoning techniques such as Lyapunov functions and model checking. In this paper we present VERIDRONE, a foundational framework for reasoning about cyber-physical systems at all levels from high-level models to C code that implements the system. VERIDRONE is a library within the Coq proof assistant enabling us to build on its foundational implementation, its interactive development environments, and its wealth of libraries capturing interesting theories ranging from real numbers and differential equations to verified compilers and floating point numbers. These features make proof assistants in general, and Coq in particular, a powerful platform for unifying foundational results about safety-critical systems and ensuring interesting properties at all levels of the stack.