VAROUMAS Steven

PhD student at Sorbonne University
Team : APR
https://perso.lip6.fr/Steven.Varoumas
https://perso.lip6.fr/Steven.Varoumas

Supervision : Tristan CROLARD

Co-supervision : TREBUCHET Philippe , CHAILLOUX Emmanuel

High-level programming models for microcontrollers with scarce resources

Microcontrollers are programmable integrated circuit embedded in multiple everyday objects. Due to their scarce resources, they often are programmated using low-level languages such as C or assembly languages. These languages don't provide the same abstractions and guarantees than higher-level programming languages, such as OCaml. This thesis offers a set of solutions aimed at extending microcontrollers programming with high-level programming paradigms. These solutions provide multiple abstraction layers which, in particular, enable the development of portable programs, free from the specifics of the hardware. We thus introduce a layer of hardware abstraction through an OCaml virtual machine, that enjoys the multiple benefits of the language, while keeping a low memory footprint. We then extend the OCaml language with a synchronous programming model inspired from the Lustre dataflow language, which offers abstraction over the concurrent aspects of a program. The language is then formally specified and various typing properties are proven. Moreover, the abstractions offered by our work induce portability of some static analyses that can be done over the bytecode of programs. We thus propose such an analysis that consists of estimating the worst case execution time (WCET) of a synchronous program. All the propositions of this thesis form a complete development toolchain, and several practical examples that illustrate the completeness of the given solutions are thus provided.

Defence : 11/05/2019

Jury members :

PAGETTI Claire, ONERA [rapporteur]
SCHMITT Alan, INRIA Rennes [rapporteur]
SIGHIREANU Mihaela, Université Paris Diderot
BOURKE Timothy, INRIA Paris
AGON Carlos, Sorbonne Université
CROLARD Tristan, CNAM
CHAILLOUX Emmanuel, Sorbonne Université
TRÉBUCHET Philippe, Sorbonne Université

Departure date : 12/31/2020

2016-2023 Publications