Swedish industry is today extremely dependent on embedded systems and software. As they are often embedded in products with high demands on safety or availability, it is of utmost importance that their functionality is ensured. Today, verification and validation (V & V) consumes a significant part of the budget for software development and maintenance. Static program analysis is a V & V technique on the rise that can help pinpointing problems early in the software development, and thus reduce V & V costs and time-to-market. Static analysis can give strong guarantees that the software does not have certain "bad" properties that can make it malfunction or crash. This makes it highly desirable to apply in the development of embedded software. The technique is however currently not very much used in Swedish industry. Also, the precision is often insufficient and the m
ethods sometimes do not scale to large software systems. In this project we will target static program analysis for complex embedded systems, such as large embedded systems for telecom or automation. Our goal is to develop methods that scale to large embedded software systems while still being precise enough to be of practical interest. The methods must also handle the heterogeneity that these systems often expose. The implementations will be made freely available as open source, for the benefit of both industry and academia. Close collaboration with com
panies will ensure that the results of the project are properly disseminated within Swedish industry, thus strengthening its competitiveness through more efficient development of embedded systems and software.
Automatic Inference of Task Parallelism in Task-graph-based Actor Models (Dec 2018) Abu Naser Masud, Björn Lisper, Federico Ciccozzi Journal of IEEE Access (Access'18)
Static Backward Program Slicing for Safety Critical Systems (Jun 2015) Husni Khanfar, Björn Lisper, Abu Naser Masud The 20th International Conference on Reliable Software Technologies (Ada-Europe-2015)