A thesis project at Scania is an excellent way of making contacts for your future working life. Many of our current employees started their career with a thesis project.

Background:

Modern vehicles increasingly rely on computers and software to operate properly, and Scania is among the companies aiming to sell fully autonomous vehicles within the next few years. Therefore, the correctness of software at Scania, and in the automotive industry in general, is essential. Exhaustively verifying system correctness can be done using formal verification, which means using computer-assisted mathematical reasoning to prove that a system fulfills some formal specification. However, current methods for formal verification require significant time, resources and knowledge to apply, and have therefore not yet been widely adopted in the automotive industry.

When driving in the physical world, probabilities often arise due to random hardware failures, stochastic communication delays, sensor ghost objects, machine learning components, rounding errors, human behavior, and use of probabilistic algorithms. Therefore, comprehensive vehicle specifications must take probabilities into account. When specifications of vehicle behavior are expressed as probabilistic contracts, it is possible to check that a vehicle fulfills its contract using the following three steps: 

1.    Decompose the top-level vehicle contract into vehicle component contracts.
2.    Verify the correctness of the decomposition.
3.    Verify that each component fulfills its contract. 

In most cases, current techniques for formal verification lack the generality and efficiency to successfully complete the above three steps.

Target:

The goal of this project is to investigate whether statistical methods such as hypothesis testing can be used to check, with some degree of certainty, if an autonomous vehicle fulfills its specification. For instance, given a black-box model M of an autonomous vehicle and a specification S stating that “reindeer on the road are correctly identified 99.9% of the time, with each correct identification resulting in a 95% probability of avoidance”, is it possible to determine whether M fulfills S? Another area of interest is floating point programs, by viewing rounding errors as probabilistic, it may be possible to verify code against specifications using statistical methods. 

Examples of assignment:

-    Modeling of vehicles or vehicle components as stochastic processes
-    Formulating probabilistic specifications for vehicles or components
-    Checking probabilistic specifications using statistical hypothesis testing
-    Specifying rounding errors in software using probabilities
-    Checking C code against formal specifications using hypothesis testing
-    Investigating hypothesis testing for nested probabilities
-    Performing Monte Carlo simulations for the verification of probabilistic specifications

We are also interested in hearing other ideas for directions of the thesis, which can be discussed before the start to better match the applicant’s interest and background.

Education:

M.Sc. in Engineering or similar with an interest in computer science, probability theory, or formal methods.

Contact person and supervisor:
Anton Hampus, PhD student, +46737787449, anton.x.hampus@scania.com
Mattias Nyberg, Adj. prof, KTH / Research Manager, Scania, mattias.nyberg@scania.com

Number of students: 1-3
Time:20 weeks, full time 40 hours per week
Start: Jan 2025
Credits: 30hp

Application:
Enclose CV, personal letter and transcript of grades.
Application shall be registered in both: Thesis project application, and the "Apply"-button on this page

A background check might be conducted for this position. We are conducting interviews continuously and may close the recruitment earlier than the date specified.