To content
Fakultät für Informatik
Correct, Safe, and Secure

Software for the 21st century

The AQUA group works on formal methods in software engineering. The one overarching question that drives our works is:

How can we ensure that a software system meets our requirements in the intended operation environment?

We work in four working directions:

  1.  modeling assumptions on the intended operation environment,
  2. formalizing expectations into testable requirements,
  3. checking requirements on a system under the made assumptions, and
  4. supporting the corresponding tasks with tools, processes, and standards.

Find out more about our research ...

to verify a system, models have to be generated for environment, system, and requirements © Falk Howar​/​TU Dortmund
Research-based Teaching

Safe Autonomous Systems

We frequently base our teaching on our research (e.g., for model learning, symbolic execution, or for deveoping and testing automotive software). This allows students to get hands-on experience with the application of formal methods in the context of software engineering. We develop multiple software systems and frameworks that serve as case studies in our labs and courses:

  • Autonomous Vehicle Lab (AV model vehicles for learning about developing and testing software for autonomous robotic systems)
  • BoardGameWork (a framework for learning application development by developing board games in Kotlin)
  • BoardGame Server (a SpringBoot web application for playing Pachisi and similar games in the browser for learning a.o.  enterprise application patterns)
  • BoardGame Editor (a Cinco meta-model from which an editor for Pachisi boards and rulesets can be generated)
  • Very Insecure Web Application (a PHP web application for learning about injection attacks and pen testing)

Find out more about our teaching ...

Open-Source Research Software

World Class Research Tools for Model Learning and Software Verification

Robust research tools are a cornerstone for scaling formal methods towards industrial applications. The group develops multiple research tools, mainly for dynamic symbolic execution and model learning. To compete, compare, and collaborate with other researchers, we compete with our tools in SV-COMP and co-organize the RERS challenge.

We believe in open science and make our tools available as open source software. We strive to provide reproduction packages and benchmark suites for our results and engange in the national research data infrastructure (NFDI) to develop corresponding standards. 

 

Software Verification

  • JDart (SVCOMP Gold 2022 for Java)
  • GDart (SV-COMP Silver 2023 for Java)
  • SPouT (Amazon Research Award 2021)

Model Learning

  • LearnLib (CAV Artifact Award 2015. We are contributors and co-developers of the library)
  • RaLib

 


News

01/24/2024

Marcel Altendeitering defends Doctoral Dissertation

Marcel successfully defended his dissertation titled "Design Principles for Data Quality Tools".

 

Marcel Altendeitering with Christian Janiesch and Falk Howar after defending his doctoral dissertation
12/15/2023

DFG funding for project StdLearnLib

The StdLearnLib project ensures the continued development and expansion of the de-facto standard library in automata learning, LearnLib.

Architectural Overview of the LearnLib Components.
11/04/2023

Autonomous Racing Showcased on AADC Model Car

At the university's 2023 open house event, AQUA group presented its work on autnomous driving on 1:8 scale model cars. Using the Jarvic racing…

The AADC model car is being prepared for an autonomous race at the TU's 2023 Open House.
07/01/2023

NFDI4Ing seed funding for project ReReSo

Project ReReSo, funded by NFDI4Ing, aims at improving the reusability of research software through standardized data formats and tool interfaces.

Three benchmarks with distinct problems and metadata of different types, together with two tools providing an execution and a configuration API, as well as two configurations for each tool.
03/30/2023

GDart wins SILVER in SV-COMP 2023 for Java

Our dynamic symbolic execution GDart wins SILVER in this year's Software Verification competition for the Java category.

03/02/2023

Malte Mues defends Doctoral Dissertation

Malte successfully defended his dissertation project on the integration of dynamic symbolic execution and multi-color taint analysis.

Malte Mues with Heinrich Müller, Ben Hermann and Falk Howar after defending his doctoral dissertation
01/01/2023

BMBF funding for Project QuARum

The QuARUm project, a collaboration between Constructor University and TU Dortmund University, aims to develop a low-code platform for geoscientists.

QuARUm-based workflow. After sampling and sample preparation, a geochemist can use the QuARUm low-code platform to perform quality assessment of analytical and literature data during data evaluation, leading to a publication.
11/01/2022

AQUA featured in CS@ TU Dortmund promotional video

Our research and teaching on software engineering for autonomous driving is featured in the promotional video of the CS department of TU Dortmund!

10/01/2022

DFG funding for project STING

The STING project researches formal methods for finding exploits in Web applications.

08/11/2022

Fabian Bruckner defends Doctoral Dissertation

Fabian successfully defended his dissertation project in which he developed concepts for enforcing data usage control requirements in DSLs.

 

Fabian Brucker with Ben Hermann and Falk Howar after defending his doctoral dissertation
03/30/2022

JDart wins GOLD in SV-COMP 2022 for Java

Our dynamic symbolic execution JDart wins GOLD in this year's Software Verification competition for the Java category.

01/01/2022

BMBF funding for project KomDatis

The projects will develop a general data escrow concept in a collaboration with Fraunhofer ISST and DKSR.

11/11/2021

ACM Distinguished Paper Award for ASE 2021

Paper "Data-driven design and evaluation of SMT meta-solving strategies: balancing performance, accuracy, and cost" receives Distinguished Paper Award

08/01/2021

Falk Howar receives Amazon Research Award

The funded project aims at scaling symbolic execution for the Java VirtualMachine.

03/30/2021

JDart wins SILVER in SV-COMP 2021 for Java

Our dynamic symbolic execution JDart wins SILVER in this year's Software Verification competition for the Java category.

01/01/2020

PG 625 "Development of Automated Tools for Solving CTFs" receives P5 Award

PG 625 "Development of Automated Tools for Solving CTFs", organized by the AQUA group at TU Dortmund University, wins this year's P5 Competition