Text

Challenge 1

Real Driver Emission

UNIVAQ is the offering this challenge!

A data-driven model is needed to be applicable to simulate human-like driving on any arbitrary test route. UNIVAQ proposes to define a solution that realizes partially the “Engagement & Analysis” component. During the Hackathon, the team will analyze the AVL_RDE dataset and develop a prototype of the prediction model (the solution component can be implemented by employing linear regression algorithm - supervised ML)








 

Challenge 2

500 nights of testing

WESTERMO is the team offering this challenge!

WESTERMO is developing robust date communication systems for industrial applications. The software is developed every day and tested every night. The tests are conducted on a number of heterogeneous test systems with different hardware setups. Test results from nightly testing have been collected in a test results database.

In the beginning, they had four test systems and developed software in two software branches. Now, they have several dozen test systems and some 50+ branches in use each quarter. They need to prioritize test cases and only have time to run about a third of the tests each night.

Western will provide the data for 500 consecutive night test results from the early days of the test results database

 

 


Challenge 3

Modeling properties constraints of the Abinsula Use Case

ABI, UNISS and UNIVAQ is the team offering this challenge!

This challenge is related to the ABI usecase: “Safety-critical systems in the automotive domain using disruption technology”.

Modern cars are connected systems and acquire inputs from the environment.

In this Case Study, the sensors are going to be the video that can be used instead of the rear-view mirror. With this new source of information, new challenges in the development process are arising.

The main goal of the Abinsula Case Study is to propose an approach, based on AIDOaRt methods, and apply it in an interesting case as the virtual mirrors in cars. The technological goals of the Case Study are related to the introduction of Artificial Intelligence and Machine Learning techniques in the modeling and testing phase of the system development life cycle.

Challenge 4

Architecture modeling pattern

VCE and Solution providers is the team offering this challenge!

At VCE, the product life cycle is integrated with architecture-based methodologies, related tools and information systems.

An “interface diagram” (IFD) represents a physical view as well a “logical architecture” of not only a single machine but also the corresponding product line.

At VCE, the IFD views are based on modular architecture methodologies to support product line engineering for enhancing modularity for design reuse for instance.

The goal of this task is to develop a modeling template to model system logical architecture (IFD views), using SysML language and corresponding tools.

An additional challenge is to be able to ‘generate’ a traceability matrix, called “sharing matrix” at VCE, showing shared variable system elements in different contexts.

Challenge 5

Big data Monitoring solution

PRO and ACO are the teams offering this challenge!

The Smart Port platform is in charge of monitoring the activities of a port in real-time, through the analysis of data coming from sensors (IoT) and information systems (legacy and external systems). Monitoring the platform is another aspect to control, also monitoring the flow of data and ensuring that the information is processed correctly without loss of data is another desired improvement.

The challenge is aiming at improving the current monitoring system of the platform that is based on ElasticSearch and Kibana, changing the current system by Prometheus and Grafana monitoring more parameters and generating alarms in an automated way.

The result of this challenge would enable the team to detect anomalies and predict them during the next iterations of the Use Case.

 

Challenge 6

Design choices exploration/verification

TEK and UNIVAQ is the team offering this challenge!

This challenge is in the context of the Use case “Agile process and Electric/Electronic Architecture of a vehicle for professional applications'' and the use case scenario “Design choices verification”. (Before the implementation; on models and/or with rapid prototyping; environment: simulated/emulated.)

This challenge aims is to verify at design time, with respect to the requirements, the adequacy (the functional aspects, as well as the response versus the resources) of the architecture and of the real components (hardware and software, make or buy or reuse) on-which/with-which the system architect has in mind to map/realize the design of the system.

This challenge would enable the team to verify in a semi-automatic manner, the coverage of the architectural models with respect to the requirements, as well as if the models “work”.

Challenge 7

Operating Life Monitoring

TEK and ROTECH is the team offering this challenge!

This challenge is in the context of the Use Case: “Agile process and Electric/Electronic Architecture of a vehicle for professional applications'' and the Use case scenario: “Operating life monitoring”. (On the product that carries out its services; environment: operating, partly simulated.)

The goal of this challenge is to find a component for communication between the On-board computing platform and the Remote Computing and Data Storage with data handling and encryption/decryption for secure communication.

 

 

 

 

 

 

Challenge 8

Improving CI/CD in Restaurants

HI Iberia and Anders Innovations Oy is the team offering this challenge!

This challenge is in the context of the Use Case 6 Restaurants. It is based on the TAMUS product by HI Iberia which has been developed in an informal DevOps manner in the past years. This approach is flexible but needs considerable human intervention in all phases of the cycle.

Two of the requirements for the UC6 are mapped to CI/CD solutions. Both necessitate the installation of a proper CI/CD solution and the undertaking of an initial proof-of-concept experiment on the use case testbed.

 

 

 

 

 

 

Challenge 9

Automate finding test values

Automated Software Testing and Universitaet Linz is the team offering this challenge!

AST’s tool suite DevMate provides automated generation of test cases through equivalence classes from parsed code or models. Test data currently needs to be provided by a user. Determination of this data through Artificial Intelligence and Machine Learning approaches is the next step in test automation.

Aalpy by TUG is an extensible active automata learning library which can build a model from black box systems and determine its states.

During this challenge, it will be attempted to create a small demo of Aalpy working with DevMate. The demo will try to generate test cases for a known code project with an existing test model.

 

Challenge 10

Requirements analysis – Recommendation system for RE

Alstom, Mälardalen University, and solution providers are the teams offering this challenge!

Alstom’s tender engineers need to evaluate and answer customer requirements in a short time. The analysis of the requirements is extensive and time-consuming. These requirements are stored in DOORS Rational Software, and they are going through several phases of requirements analysis.

Alstom’s goal is to minimize time spent on analyzing the requirements and maximize the number of correctly evaluated and answered requirements.

That is why the goal of this challenge is to investigate whether finding semantic similarities between new requirements and requirements from previous projects could help us create an automated solution that would enhance the engineers’ analysis.

Challenge 11

Proof obligations classifications

CSY is the team offering this challenge!

Proof obligations are side artifacts from proven development that need to be solved to ensure correctness. Several automatic tools exist that can solve proof obligations (PO) but until now no methods or criterion can tell which tool will be efficient on one PO. We need to try tools one by one until one of them solves the PO.

A large bank of PO from several projects is collected , anonymized and made available on a github repository. Also, running scripts are run to label them the old fashion way while benchmarking the operation. The PO are modeled into xml based files.

This challenge aims at training a neural network able to classify PO with the correct tool in less time than what will be saved by avoiding running unnecessary tools.

Challenge 12

Trello interface for AIOps

HIB is the team offering this challenge!

The Restaurants UC by HIB uses as a starting point the TAMUS application by HI Iberia. The development team for this uses Trello to manage requirements and assign development tasks. For now, the process is entirely manual, and the product manager creates tasks and assigns them based on his experience and does so based on his available time for analysis. This leaves room for automation to be implemented.

The goal for this challenge would be to have a first dummy working demo of the Trello analysis. This would entail for now:

-Connecting to the Trello via the API

-Reacting to new cards appearing of the board and sending it to the analysis AI pipeline.

-Perform a (dummy) analysis and make changes on the board automatically.