Text
Team: CAMEA
Challenge: Power-aware radar configuration
Context
- Traffic monitoring system – a complex mix of various sensors and components
- UC includes ITS systems that are mostly video-based or radar-based
- Applications as travel time estimation or vehicle detection and classification
- Sensors are mostly standalone and should operate with a limited power supply.
- The radar sensor is placed in a hermetically sealed box (lacks active cooling)
- Low power consumption + reduced chip heat dissipation is the key feature.
Goals
- Targeting to low-power requirements using and AI guided configuration and setup
- Allow battery supply (or solar power) and reduce heat dissipation.
External link, opens in new window.
Team: ABI, UNISS, INTECS
Challenge: Formal verification of Neural Networks
Context
- Automotive domain
- The combination of AI and ML can enhance the entire development of safety-critical systems and support the prediction of new scenarios that might be considered as safety critical.
Goals
- Identify a plausible network architecture presenting satisfactory accuracy and robustness when applied to the case study of interest and possible to verify leveraging the current state-of-the-art verification tools
Read the LinkedIn newsletter about this challenge External link, opens in new window.
Team: TEK, UNIVAQ
Challenge: Design choice exploration/verification
Context
- Agile process and Electric/Electronic Architecture of a vehicle for professional applications.
- The use case deals with diagnostics and prognostics of power electronics
Goals
- Design choices verification
- Before the implementation; on models and/or with rapid prototyping; environment: simulated/emulated
- Verification of the models: do they cover the requirements? do they "work"?
- Verify, at design time, the adequacy (the functional aspects, as well as the response versus the resources) of the target components that the system architect has in mind to map/realize the design of the system.
Read the LinkedIn newsletter about this challenge External link, opens in new window.
Team: TEK, AST
Challenge: Runtime Verification
Context
- Agile process and Electric/Electronic Architecture of a vehicle for professional applications.
- The use case deals with diagnostics and prognostics of power electronics
Goals
- A run time verification on the implemented module/component/sub-system/system Unit test.
External link, opens in new window.
Team: TEK, ROTECH
Challenge: Operating life monitoring
Context
The TEKNE srl
External link. case study “Agile process and Electric/Electronic Architecture of a vehicle for professional applications” deals with diagnostics and prognostics of power electronics.
- AI-based solutions whose functionalities are monitoring CPS (the software system as well as the physical system on which the former acts)
- interpreting the data so produced (diagnostics and prognostics of the systems), and that can be deployed in the Cloud.
Goals
- Analysis and design of the data cleaning module.
- Analysis of the board constraints for the deployment.
External link, opens in new window.
Team: VCE, MDU, JKU, UNIVAQ, SOFT, IMTA
Challenge: Modeling patterns for AI enhanced architecture modeling
Context
- Thermal management system modeling with recommender systems
- Various VCE artefacts as input
- Elaboration of previous hackathon challenge
Goals
- Modeling patterns for the VCE context
- DEMO of the proposed solution
External link, opens in new window.
Team: VCE, MDU, JKU, SOFT, DT, IMTA
Challenge: Continuous delivery of SysML models and testing in Simulink
Context
- Thermal management system modeling
- Many components are variants
- SysML and Simulink are the two languages of utilized
- Simulation is required in the process
Goals
- The goal is to instantiate a simulink model from SysML based on requirements
- Further it should be validated via simulation
External link, opens in new window.
Team: AVL,JKU,UCAN,AIT,UNIVAQ
Challenge: Optimization of Development Processes
Context
- Automotive development processes
- Systematic data management (traceability!) will produce massively linked and highly structured data
- We want to mine it in order to optimize development processes
- Problem: Structured data is not yet available
Goals
- We want to artificially create realistic structured data
- We need to emulate vehicle development projects
- Only consider product optimization–i.e., iteratively optimize vehicle parameters until vehicle KPIs meet targets
- Create time series data of vehicle parameters and vehicle KPIs
- Reproduce complex non-monotonic behavior of KPIs and parameters over time
- Ensure scalability to large number of KPIs and parameters
Team: BT/ALSTOM + AVL+ABO
Challenge: Automatic parametrization of PS controller
Context
- Parametrizing control models of the propulsion system is done manually during the physical system testing
- Alstom’s control/test engineers need to automatically parametrize control models of the propulsion system
Goals
- Automate the parametrization process using AI/ML models
- Minimize time and effort spent on parametrizing control models
- Extend the parametrization process to new propulsion systems using AI/ML
External link, opens in new window.
Team: WMO + RISE + COPADO
Challenge:AI-Enhanced Test Results Exploration of Nightly Test Results Data
Context
- Westermo develops embedded systems in
an agile process with nightly testing - Nightly test results are stored to database
- This challenge also involves code changes
in an abstract format from the software
under test and from the test framework.
Goals
- When having this data, what can we see
in and learn from it, e.g.
- Do some tests pass and fail together?
- How are testing becoming flaky, and for how long are they flaky?
- Do test cases change verdicts when code changes appear?
- How could one present test results in a suitable way?
External link, opens in new window.
Team: CSY
Challenge: PO Classification Context
- Automatic proving can solve 85% of artifacts produced by B development.
- Several tools can be used and finding the good one is time consuming
- Classification and Machine Learning can be used to avoid try and fail
- POG (proof obligations in XML form) are available in a GitHub repository
Goals
- Classify POG to simplify automatic proving
- Prepare field for full AI proving
Team: CSY
Challenge: Game of Proof
Context- After automatic proving, PO have to be solved manually.
- Manual proof can be compared to winning a game.
- We have banks of manual proofs to work on.
- The way to go is not clear
- Determine a way to train a neural network with proofs created during manual proving.
- Organize data
- Create pipelines in current software
Team: HIB+AND
Challenge: CI/CD for restaurant management level 2
Context- Integration of AND’s tools into HIB’s TAMUS development environment was started in the first hackathon but never completed (signing an NDA is no easy task). But now that is solved so we jump into full automation for CD and CI
- Analyze current state of practice for TAMUS development wrt CD and CI.
- Identify tools in AND’s portfolio to help solve low hanging fruit for now.
Team: PRO + ACORDE + ITI + UOC
Challenge: Anomaly Detection for smart platform
Context
- The Smart Port monitoring platform is responsible for collecting data from many different sensors. Currently, the quality of the data leaves much to be desired, not only because of problems in receiving the data (frequency of sending and quality of the data received), but also because of problems in the infrastructure
Goals
- Detect problems as soon as possible in order to try to prevent them from occurring, by using anomaly detection techniques.
- Simulation of the platform needs.
- Monitor data quality of IoT sensors and platform.
%20European_Commission%20logo.png)
%20Picture2.png)