Remaining time:
The second joint hackathon of the University of Primorska (UP FAMNIT) and the HTW Dresden (Mechlab) will take place from:
15-19.09.2025: University of Primorska (Koper, FAMNIT-Building)
- 15.09: 10:00 – Opening, Introduction and Group assignment
- 15.09: 13:00 – Start programming
- 16.09: Programming
- 17.09: Programming and interim presentation
- 18.09: Programming
- 19.09: Final presentations, selection for Dresden 2026
22-26.09.2025: University of Applied Sciences (Dresden, Germany)
The first week provides an introduction to the topics and the creation of first custom functions. In the second week (Dresden), these functions are further developed and tested in real-life applications on the test area for automated vehicles. Five students have already been selected during the 2024 Hackathon.
Application for participation: informally by email to trautmann@htw-dresden.de
Example function 1 – Live display for an autonomous shuttle (supported by FSD GmbH)
An autonomous shuttle has recently been deployed at the HTWD. This will be used to develop test procedures that can/must be used for safe daily operation. Currently, there is no live display of the position and important operating data. This is necessary so that an assistant can monitor the journey and intervene if necessary. Furthermore, it may be necessary to repeat certain tests. Therefore, a live display based on an open-source map (e.g., Open Street Map) is to be developed that offers interactive options for intervention. The concept is intended to be expandable so that multiple shuttles can be monitored simultaneously. The data exchange of all sources can be done, for example, via MQTT server.
Example function 2 – Person detection and gesture recognition (supported by eVersum)
In an automated vehicle, it is important to know the number and reactions of the occupants. Camera-based methods with AI support are particularly well-suited for this. The aim here is to develop a feasible and expandable concept to obtain as much usable information as possible from the video image.
Example function 3 – KI-based analysis of large data streams (supported by tracetronic GmbH)
The data collected during the test drives in Example 1 (driving motion, environment, weather, etc.) must be analyzed automatically. This data must be used to generate a report for the test engineer with suggestions for possible deviations from the normal operating procedure. Suggestions for the derivation of further test cases should be made.
https://www.tracetronic.com/products/extras/et-agent
Example function 4 – Traffic sign recognition
Reliable traffic sign recognition still poses a major challenge for autonomous vehicles. Despite significant progress, reliable recognition of critical traffic signs (stop sign, yield) is still lacking. This can be clearly seen in Tesla vehicles, which have had stop sign assistance for several years.
Other manufacturers do not yet offer such a function. The hackathon will therefore initially present the existing methods (e.g. Yolo object detection). These will then be expanded for use with other traffic signs (give way). The driver should be informed via an appropriate display. Active braking intervention is also possible on the test area in Dresden. The methodology can then be expanded to include other safety functions such as pedestrian detection. Furthermore, concepts are being developed to share the information with other road users. This requires setting up the appropriate servers that enable secure and tamper-proof data exchange. An initial concept was developed as part of the 4th GDG Hackathon at the University of Primorska. This will serve as the basis for further development.
Application for participation: informally by email to trautmann@htw-dresden.de