The Customer’s Challenge
FSD develops innovative testing technologies for vehicle inspections, vehicle diagnostics and safety inspections. Control units in vehicles issue error messages. The associated error descriptions and designations vary from manufacturer to manufacturer. The ZZ Tool (status allocation tool) was a tool that FSD already had in use to allocate between manufacturer fault descriptions and FSD’s generic fault descriptions. This allocation process was previously done completely manually.
The objective is to support the allocation through the use of artificial intelligence in the future. There was already a collaboration agreement with a third-party contractor for this purpose. To enable the work with artificial intelligence, FSD needed a data integration service for provisioning and processing in the context of its existing tool set. Without that interface, it would not be possible to work with and evaluate the AI suggestions in practice.
The Avantgarde Labs solution
Avantgarde Labs was engaged during the current project to develop an interface that could import the available AI suggestions into the ZZ Tool and make them usable. Modifications in the processing and frontend of the existing ZZ Tool were implemented so that the AI suggestions could be used directly by the FSD staff. AI suggestions displayed can be selected and/or allocated to the respective fault descriptions.
Furthermore, it is also possible to export data in the correct format at the push of a button to train the AI solution and thereby iteratively improve the AI solution. In addition, improvements to the UX were made in the existing tool, such as the ability to add filters, short-cut functions and also display improvements, to simplify the FSD staff’s work.
Project results & customer benefits
The development of the interface allowed the AI suggestions to be utilized and the previous manual allocation process to be semi-automated. This function greatly minimizes the amount of searching required between manufacturer fault descriptions and suitable FSD generic fault descriptions. It was only as a result of the fast deployment of the interface that FSD was able to use and also evaluate the AI suggestions of the third-party contractor. Feedback concerning the existing AI suggestions leads to an ongoing adaptation and improvement of the AI suggestions. After completion of the project, Avantgarde Labs was engaged to develop a new ZZ Tool because the existing tool had grown historically and needed to be given a modern architecture through redevelopment. See these projects as well: OTX Ninja at FSD for automated creation of test procedures ODX Ninja at FSD – standardization of the entire data pool for vehicle diagnostics data
"Together with Avantgarde Labs, we developed two phases for the redesign of the ZZ Tool. The first phase was a small modification of the existing ZZ Tool, which was implemented quickly and easily. Currently, a complete redevelopment of the ZZ tool is being carried out and implemented by Avantgarde Labs. We exchange information regularly and are collaborating to develop a direct implementation."
Gunar Hubrich
Product owner for testing in the context of "Investigation of components and systems"
