WireInstruction: AR for precise automotive repair processes

Purpose and goal:

The project WireInstruction aimed to further develop the image identification functionality in Wirevision to also provide interactive instructions for disassembly/assembly through the stages in a work instruction using vision-based machine learning and AI. We have developed a demonstrator app that guides a user by identifying the state of a component, i.e. how far the disassembly/assembly has progressed and at the same time interactively instruct the user at each step. We have hereby verified the functions in Wiretronic´s paten application from 2020.

Expected results and effects:

The result is a demonstrator app of the function described above, as well as verification of an earlier paten application. We have also written an application for design protection for a novel hand-tool design for better visual tool recognition. We have received very positive feedback from selected customers the system was shown to.

Going forward, we will continue the development and concentrate on developing systems that simplify industrialization and large-scale operation for e.g. effectively automating the generation of interactive instructions for WireInstruction. Approach and implementation:

The project has largely been carried out according to plan without any real complications. The response from stakeholders has been very positive. We have identified several interesting paths forward to move towards industrialization and several exciting features such as: Inspection of objects to verify if installations are correct; 5S inspection-i.e. are components and tools where they should be?; Stock inventory etc. etc.

We have also realized how best use the right technology not to make an application too slow or complicated.