Research project: "Livescan" -

Live Interactive Visual Evaluation System for Component Analysis and Assembly Navigation

Research area: Human Action Recognition, Progress Monitoring, Artificial Intelligence
Funded by: IFB Hamburg as part of the "Green Aviation Technologies II (GATE II)" program
In collaboration with: Synergeticon GmbH,
 

Lufthansa Technik AG

Start of project: January 2025
End of project: December 2026

Description:

LIVESCAN – Visual Progress Detection and Human Action Recognition to Support Aviation Maintenance Processes

The LIVESCAN project is developing AI-supported assistance systems for manual maintenance in aviation. The goal is to enable digital feedback on the current work status through visual progress detection and automatic analysis of human actions (human action recognition), thereby increasing transparency, safety, and process quality in MRO (Maintenance, Repair, and Overhaul) processes.

This sub-project of the IFPT focuses on the research and development of non-invasive systems for image-based progress detection. The focus is on analyzing camera data using artificial intelligence methods to automatically detect and classify activities such as disassembly, cleaning, inspection, and assembly. The unique feature lies in its non-invasive implementation: The systems used do not require any special sensors on the human or the component and do not interfere with existing processes.

A central research goal is the robust, context-sensitive recognition of typical work steps based on visual data in complex, partially unstructured maintenance environments. The developed methods should enable the automated evaluation of assembly progress and the integration of the information into digital process representations. Furthermore, this data can be used to provide targeted support for digital assistance systems, e.g., in the form of context-dependent documentation or step-by-step instructions.

The IFPT contributes its comprehensive expertise in the fields of process monitoring, human-machine interaction, and AI-supported data processing. This builds on existing research results from related projects (including ADAPT, HyPLANT100, Hi-Digit Pro 4.0). The developed methods will be designed to be compatible with the project partners' systems and integrated into the overarching goal: the establishment of an intelligent, adaptive assistance system for the sustainable improvement of quality and efficiency in aviation maintenance.

 

Contact at the Institute: M.Sc. Ehssan Roshankar