A development project at the Giessen -Friedberg University of Applied Sciences is focusing on a quality control method for the automotive industry. A team from the Department of Mathematics, Natural Sciences and Computer Science has created a system for automatically checking the fit of assembled body parts.
The quality standards of automotive manufacturers require, for example, that doors, hoods, and tailgates are fitted precisely. To verify this, the remaining gap between the moving functional part and the supporting frame is measured. The project group at the Giessen -Friedberg University of Applied Sciences has developed a novel solution for this measurement process. An industrial robot is used, whose movement as it scans the gap is controlled by a stereo camera.
Professors Dr. Berthold Franzen, Dr. Klaus Rinn, and Dr. Klaus Wüst contributed their expertise in protocol design, robotics, and image processing to the development. Research assistant Michael Kreutzer served as coordinator. Two graduate students from the computer science program, Jan Hain and Margarita Haubetz, also participated, solving sub-tasks. Exact Vision GmbH, a Bischoffen-based company specializing in measurement technology for the automotive industry, acted as a cooperation partner, investing €50,000 in the project from the HessenAgentur technology funding program.
{mosimage}Two measurements are of primary importance when inspecting a gap in the car body: "Gap" refers to its width, and "Flush" refers to the offset of the planes on both sides of the gap. To determine these values precisely and to ascertain whether they meet the automotive manufacturer's specifications, the curvature of the metal sheets within the recess must be measured. This is accomplished using special dual-axis optical sensors, which, however, only function as intended if they are guided across the gap at the correct angle and distance. The FH team successfully tackled this challenge – automatically guiding a dual-axis optical sensor, optimally aligned, along an ideal path and without contact, across an unknown, arbitrarily curved gap using a robot.
The development focused on designing the system to perform a series of work steps in the smoothest possible sequence. These steps include the photogrammetric pre-measurement of the position and orientation of the unknown gap, the calculation of the most favorable measurement path for the next section and the optimal orientation of the dual light section sensor, the computational prevention of a collision with the vehicle body, the transmission of this data to the robot controller, and the actual measurement for quality control.
The prototype now performs all these sub-functions fully automatically. A first step towards applying this method in automotive production has already been taken. A delegation from Toyota, as guests of the company Exact Vision, visited the FH laboratory and showed great interest in the innovative measurement technology.
