RUTILEA

CASE STUDY: Manufacturer of automobile

Realize picking of stacked crankshafts using a robot arm

Theme

Solution

Result

Forged materials such as crankshafts are heavy objects, and repeatedly removing these materials places a heavy burden on operators. Accordingly, there is a strong demand for robot arm automation. Removing forged materials, machinery components, and other heavy objects not only requires a great deal of labor, it also poses a risk of accidents or injuries resulting from falls. By achieving stable picking with a robot, we can drastically reduce the physical burden on operators, contributing to improved productivity for the manufacturing line as a whole. The client had a need for robot arm central point picking of stacked crankshafts in order to achieve efficient operations, making the handling of various processes smoother.

Requirements for crankshaft picking

Crankshaft picking has the following two requirements.

  • A robot system that can provide stable component removal operations is required for reducing the burden on operators to improve efficiency and safety.
  • For stacked crankshafts, we want to achieve smooth picking movement, grabbing onto the center axis.

Realize secure picking with our algorithm

Using pictures of multiple stacked crankshafts, we extract working images for single picking targets, carry out authentication and matching with the template, and determine a picking point from the central axis. In addition, we used a transformation matrix to convert the three-dimensional camera coordinates to robot coordinates based on the information from several sampling points, and also created a calibration program. With these materials, we determined the positional relationship between the crankshafts and the robot arm using images from the camera photography, enabling accurate picking operations to be performed.

Successful rapid picking using prescribed movements contributes to manufacturing line efficiency

Picking holding the central axis was performed 30 times on stacked crankshafts with a 99.7% success rate (29 successes). In addition, the calibration program created to convert camera coordinates to robot coordinates successfully converted this data with a 1.3 mm margin of error. As a result, we successfully achieved crankshaft picking using a robot arm and contributed to manufacturing line efficiency.