Stellantis has fixed a bolt-tightening problem on an engine assembly line at its Trenton, Michigan complex by deploying vision-guided robot software from Inbolt, called the Inbolt Platform, according to a report from Assembly. A six-axis robot on the line was missing 2.55 percent of the bolts it needed to tighten, because a lift-and-locate system could not position each moving engine consistently enough for the robot to find every fastening point. Since installing Inbolt's system, the miss rate has fallen to 0.24 percent.
The platform pairs a 3D camera mounted directly on the robot with an AI model pretrained and fine-tuned on CAD data, letting it recognize part geometry without a custom vision setup for each station. It reads a part's position in real time, including shifts in X, Y, Z, yaw, pitch and roll, and adjusts the robot's path on the fly, even while the engine keeps moving on the line. Stellantis installed it without changing the plant's existing infrastructure.
For a line that has assembled V6 engines since the 1950s, cutting the miss rate by a factor of ten changes the economics of the station. Part rejections overall fell 90.5 percent, according to Assembly, and rework has nearly disappeared. A worker is still posted at the end of the line to catch anything missed, but does so far less often now. Stellantis recouped its investment in 12 months.
Stellantis has since applied the same system elsewhere. At its Detroit Assembly Complex-Jefferson plant, Inbolt's guidance corrected an adhesive dispensing operation where misalignments as small as 4 millimeters were forcing manual cleanup, cutting unplanned downtime 70 percent and paying back its cost in six months. At a factory in Szentgotthárd, Hungary, the system locates cylinder heads that arrive stacked unevenly on pallets delivered by a mobile robot with an inconsistent stopping position. In Valenciennes, France, Stellantis uses it to guide cobots mounted on mobile carts that shift position slightly each time they're relocated between machines.
