Robotic manipulation of flexible components with high variance and process uncertainties in the production of electric vehicles
The vehicle electrical system is, in a sense, the nervous system of the car, and its importance continues to grow as a result of increasing electrification. It consists of cables with a total length of several kilometers and weighs around 60 kg. Due to the high degree of individualization of today's vehicles, wiring systems are designed to be vehicle-specific, which places high demands on assembly personnel and generates a high workload. Around 80% of errors in the production and assembly of wiring harnesses occur during manual work steps. The manual activity also causes high labor costs, high physical stress and requires a lot of time. However, automation is difficult to implement.
Flexurally loose components such as cables and wire harnesses are of particular importance in the production of electric vehicles. However, due to a lack of robust processes, they cannot be handled in an automated way so far. Therefore, novel AI methods for robot-assisted manipulation of flexurally slack components are explored in RoboCable.
For this purpose, a combination of the following technologies is considered:
Through this combination, the robot learns new assembly tasks virtually in the simulation environment to account for the high variances of flexurally slack components and inherent process uncertainties. Existing tasks can be performed in parallel without interruption. In the sense of a VISION ZERO, programming, downtime and start-up times as well as defects are minimized.