Full Automatic Path Planning of Cooperating Robots in Industrial Applications

Kurzfassung

Carbon fiber reinforced plastic (CFRP) parts for airplane components can be so huge that a single industrial robot is no longer able to handle them, and cooperating robots are required. Manual programming of cooperating robots is difficult, but with large numbers of different sized and shaped cut-pieces, it is almost impossible. This paper presents an automated production system consisting of a camera for the precise detection of the position of each cut-piece and a collision-free path planner which can dynamically react to different positions for the transfer motions. The path is planned for multiple robots adhering to motion constrains, such as the requirement that the textile cut-piece must form a catenary which can change during transport. Different existing path planning algorithms are evaluated and compared. Additionally a technique based on machine learning has been implemented which correctly resolves redundancy for a linear axis during planning. Finally, all components are tested on a real robot system in industrial scale.