Path Planning of Cooperating Industrial Robots Using Evolutionary Algorithms

Abstract

Nowadays, aerospace components made of carbon fiber reinforced plastics (CFRP) which are manufactured by a dry-fiber depositing process are produced almost exclusively by hand. In this type of process, hundreds of textile blanks with different contours are stored in a tool mould. For large components such as an aircraft fuselage, it is also necessary to transport the flexible blanks with cooperating robots. For this purpose, a system was developed which allows to generate robot paths on the basis of Computer Aided Design (CAD) data, which is automatically extended by information such as pickup and drop points for the robot. The system has already been successfully tested with sampling-based algorithms on a robot system with two KUKA KR210 R3100 robots on a common linear axis. However, the use of sampling-based algorithms a heuristic must be developed to solve the redundancy of linear axis which is usually only functional for the current scenario and must be adapted for another one. Evolutionary algorithms offer the advantage that they are also able to find a solution for the linear axis independent of the scenario. In this work. A path planning system that uses evolutionary algorithms to calculate collision-free paths reliably for cooperating robots on a common linear axis will be presented.