Abstract:
The project achieved the automation of the rolling process for the application of automated layup, specifically for the rolling of the pre-preg layer. However, it is important to note that the actual laying of the layers and stacking was still done manually (although the ultimate objective of the project is to fully automate the layup process with three collaborative robots). The project includes the design and construction of a rolling-end effector, which was not the primary objective of the project.
To analyze the feasibility and efficiency of the automated layup application, the project extensively conducted software simulations using ROS, MoveIt, and RoboDK. Various path planning algorithms, including A*, RRT, and Dijkstra, were tested, and valuable insights were gained from the simulation results. These findings enhanced the project's understanding of the capabilities of the algorithms and contributed to improved robotic control throughout the design process. The project compared the performance of the algorithms based on planning time, collision avoidance, and overall trajectory optimization.
Furthermore, the project emphasized the integration of path planning algorithms into real-world hardware, specifically the UR5 manipulators. This integration process involved evaluating the robots, optimizing the algorithms, and ensuring smooth communication between the robot controller and the system software. Rigorous testing and iterations were conducted during this phase to address any discrepancies between the simulations and the execution in the real-world environment.
