Bachelor’s/Master’s Thesis Trajectory Planning with a Scoring Function for Robot-Based Prepreg Draping

Trajectory Planning with a Scoring Function for Robot-Based Prepreg Draping
Contact Name:  Kevin Heinrich Email: kevin.heinrich@ifu.rwth-aachen.de Type of Thesis: Bachelor’s & Master’s thesis

The robot-based draping of carbon-fiber prepregs is a complex process in the manufacturing of fiber-reinforced polymer components and requires careful planning of robot trajectories.

While sequence planning defines the processing order, it must be translated into a concrete, robot-executable trajectory that ensures wrinkle-free layup while simultaneously minimizing process time. As part of the IGF research project IntelliDrape, this challenge is to be addressed through intelligent trajectory planning using multi-criteria optimization.

The objective of this thesis is to develop and implement an intelligent scoring function that generates optimal robot trajectories from geometry-based motion primitives. A particular challenge is to integrate not only technical optimization criteria, but also human expertise derived from interviews and observational studies with experienced specialists into the evaluation and optimization strategy.

Various trade-offs must be considered, including minimum cycle time, optimal end-effector orientation relative to the fiber direction, smooth force progression, and the avoidance of singularities. A key aspect is the application of conformal mapping methods to project complex 3D surfaces onto a 2D plane, compute optimal paths there, and subsequently map them back onto the original surface.

The practical applicability of the approach will be demonstrated through simulation in RViz/MoveIt and experimental validation.

Possible research questions / directions:

• Which criteria should be included in the scoring function, and with what weighting?
• How can motion primitives from sequence planning be transformed into smooth, continuous robot trajectories using conformal mapping?
• How can trajectories be parameterized to be robust with respect to geometric variations while also taking the robot’s inverse kinematics into account?
• How do different optimization methods (gradient-based, evolutionary, sampling-based) compare in terms of solution quality and computation time?
• How can the optimized trajectories be visualized in MoveIt/RViz and prepared for real robotic systems?

Requirements:

• Degree program in CES, Mechanical Engineering, Industrial Engineering, or a related field.
• Interest in robotics, algorithm development, and AI methods.
• Programming skills in Python.
• Initial experience with ROS is an advantage.

What we offer:

• An exciting thesis/project in an innovative field of research.
• Close and regular supervision in German or English.
• An excellent working atmosphere in a highly motivated team, located in a unique backyard office setting with an office dog.