Choudhary, Pranav Praveen (2026) Feature Engineering to Optimize the 6-DOF Trajectories for Robotic Additive Manufacturing. Masters thesis, Universität Rostock.
Full text not available from this repository.Abstract
In recent times, innovative approaches in additive manufacturing and robotics, two of the main elements of Industry 4.0, are transforming the modern production chain by fabricating complex, highly adaptable, and lightweight components on a large scale. The concept of “Print what you think” has become a reality by employing a robot with multiple degrees of freedom (DOF). In particular, robotassisted Screw Extrusion Additive Manufacturing (SEAM) offers significant advantages for industrial applications, providing scalability and efficiency. However, the trajectory of the robotic SEAM while printing complex geometries is driven by inaccuracies that are dependent on multiple factors. This thesis presents a data-driven approach to overcome the trajectory deviation of the 6-axis robot used for SEAM. Experiments with critical geometries are carried out on a SpaceA production cell installed at DLR’s Future Lab for Additive Manufacturing and Engineering in ARENA2036, Stuttgart, for validating the intended trajectory optimization research. The proposed research focuses on the analysis of different production phase data collected, and consists of a systematic feature engineering framework that details the Robotic AM domain. These features are used to represent the robot controller behavior in relation to the artifacts for the ML models that can be employed for trajectory optimization. Subsequently, XGBoost, LightGBM, and LSTM are trained on the recorded as well as the planned trajectory of the robot to determine the deviation with respect to the path and velocity. The factors most associated with the deviation are identified and analyzed in both planned and actual trajectories and the models are validated as well. Ultimately, this research aims to elevate the robotic additive manufacturing processes for the production of intricate parts and establish the roadmap for future data-driven solutions.
| Item Type: | Thesis (Masters) |
|---|---|
| Subjects: | Autorenart > Studentische Arbeiten > Masterarbeit Autorenart > Studentische Arbeiten |
| Depositing User: | Dbis Admin |
| Date Deposited: | 08 Sep 2026 08:27 |
| Last Modified: | 08 Sep 2026 08:27 |
| URI: | https://eprints.dbis.informatik.uni-rostock.de/id/eprint/1159 |
Actions (login required)
 |
View Item |
