Kinematics-Informed Reinforcement Learning for Trajectory Optimization in CNC Machining
Keywords: Trajectory Optimization, Reinforcement Learning, CNC Machining
TL;DR: Kinematics-Informed Reinforcement Learning (KIRL) optimizes CNC machining by integrating toolpath smoothing and feedrate planning for improved accuracy and efficiency.
Abstract: Toolpath smoothing and feedrate planning are key techniques in Computer Numerical Control (CNC) machining, and play a significant role in machining accuracy, efficiency, and tool life.
Traditional methods typically decouple path smoothing from feedrate planning, without considering the kinematic constraints during the smoothing process.
As a result, the subsequent feedrate planning process is subject to more stringent kinematic limitations, which hinders the achievement of optimal speed execution.
However, the integration of these two processes presents a significant challenge due to severe complexity and nonlinearity of the problem. Here, we propose a novel Reinforcement Learning (RL) based method, termed KIRL, to address the integrated optimization problem.
Experimental results demonstrate that KIRL can generate smoother trajectories and optimize machining time compared to traditional decoupled methods.
To our best knowledge, KIRL is the first RL-based method for solving the integrated toolpath smoothing and feedrate planning optimization problem in CNC machining.
Primary Area: applications to robotics, autonomy, planning
Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.
Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2025/AuthorGuide.
Reciprocal Reviewing: I understand the reciprocal reviewing requirement as described on https://iclr.cc/Conferences/2025/CallForPapers. If none of the authors are registered as a reviewer, it may result in a desk rejection at the discretion of the program chairs. To request an exception, please complete this form at https://forms.gle/Huojr6VjkFxiQsUp6.
Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.
No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.
Submission Number: 3325
Loading