| About this Abstract |
| Meeting |
2024 Annual International Solid Freeform Fabrication Symposium (SFF Symp 2024)
|
| Symposium
|
2024 Annual International Solid Freeform Fabrication Symposium (SFF Symp 2024)
|
| Presentation Title |
Design and Additive Manufacturing of Compliant Door-latch Mechanism Based on Reinforcement Learning |
| Author(s) |
Yejun Choi, Keun Park, Yeonung Kim |
| On-Site Speaker (Planned) |
Yejun Choi |
| Abstract Scope |
Recent advancements in additive manufacturing (AM) technology have spurred active research into mechanical metamaterials. In this study, we investigated a micro-lattice-based compliant mechanism capable of altering motion through interaction of the lattice structure constituting the mechanical metamaterial. Finite element analysis (FEA) was conducted on square and parallelogram-based lattice structures to examine their deformation characteristics. The design domain of the compliant mechanism employing the lattice structure, was subjected to FEA, and the analysis results were utilized for reinforcement learning (RL) for optimal design of the mechanism. The dueling deep Q-network (DQN) algorithm was implemented to enable efficient decision-making. Additionally, a reward function was suggested to design a compliant door latch mechanism capable of converting rotational motion into linear motion. The compliant mechanisms designed through the RL with auxiliary guidance of human experience were fabricated using fast-filament fabrication (FFF) type AM, and the relevant deformation behaviors were validated experimentally. |
| Proceedings Inclusion? |
Definite: Post-meeting proceedings |