|
Data-Driven Disturbance Compensation for DC Motors |
|
Jaeho Lee, Yongsoon Eun(DGIST, Korea) |
| This paper presents an experimental result of applying data driven disturbance observer. A DC motor testbed is used and the experimental results indicate successful disturbance compensation as predicted by the theoretical results |
|
|
Data-driven Fault Detection and Identification of Nonlinear Systems using Weighted Window Extended Dynamic Mode Decomposition |
|
Jayden Dongwoo Lee, Lamsu Kim(KAIST, Korea), Hosun Lee, Hyochoong Bang(Korea Advanced Institute of Science and Technology, Korea) |
| In this study, we designed a data-driven fault detection and identification method for time-varying nonlinear systems using the Koopman operator. Koopman operator is an infinite-dimensional linear operator that transforms a nonlinear dynamical system. In this paper, Weighted Window Extended Dynamic Mode Decomposition (WW-EDMD) is used to obtain the Koopman operator through a recursive procedure to reduce the computation time and memory. Numerical simulation results show that the proposed method has better fault detection ability for time-varying nonlinear systems than window extended dynamic mode decomposition. |
|
|
Reinforcement Learning Approach to Velocity and Position Control of Metro Trains |
|
Kyungbae Lee, Seungyeop Lee, Seunghyeon Kim, Yongsoon Eun(DGIST, Korea) |
| This paper investigates the feasibilty of applying reinforcement learning for position and velocity control of metro trains. |
|
|
Data-Driven Inverse Dynamics of Single-Input Single-Output Linear System |
|
Juwon Lee, Juhoon Back(Kwangwoon University, Korea) |
| In this paper, we propose a data-driven inverse dynamics for SISO LTI systems. We find an input trajectory from the corresponding output trajectory from input and output data obtained by an experiment. To solve the problem, we consider a relative degree of the system related to invertibility of the system matrix D. Simulation results are given to validate a proposed method. |
|
|
Sparse Bayesian Network-based Disturbance Observer for Policy-based Reinforcement Learning |
|
HyeonBeen Park(POSCO Holdings, Korea) |
| We proposed the Sparse Bayesian Network-based Disturbance Observer (SBN-DOB) to enhance the robustness of policy-based reinforcement learning. SBN-DOB utilizes sparse Bayesian learning to estimate the nominal inverse model dynamics, effectively mitigating model uncertainty and disturbances without relying on physical modeling. The SBN-DOB can be compressed by inducing sparsity in network parameters through sparse Bayesian learning, and the Bayesian model reduces the risk of overfitting during inference. SBN-DOB is expected to minimize the simulation-to-reality gap of reinforcement learning by used in embedded systems with limited performance and capacity. |
|
|
Data Dependency of DeePC Performance: Case Study with Metro Trains |
|
Seunghyeon Kim, Yongsoon Eun(DGIST, Korea) |
| This paper investigates the effect of data selection used to form the Hankel matrix for data enabled predictive control. Specifically, a metro train simulator is used to evaluate two Hankel matrices obtained from two different data set. The results illustrate the importance of data characteristics when the method is applied in practice for systems that allow linear approximation. |
|