Model Predictive Control with Variational Autoencoders for Signal Temporal Logic Specifications
Author:
Im Eunji1, Choi Minji1, Cho Kyunghoon1ORCID
Affiliation:
1. Department of Information and Telecommunication Engineering, Incheon National University, Incheon 22012, Republic of Korea
Abstract
This paper presents a control strategy synthesis method for dynamical systems with differential constraints, emphasizing the prioritization of specific rules. Special attention is given to scenarios where not all rules can be simultaneously satisfied to complete a given task, necessitating decisions on the extent to which each rule is satisfied, including which rules must be upheld or disregarded. We propose a learning-based Model Predictive Control (MPC) method designed to address these challenges. Our approach integrates a learning method with a traditional control scheme, enabling the controller to emulate human expert behavior. Rules are represented as Signal Temporal Logic (STL) formulas. A robustness margin, quantifying the degree of rule satisfaction, is learned from expert demonstrations using a Conditional Variational Autoencoder (CVAE). This learned margin is then applied in the MPC process to guide the prioritization or exclusion of rules. In a track driving simulation, our method demonstrates the ability to generate behavior resembling that of human experts and effectively manage rule-based dilemmas.
Reference40 articles.
1. Camacho, E.F., and Alba, C.B. (2013). Model Predictive Control, Springer Science & Business Media. 2. Dantec, E., Naveau, M., Fernbach, P., Villa, N., Saurel, G., Stasse, O., Taix, M., and Mansard, N. (2022, January 28–30). Whole-body model predictive control for biped locomotion on a torque-controlled humanoid robot. Proceedings of the IEEE-RAS International Conference on Humanoid Robots, Ginowan, Japan. 3. Hybrid iLQR model predictive control for contact implicit stabilization on legged robots;Kong;IEEE Trans. Robot.,2023 4. Fast contact-implicit model predictive control;Howell;IEEE Trans. Robot.,2024 5. Jang, E., Irpan, A., Khansari, M., Kappler, D., Ebert, F., Lynch, C., Levine, S., and Finn, C. (2021, January 8–11). Bc-z: Zero-shot task generalization with robotic imitation learning. Proceedings of the Conference on Robot Learning, London, UK.
|
|