Multi-Objective Physics-Guided Recurrent Neural Networks for Identifying Non-Autonomous Dynamical Systems*

Schön, Oliver; Götte, Ricarda-Samantha; Timmermann, Julia

doi:10.1016/j.ifacol.2022.07.282

Cited by 6 publications

(4 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In the experiments, we compare to GRUs that receive the simulator as additional control input (Schön, Götte, and Timmermann 2022;Jia et al 2021). Under certain conditions, GRUs and other RNNs are a contraction (Bonassi, Farina, and Scattolini 2021;Miller and Hardt 2019).…”

Section: Distinction To Other Architecturesmentioning

confidence: 99%

“…Furthermore, none of these works informs the latent states of the learning-based component. Another possibility is to provide the simulator as control input to an RNN (Schön, Götte, and Timmermann 2022;Jia et al 2021). As discussed in detail in Sec.…”

Section: Related Workmentioning

confidence: 99%

“…In the experiments, we show that our model produces accurate short and long-term predictions and is on par or better than baselines. In particular, it also outperforms a recurrent architecture that receives the simulator as control input (Schön, Götte, and Timmermann 2022;Jia et al 2021), especially when the simulator is only partially informative. We also derive the differences and advantages with respect to this architecture intuitively and mathematically and support the empirical findings.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Learning Hybrid Dynamics Models with Simulator-Informed Latent States

Ensinger,

Ziesche,

Trimpe

2024

AAAI

View full text Add to dashboard Cite

Dynamics model learning deals with the task of inferring unknown dynamics from measurement data and predicting the future behavior of the system. A typical approach to address this problem is to train recurrent models. However, predictions with these models are often not physically meaningful. Further, they suffer from deteriorated behavior over time due to accumulating errors. Often, simulators building on first principles are available being physically meaningful by design. However, modeling simplifications typically cause inaccuracies in these models. Consequently, hybrid modeling is an emerging trend that aims to combine the best of both worlds. In this paper, we propose a new approach to hybrid modeling, where we inform the latent states of a learned model via a black-box simulator. This allows to control the predictions via the simulator preventing them from accumulating errors. This is especially challenging since, in contrast to previous approaches, access to the simulator's latent states is not available. We tackle the task by leveraging observers, a well-known concept from control theory, inferring unknown latent states from observations and dynamics over time. In our learning-based setting, we jointly learn the dynamics and an observer that infers the latent states via the simulator. Thus, the simulator constantly corrects the latent states, compensating for modeling mismatch caused by learning. To maintain flexibility, we train an RNN-based residuum for the latent states that cannot be informed by the simulator.

show abstract

Section: Distinction To Other Architecturesmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Learning Hybrid Dynamics Models with Simulator-Informed Latent States

Ensinger,

Ziesche,

Trimpe

2024

AAAI

View full text Add to dashboard Cite

show abstract

“…) with g as gravity constant and the same parameters p = (m, a, d, J, r, µ) T utilized as in Götte and Timmermann (2022); Schön et al (2022). Here, the joint model ( 2) is the model that completely lacks of the term M F .…”

Section: Golf Robotmentioning

confidence: 99%

Estimating States and Model Uncertainties Jointly by a Sparsity Promoting UKF

Götte¹,

Timmermann²

2022

Preprint

Self Cite

View full text Add to dashboard Cite

State estimation when only a partial model of a considered system is available remains a major challenge in many engineering fields. This work proposes a joint, square-root unscented Kalman filter to estimate states and model uncertainties simultaneously by linear combinations of physics-motivated library functions. Using a sparsity promoting approach, a selection of those linear combinations is chosen and thus an interpretable model can be extracted. Results indicate a small estimation error compared to a traditional square-root unscented Kalman filter and exhibit the enhancement of physically meaningful models.

show abstract

Very high cycle fatigue life prediction of Ti60 alloy based on machine learning with data enhancement

Qian

Huang

et al. 2023

Engineering Fracture Mechanics

View full text Add to dashboard Cite

Multi-Objective Physics-Guided Recurrent Neural Networks for Identifying Non-Autonomous Dynamical Systems*

Cited by 6 publications

References 6 publications

Learning Hybrid Dynamics Models with Simulator-Informed Latent States

Learning Hybrid Dynamics Models with Simulator-Informed Latent States

Estimating States and Model Uncertainties Jointly by a Sparsity Promoting UKF

Very high cycle fatigue life prediction of Ti60 alloy based on machine learning with data enhancement

Contact Info

Product

Resources

About