Details
| Originalsprache | Englisch |
|---|---|
| Aufsatznummer | 11609 |
| Seiten (von - bis) | 11609-11616 |
| Seitenumfang | 8 |
| Fachzeitschrift | IEEE Robotics and Automation Letters |
| Jahrgang | 9 |
| Ausgabenummer | 12 |
| Frühes Online-Datum | 11 Nov. 2024 |
| Publikationsstatus | Veröffentlicht - Dez. 2024 |
Abstract
Soft robots pose difficulties in terms of control, requiring novel strategies to effectively manipulate their compliant structures. Model-based approaches face challenges due to the high dimensionality and nonlinearities such as hysteresis effects. In contrast, learning-based approaches provide nonlinear models of different soft robots based only on measured data. In this letter, recurrent neural networks (RNNs) predict the behavior of an articulated soft robot (ASR) with five degrees of freedom (DoF). RNNs based on gated recurrent units (GRUs) are compared to the more commonly used long short-term memory (LSTM) networks and show better accuracy. The recurrence enables the capture of hysteresis effects that are inherent in soft robots due to viscoelasticity or friction but cannot be captured by simple feedforward networks. The data-driven model is used within a nonlinear model predictive control (NMPC), whereby the correct handling of the RNN's hidden states is focused. A training approach is presented that allows measured values to be utilized in each control cycle. This enables accurate predictions of short horizons based on sensor data, which is crucial for closed-loop NMPC. The proposed learning-based NMPC enables trajectory tracking with an average error of 1.2° in experiments with the pneumatic five-DoF ASR.
ASJC Scopus Sachgebiete
- Ingenieurwesen (insg.)
- Steuerungs- und Systemtechnik
- Ingenieurwesen (insg.)
- Biomedizintechnik
- Informatik (insg.)
- Mensch-Maschine-Interaktion
- Ingenieurwesen (insg.)
- Maschinenbau
- Informatik (insg.)
- Maschinelles Sehen und Mustererkennung
- Informatik (insg.)
- Angewandte Informatik
- Mathematik (insg.)
- Steuerung und Optimierung
- Informatik (insg.)
- Artificial intelligence
Zitieren
- Standard
- Harvard
- Apa
- Vancouver
- BibTex
- RIS
in: IEEE Robotics and Automation Letters, Jahrgang 9, Nr. 12, 11609, 12.2024, S. 11609-11616.
Publikation: Beitrag in Fachzeitschrift › Artikel › Forschung › Peer-Review
}
TY - JOUR
T1 - Learning-Based Nonlinear Model Predictive Control of Articulated Soft Robots Using Recurrent Neural Networks
AU - Schäfke, Hendrik
AU - Habich, Tim-Lukas
AU - Muhmann, Christian
AU - Ehlers, Simon
AU - Seel, Thomas
AU - Schappler, Moritz
N1 - Publisher Copyright: © 2024 IEEE.
PY - 2024/12
Y1 - 2024/12
N2 - Soft robots pose difficulties in terms of control, requiring novel strategies to effectively manipulate their compliant structures. Model-based approaches face challenges due to the high dimensionality and nonlinearities such as hysteresis effects. In contrast, learning-based approaches provide nonlinear models of different soft robots based only on measured data. In this letter, recurrent neural networks (RNNs) predict the behavior of an articulated soft robot (ASR) with five degrees of freedom (DoF). RNNs based on gated recurrent units (GRUs) are compared to the more commonly used long short-term memory (LSTM) networks and show better accuracy. The recurrence enables the capture of hysteresis effects that are inherent in soft robots due to viscoelasticity or friction but cannot be captured by simple feedforward networks. The data-driven model is used within a nonlinear model predictive control (NMPC), whereby the correct handling of the RNN's hidden states is focused. A training approach is presented that allows measured values to be utilized in each control cycle. This enables accurate predictions of short horizons based on sensor data, which is crucial for closed-loop NMPC. The proposed learning-based NMPC enables trajectory tracking with an average error of 1.2° in experiments with the pneumatic five-DoF ASR.
AB - Soft robots pose difficulties in terms of control, requiring novel strategies to effectively manipulate their compliant structures. Model-based approaches face challenges due to the high dimensionality and nonlinearities such as hysteresis effects. In contrast, learning-based approaches provide nonlinear models of different soft robots based only on measured data. In this letter, recurrent neural networks (RNNs) predict the behavior of an articulated soft robot (ASR) with five degrees of freedom (DoF). RNNs based on gated recurrent units (GRUs) are compared to the more commonly used long short-term memory (LSTM) networks and show better accuracy. The recurrence enables the capture of hysteresis effects that are inherent in soft robots due to viscoelasticity or friction but cannot be captured by simple feedforward networks. The data-driven model is used within a nonlinear model predictive control (NMPC), whereby the correct handling of the RNN's hidden states is focused. A training approach is presented that allows measured values to be utilized in each control cycle. This enables accurate predictions of short horizons based on sensor data, which is crucial for closed-loop NMPC. The proposed learning-based NMPC enables trajectory tracking with an average error of 1.2° in experiments with the pneumatic five-DoF ASR.
KW - and learning for soft robots
KW - control
KW - machine learning for robot control
KW - Modeling
KW - optimization and optimal control
UR - http://www.scopus.com/inward/record.url?scp=85209654935&partnerID=8YFLogxK
U2 - 10.1109/LRA.2024.3495579
DO - 10.1109/LRA.2024.3495579
M3 - Article
VL - 9
SP - 11609
EP - 11616
JO - IEEE Robotics and Automation Letters
JF - IEEE Robotics and Automation Letters
SN - 2377-3766
IS - 12
M1 - 11609
ER -