Dynamic analysis and parameter identification for robotic manipulators
Wang, Qi (2023-11-20)
Väitöskirja
20.11.2023
Lappeenranta-Lahti University of Technology LUT
Acta Universitatis Lappeenrantaensis
School of Energy Systems
School of Energy Systems, Konetekniikka
Kaikki oikeudet pidätetään.
Julkaisun pysyvä osoite on
https://urn.fi/URN:ISBN:978-952-412-003-6
https://urn.fi/URN:ISBN:978-952-412-003-6
Kuvaus
ei tietoa saavutettavuudesta
Tiivistelmä
This paper presents a comparative advanced study on parameter identification and dynamic analysis of lightweight and heavy-duty robotic manipulators. The ultimate goal is to enhance the performance and control of these manipulators through precision in parameter estimation.
The initial section investigates the dynamics of harmonic drive systems in lightweight arms. Here, a data acquisition-based algorithm is introduced for offline dynamic parameter analysis. The Markov chain Monte Carlo (MCMC) method is central to this analysis, facilitating accurate parameter estimation for a friction model while considering uncertainty and sensitivity. Furthermore, the prediction of the GMS model, based on the MCMC method, exhibited a commendable improvement in accuracy, further emphasizing the efficacy of the chosen methods.
Subsequent sections explore heavy-duty arms, particularly those with planetary gearboxes. These discussions focus on friction, hysteresis issues, and the complexities of parameter estimation. The Bouc-Wen model is highlighted as a useful tool for identifying and addressing errors caused by hysteresis.
Beyond core analysis, the paper shows the potential combination of digital twin and control robot technologies for remote maintenance, especially in fusion reactor circumstances. Integration of these technologies promises to improve the operational capabilities of robotic systems, resulting in more reliable remote maintenance in challenging conditions.
The results of our various investigations provided a thorough understanding of robot dynamics and advanced parameter estimations. The focus of the first study was on identifying the joint dynamics of the robot. Incorporating the friction model for simulation of harmonic drives, along with the unique perspective of hysteresis characteristics, has enriched this understanding even further. Furthermore, the MCMC and SGHMC algorithms were thoroughly evaluated and validated, with the first exhibiting enhanced prediction accuracy by more than 5% and the second demonstrating robustness in parameter estimation for heavy-duty manipulators.
Based on our findings, integrating these methods, especially the GMS model, significantly impacts the field. In our subsequent research, the fractional-order Bouc-Wen (FOBW) model has emerged as a key tool for illustrating hysteresis behaviors in different systems. It is based on the popular Bouc-Wen model and has been expanded to include more features. Our research underscores the importance of these advanced tools and models, suggesting an important development in the reliability and accuracy of robotic dynamic control in complex real-world situations.
The initial section investigates the dynamics of harmonic drive systems in lightweight arms. Here, a data acquisition-based algorithm is introduced for offline dynamic parameter analysis. The Markov chain Monte Carlo (MCMC) method is central to this analysis, facilitating accurate parameter estimation for a friction model while considering uncertainty and sensitivity. Furthermore, the prediction of the GMS model, based on the MCMC method, exhibited a commendable improvement in accuracy, further emphasizing the efficacy of the chosen methods.
Subsequent sections explore heavy-duty arms, particularly those with planetary gearboxes. These discussions focus on friction, hysteresis issues, and the complexities of parameter estimation. The Bouc-Wen model is highlighted as a useful tool for identifying and addressing errors caused by hysteresis.
Beyond core analysis, the paper shows the potential combination of digital twin and control robot technologies for remote maintenance, especially in fusion reactor circumstances. Integration of these technologies promises to improve the operational capabilities of robotic systems, resulting in more reliable remote maintenance in challenging conditions.
The results of our various investigations provided a thorough understanding of robot dynamics and advanced parameter estimations. The focus of the first study was on identifying the joint dynamics of the robot. Incorporating the friction model for simulation of harmonic drives, along with the unique perspective of hysteresis characteristics, has enriched this understanding even further. Furthermore, the MCMC and SGHMC algorithms were thoroughly evaluated and validated, with the first exhibiting enhanced prediction accuracy by more than 5% and the second demonstrating robustness in parameter estimation for heavy-duty manipulators.
Based on our findings, integrating these methods, especially the GMS model, significantly impacts the field. In our subsequent research, the fractional-order Bouc-Wen (FOBW) model has emerged as a key tool for illustrating hysteresis behaviors in different systems. It is based on the popular Bouc-Wen model and has been expanded to include more features. Our research underscores the importance of these advanced tools and models, suggesting an important development in the reliability and accuracy of robotic dynamic control in complex real-world situations.
Kokoelmat
- Väitöskirjat [1212]