التفاصيل البيبلوغرافية
| العنوان: |
CNN–AUPI-Based Force Hysteresis Modeling for Soft Joint Actuator. |
| المؤلفون: |
Chen, Shitao1 (AUTHOR), Xu, Ming1 (AUTHOR) jxxuming@hdu.edu.cn, Liu, Shuo1 (AUTHOR), Liu, Hui1 (AUTHOR), Su, Lirong1 (AUTHOR) |
| المصدر: |
Arabian Journal for Science & Engineering (Springer Science & Business Media B.V. ). Nov2024, Vol. 49 Issue 11, p14577-14591. 15p. |
| مصطلحات موضوعية: |
*CONVOLUTIONAL neural networks, *COMPRESSIBILITY, *ACTUATORS, *GENERALIZATION, *PNEUMATIC actuators |
| مستخلص: |
The output force of soft-pneumatic actuators is critical for their applications. However, the hyperelasticity of its material and the high compressibility of the actuation medium, i.e., air, lead to a strong asymmetric hysteresis, which poses a great challenge for its force control. An attention CNN (Convolution Neural Network)–AUPI (Amplitude-dependent Un-parallel Prandtl–Ishlinskii)-based force-position hysteresis modeling method for soft actuators is proposed. Based on the preliminary fitting of the asymmetric hysteresis curve by AUPI, the attention CNN mechanism extracts the global features of the hysteresis information, prevents the model from overfitting, and greatly improves the accuracy and generalization ability of the composite model. Experiments show that the CNN–AUPI model has excellent hysteresis fitting for soft joint actuators, with a maximum relative error of only 6.1% and a goodness-of-fit of more than 0.99. The proposed CNN–AUPI model is also compared with other hysteresis models, and the results show that the CNN–AUPI model not only has a high modeling accuracy but also has a strong prediction capability, which provides a promising method for hysteresis modeling of soft actuators. [ABSTRACT FROM AUTHOR] |
| قاعدة البيانات: |
Academic Search Index |