Improving Robot Model Predictions with NARX Bayesian Neural Networks
Author Information
Author(s): Alaa Aldeen Joumah, Jafar Assef, Chadi Albitar
Primary Institution: Higher Institute for Applied Sciences and Technology (HIAST), Damascus, Syria
Hypothesis
Can a modified NARX Bayesian Neural Network enhance the prediction of the forward geometric model for a 6-RSU parallel manipulator?
Conclusion
The NARX-BNN model significantly improves prediction accuracy and reduces uncertainty in robotic applications compared to traditional methods.
Supporting Evidence
- The NARX-BNN reduces the RMSE of predicted values by up to 11%.
- The Average Width indicator of the prediction interval is reduced by approximately 12.7%.
- The proposed model outperforms traditional Bayesian shallow neural networks in accuracy.
- Experimental results validate the effectiveness of the NARX-BNN model.
Takeaway
This study shows that a special type of neural network can help robots understand their movements better, making them safer and more accurate.
Methodology
The study used simulations and experiments to compare the performance of a modified NARX Bayesian Neural Network with traditional Bayesian neural networks.
Potential Biases
Potential biases may arise from the selection of datasets and the specific configurations of the neural networks used.
Limitations
The study's findings may not be generalizable to all types of robotic systems or manipulator configurations.
Statistical Information
P-Value
p<0.05
Confidence Interval
95%
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website