Related Experiment Videos
A Novel Recurrent Neural Network for Manipulator Control With Improved Noise Tolerance
IEEE Transactions on Neural Networks and Learning Systems
|April 20, 2017
Summary
This study introduces a new recurrent neural network for manipulator kinematic control, effectively eliminating polynomial noise to achieve accurate trajectory tracking. The proposed method ensures system stability and minimal tracking errors, outperforming existing solutions.
Area of Science:
- Robotics
- Control Systems
- Artificial Intelligence
Background:
- Manipulator kinematic control is essential for robotic tasks.
- Noise in polynomial form can degrade control accuracy.
- Existing neural network solutions face stability issues with certain noise types.
Purpose of the Study:
- To propose a novel recurrent neural network for robust manipulator kinematic control.
- To address and eliminate the impact of polynomial noises on trajectory tracking.
- To enhance system stability and reduce position tracking errors.
Main Methods:
- A novel recurrent neural network architecture is designed.
- The network leverages high-order derivative properties of polynomial noises.
- Theoretical analysis and extensive simulations are used for verification.
Main Results:
- The proposed neural network stabilizes system dynamics under polynomial noise.
- Position tracking errors converge to zero even with noise present.
- The approach demonstrates superior performance compared to existing dual neural solutions, especially with large or time-varying noises.
Conclusions:
- The novel recurrent neural network effectively resolves manipulator redundancy and noise issues in kinematic control.
- The proposed method achieves accurate trajectory tracking with stability guarantees.
- This work offers a robust solution for robotic control in noisy environments.