Comparison between cascade forward and multi-layer perceptron neural networks for NARX functional electrical stimulation (FES) based muscle model / Abu Huzaifah Abu Hassan
This paper exhibits the improvement and correlation of muscle models taking into account FES incitement parameters utilizing the Nonlinear Auto-Regressive model with Exogenous Inputs (NARX) using Multi-Layer Perceptron (MLP) and Cascade Forward Neural Network (CFNN). FES stimulations with varying fr...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English |
| Published: |
2016
|
| Subjects: | |
| Online Access: | https://ir.uitm.edu.my/id/eprint/69020/1/69020.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | This paper exhibits the improvement and correlation of muscle models taking into account FES incitement parameters utilizing the Nonlinear Auto-Regressive model with Exogenous Inputs (NARX) using Multi-Layer Perceptron (MLP) and Cascade Forward Neural Network (CFNN). FES stimulations with varying frequency, pulse width and pulse duration were utilized to evaluate the muscle torque. 722 data points' focuses were utilized to make muscle model. One Step Ahead (OSA) prediction, correlation tests, and residual histogram analysis were performed to accept the model. The ideal MLP results were obtained from input lag space of 1, output lag space of 43, and hidden units 30. A total of three terms were selected to construct the final model, namely ul (t - 1), y (t - 1), and u4 (t - 1). The last MSE delivered was 1.1299. The optimal CFNN results were gained from input lag space of 1, output lag space of 5, and hidden units 20. The terms selected are similar to that of the MLP model. The final MSE produced was 1.0320. The proposed approach figured out how to rough the system well with unbiased residuals, with CFNN demonstrating 8.66% MSE change over MLP with 33.33% less hidden units. |
|---|