Datasets
Standard Dataset
Design and Test of Pneumatic Artificial Muscle
- Citation Author(s):
- Submitted by:
- Carolyn Ren
- Last updated:
- Tue, 04/30/2024 - 17:28
- DOI:
- 10.21227/9ea0-xh30
- Data Format:
- License:
54 Views- Categories:
- Keywords:
Abstract
Description: This supplementary materials file contains:
-All raw data generated in the experiments.
-All MATLAB files required to update the ForceSight application with new data.
-A video of the test procedure for a contraction force measurement of an example actuator.
-License files for BioRender.com
Size: Less than 250 MB
Platform: MATLAB
Environment: Microsoft, Linux, MacOS
Major component description:
-Files named "AllRegressionModelData_V1.0_XX%" are the trained neural networks with XX% of the data set aside for validation.
-"ForceSight_AppDesignerFile" is the MATLAB app design file that is used to make the user interface for the tool.
-"ForceSightInstaller1.0_web.exe" is the executable file to donload the compiled ForceSight tool.
-"Contraction Force Data [COMPLETE]" contains all of the contraction force data in excel file format.
-"Blocked Force Data [COMPLETE]" contains all of the blocked force data in excel file format.
-"PAM Test Procedure"is a video of the testing procedure.
-"RawData.mat" is the all of the raw blocked force data in MATLAB table format.
-"WideNeuralNet_Model" is the packaged wide neural net used to create the forcesight application in MATLAB.
-"Licenses" folder contains the BioRender Licenses.
Detailed setup instructions:
RegressionApp Executable
1. Prerequisites for Deployment
Verify that MATLAB Runtime(R2023a) is installed.
If not, you can run the MATLAB Runtime installer.
To find its location, enter
>>mcrinstaller
at the MATLAB prompt.
NOTE: You will need administrator rights to run the MATLAB Runtime installer.
Alternatively, download and install the Windows version of the MATLAB Runtime for R2023a
from the following link on the MathWorks website:
https://www.mathworks.com/products/compiler/mcr/index.html
For more information about the MATLAB Runtime and the MATLAB Runtime installer, see
"Distribute Applications" in the MATLAB Compiler documentation
in the MathWorks Documentation Center.
2. Files to Deploy and Package
Files to Package for Standalone
================================
-RegressionApp.exe
-MCRInstaller.exe
Note: if end users are unable to download the MATLAB Runtime using the
instructions in the previous section, include it when building your
component by clicking the "Runtime included in package" link in the
Deployment Tool.
-This readme file
3. Definitions
For information on deployment terminology, go to
https://www.mathworks.com/help and select MATLAB Compiler >
Getting Started > About Application Deployment >
Deployment Product Terms in the MathWorks Documentation
Center.
Detailed run instructions:
-to add data to the model simply combine it with the RawData.mat file and
re-run the regression learner application with your desired parameters.
Output description:
-The regression learner ap will output a .mat file for the model that can replace
the "WideNeuralNet_Model.mat" file and be integrated into the "ForceSight_AppDesignerFile.mlapp"
Contact Information: Jordan Savage: ja2savag@uwaterloo.ca
1 Raw Data Files
Relevant files/folders: “Blocked Force Data [COMPLETE]”, “Contraction Force Data [COMPLETE]”, “RawData.mat”, “TrainingData20.mat”, “TestData20.mat”, “NeuralNet_20%_TestData.mat”
A lot of data was required to form the predictive neural network. All the blocked force and contraction force data can be found in their associated folders categorized by the manufacturing type and bladder thickness first then by diameter and length. The standard name format for the csv files is as follows: “ManufacturingMethod_Thickness_Diameter_Length.csv”.
Within the “RegressionModelData” folder the following files: “RawData.mat”, “TrainingData.mat”, and “TestData.mat” contain all the blocked force data in a .mat format, the training data, and the randomly selected test data generated from the “RawDataFormatting.m” file respectively. The regression learner session is saved as: NeuralNet_20%_TestData.mat. Which contains all the trained models examined within the paper.
2 App Files
Relevant files/folders: “AppFiles”
This folder contains the app design file “ForceSight_AppDesignerFile.mlapp” used to generate the user interface for ForceSight. It also contains the trained model file “WideNeuralNet_V2.mat”, and the executable installation file for the app “ForceSightInstaller1.0_web.exe”.
To update the model being used replace the “WideNeuralNet_V2.mat” file with your exported model from the regression learner app. Then open the MATLAB App designer tool and load the “ForceSight_AppDesignerFile.mlapp” file in the AppDesignFiles folder. Transition to code view and scroll down to line 117 to change the name of the model file being loaded. On line 174 the name of the model should be changed from ‘rainedModelWide100’ to match your model’s variable name in the MATLAB workspace.
In the “ForceSightV1.0 Download files” folder the packaged app can be downloaded onto any computer by installing it using the “ForceSightInstaller1.0_web.exe” file.
3 Licenses
Relevant files/folders: “Licenses”
Contains the required BioRender.com license files for the images generated with the software that are contained within the publication.
4 Licenses
