Multi-Sensor Dataset of Ultrasonic and mmWave for Material Classification (MatSense2025)

Dataset Folder Structure:

datasets/
│
├── README-for-all.txt
├──Materials' Thicknesses Details
├── C4001 - Dataset/
│   │
│   ├── C4001 Reflected Signal Dataset – Multiple Materials & Thickness Levels/
│   │   ├── C4001_AllMaterials_AllThickness.csv
│   │   └── README.txt
│   │
│   └── C4001 Reflected Signal Dataset – Multiple Materials/
│       ├── C4001_FiveMaterials_Only.csv
│       └── README.txt
|    |------ Raw Data
|            |----- All Materials Raw Data
│
└── URM09 - Dataset/
    │
    ├── URM09 Reflected Signal Dataset – Multiple Materials & Thickness Levels/
    │   ├── URM09_AllMaterials_AllThickness.csv
    │   └── README.txt
    │
    └── URM09 Reflected Signal Dataset – Multiple Materials/
        ├── URM09_SixMaterials_Only.csv
        └── README.txt
    |   |------ Raw Data
    |            |----- All Materials Raw Data
/////////////////////////////////////////////////////////////////////////////////////////////

Notes:
1- Check the file "Materials' Thicknesses Details" to know the materials thicknesses used in this experiment.
2- Read the methodology for data collection in paper.

--------------------------------------------------------------------------------------------------------------

\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\
How to use datasets:

1- Download the CSV files from this dataset.

2- Load the CSV files into your preferred programming environment
(Python, MATLAB, R, Weka, etc.).

3- Select one or more datasets depending on your experiment needs:

A- AllMaterials_AllThickness → for general classification with multiple thickness levels.

B- Five/SixMaterials_Only → cleaner classification without thickness effects.

4- The label column shows the material and thickness for each sample
(e.g., Plastic-2).

5- Use the feature columns (Mean, RMS, Energy, etc.) as inputs to machine learning algorithms.

6- Split the dataset into training and testing sets (e.g., 80% / 20%) or use N-Folds Cross Validation.

7- Train your ML model and evaluate performance (accuracy, precision, recall).

9- Cite this dataset in your research/publication when using it.
///////////////////////////////////////////////////////////////////////////////////////////////////////