Vibration Suppression in Robotic Polishing Task Data

<pre>This dataset contains the real-world robotic polishing experiments done for the paper titled "Online Learning for Vibration Suppression in Physical Robot Interaction using Power Tools" by Gokhan Solak and Arash Ajoudani.<br>[arXiv]<br><br>A video of the experiment footage and the paper summary can be found here.<br><br>The data contains the control and motion variables logged during execution, as <i>mat</i> files. <br><br><b>Folder structure</b><br><br>The folders under <i>data</i>: <i>drill080</i>, <i>drill100</i> and <i>drill120</i> contain the parameter selection and comparison experiment logs for each drill mode.<i>data/params</i> contains example parameter values for required to run some scripts.<br><br>The <i>scripts </i>folder contains the Matlab scripts for reading and analysing the files.<i>scripts/bmflc</i> contains the example implementations of the methods described in the paper.<br><br>├── data<br>│   ├── drill080<br>│   │   ├── parameter_selection<br>│   │   │   ├── damped<br>│   │   │   └── original<br>│   │   └── stepsize_comparison<br>│   │   ├── damped<br>│   │   ├── none<br>│   │   └── original<br>│   ├── drill100<br>│   │   ├── parameter_selection<br>│   │   │   ├── damped<br>│   │   │   └── original<br>│   │   └── stepsize_comparison<br>│   │   ├── damped<br>│   │   ├── none<br>│   │   └── original<br>│   ├── drill120<br>│   │   ├── parameter_selection<br>│   │   │   ├── damped<br>│   │   │   └── original<br>│   │   └── stepsize_comparison<br>│   │   ├── damped<br>│   │   ├── none<br>│   │   └── original<br>│   └── params<br>└── scripts<br> ├── bmflc<br> └── utils<br><br><b>Running</b><br><br>In Matlab, <br>- change directory to the top folder and add all subfolders to path.<br>- run the script <code><em>analyze_results.m</em></code>. It will create the real-world experiment figures from the paper.<br><code>- </code><code><em>bmflc_time_analysis.m</em></code> runs the step-size timing benchmark and creates <i>Figure 6</i>.<br><code>- </code><code><em>plot_logistic_param_effects.m</em></code> visualizes the logistic step-size by creating <i>Figure 3</i>.</pre><br>