ROS - Shimmer3 Project

This project has been implemented for University of Bristol MSc Robotics Dissertation module under the supervision of Alex Smith and Anouk Van Maris.

There are a couple of dependencies, and you will require them to run this project.

Please install the following dependencies:

• https://github.com/dpkoch/async_comm

How to install the project

After installing related dependecities, get the "my_example" directory and source the ROS files by typing:

source devel/setup.bash

Then, the ROS files need to be built by using the following command:

catkin_make

After successfully making the project, we can run the ROS nodes using the rosrun command. However, the com port needs to be configured in Linux environment for the Bluetooth setup before running the publisher node. The following steps will set the com port on linux environment.

• At first, we need to find the shimmer3 MAC address by searching via the Blueooth (Please be ensure the Shimmer3 is open). So, the following command can be used on terminal to find the possible Bluetooth devices.

  hcitool scan

  This command will search all the Bluetooth devices with their MAC address around. After finding the shimmer3, copy the corresponding MAC address to open a new port.

• After coying the MAC address, the following code will open a new port between Shimmer3 and Linux environment. The port_numer can be selected any, but if there is no port opened, we can set it simply 0.

  sudo rfcomm bind port_number MAC_address

• To check whether the intended port has been opened or not, please simply type rfcomm command on the terminal. This should show you the related port with the settled parameters.

• After opening the new port between Shimmer3 and Linux environment, we need to give permission to the opened port by using the command follow(The same port_number needs to be put as we did above, for example, rfcomm0):

  sudo chmod a+rw /dev/rfcomm(port_number)

• Now we can run the publisher node as follows:

  rosrun my_example_pkg publisher /dev/rfcomm(port_number)

• The robot_controller node can also be run using the same command:

  rosrun my_example_pkg robot_controller

• The stress_detection.py python script can be found under the dissertation_project/my_example/src/my_example_pkg/src/, and it runs separately on the VScode environment (you can use it in your favourite IDE). Please be aware that the python script has its library dependencies. You should install all of them before running the node.

Example images from the project

The Gazebo simulation utilized during the experiment can be found following the link:

https://github.com/Frkncm/Gazebo_design

Regarding WESAD_analysis_and_ML.ipynb file

This file is created to analyse the WESAD dataset and train the SVM and KNN models. You can find all the code steps, but to run it, you will need to download the WESAD dataset that can be found from the following link below (Since the WESAD file is too large, it is left to the user who wants to run the codes).

https://archive.ics.uci.edu/ml/datasets/WESAD+%28Wearable+Stress+and+Affect+Detection%29

Alternatively, to directly download it, please follow the link below:

https://uni-siegen.sciebo.de/s/HGdUkoNlW1Ub0Gx