Interactive Posture Game System for Children with ADHD

This project implements a therapeutic system based on ROS 2, computer vision, and a simulated humanoid robot. Its purpose is to support the treatment of children with ADHD through a posture memory game guided by the robot, with emotional tracking, auditory feedback, and a mobile application for therapists.

---

Overview

The system consists of three main components:

1. A set of ROS 2 nodes responsible for game logic, posture validation, and emotion recognition.
2. A simulation environment in Gazebo that represents the humanoid robot.
3. An Android mobile application connected via WebSocket to visualize the child’s results and control the game.

---

System Requirements

Recommended Operating System

• Ubuntu 22.04 LTS (Linux)

ROS 2

Developed for ROS 2 Humble. Follow the official installation guide.

---

Step-by-step Installation

1. Create the workspace

mkdir -p ~/ros2tesis_ws/src
cd ~/ros2tesis_ws

2. Clone the repository

cd src
git clone https://github.com/RAMEL-ESPOL/Materia-Integradora-Gonzalez-Lucin.git
cd ..

3. Install system dependencies

Option A — ROS-GZ (Ignition/modern Gazebo)

sudo apt update && sudo apt install -y   python3-colcon-common-extensions python3-pip python3-opencv   ros-humble-desktop   ros-humble-moveit   ros-humble-ros2-control ros-humble-ros2-controllers   ros-humble-ros-gz ros-humble-gz-ros2-control   ros-humble-rosbridge-suite   ros-humble-warehouse-ros-mongo   ros-humble-usb-cam ros-humble-xacro

Option B — Classic Gazebo

sudo apt update && sudo apt install -y   python3-colcon-common-extensions python3-pip python3-opencv   ros-humble-desktop   ros-humble-moveit   ros-humble-ros2-control ros-humble-ros2-controllers   ros-humble-gazebo-ros   ros-humble-rosbridge-suite   ros-humble-warehouse-ros-mongo   ros-humble-usb-cam ros-humble-xacro

4. Install ROS 2 dependencies for the project

rosdep update
rosdep install --from-paths src --ignore-src -r -y

4.5 Optional — Create a virtual environment for dependencies

python3 -m venv .venv && source .venv/bin/activate
pip install --upgrade pip
pip install -r requirements.txt

5. Build the project

colcon build
source install/setup.bash

Add to .bashrc:

echo "source ~/ros2tesis_ws/install/setup.bash" >> ~/.bashrc

---

Running the System

1. Start the robot simulation

ros2 launch yaren_gazebo_sim coco_robot.launch.py

Alternative (for low-resource CPU/GPU computers):

LIBGL_ALWAYS_SOFTWARE=1 ros2 launch yaren_gazebo_sim coco_robot.launch.py

2. Start the main service

ros2 run yaren_services launch_master_service.py

3. Connect the mobile app via WebSocket

ros2 launch rosbridge_server rosbridge_websocket_launch.xml

---

Mobile Application

Developed in Kotlin (Android Studio).

1. Open the project in Android Studio.
2. Connect a physical device or emulator.
3. Run the project.

Inside the app:

• Enter the IP of the computer running ROS 2.
• Press Connect.
• Start the game or access reports.

---

Project Package Structure

Package	Function
posture_game	Game logic, feedback, statistics, posture validation, robot movement, emotion detection
coco_gazebo_sim	Gazebo simulation environment to visualize robot movements
coco_services	Service to send patient names and remotely start launches
rosbridge_server	WebSocket communication with the mobile app
coco_ws	Contains robot information such as URDF, xacro, RViz and Gazebo communication, plus additional implementations
dynamixel_sdk	Communication layer between ROS 2 and ROBOTIS Dynamixel servomotors

---

Package Structure: yaren_memory (posture memory game)

File	Function
game_manage_node.py	Node that controls game logic, feedback, statistics, and assistance
detect_emotions_node.py	Node that processes images and classifies them using a trained machine learning model
mediapipe_node.py	Node that performs image processing with MediaPipe and publishes raw topics for other nodes
speaker_node.py	Node that converts text to speech via a TTS system
checker_node.cpp	Node that validates postures according to response time and required pose
movement_publisher_node.cpp	Node that publishes robot joint configuration commands to move the robot
postures.cpp	Script containing conditions to validate a posture based on MediaPipe landmarks

---

FAQ

Can I use the system without the physical robot?
Yes. It works entirely in simulation using Gazebo.

Which camera is recommended?
Any USB camera compatible with usb_cam.

Where are the reports stored?
As .json files, accessible from the mobile app.

Where can I find the mobile application repository?
The repository is available at: https://github.com/RAMEL-ESPOL/YarenApp.git