Hovercraft Target Follower (ROS 2 Remake)

This repository is a remake of the SEMEAR's "Projeto de Bixo" (2nd group of 2025). It was developed with the intention of porting the original project to ROS 2 Kilted, improving Python implementation standards, and deepening knowledge in the ROS2 framework.

Overview

The project consists of a hovercraft (simulated in Gazebo) capable of:

1. Searching: Rotating in place to find a specific color target (blue).
2. Tracking: Once detected, the robot uses a PID controller to center the target in the camera's FOV and moves towards it.
3. Autonomous State Management: Managed via a Finite State Machine (FSM).

---

Tech Stack

• Middleware: ROS 2 Kilted
• Language: Python 3
• Vision: OpenCV (HSV Filtering, Contour Detection)
• Simulation: Gazebo
• Key Packages: cv_bridge, geometry_msgs, sensor_msgs, std_msgs

---

System Architecture

The project is divided into two main nodes:

1. Vision Node (camera_node):

   • Subscribes to /camera/.
   • Processes images using HSV masks and morphological operations.
   • Publishes object center, area, and estimated distance to /object_info.

2. Control Node (control_node):

   • Implements a Finite State Machine (FSM).
   • Calculates angular velocity using a PID controller to minimize the centering error.
   • Publishes velocity commands to /cmd_vel.

How to run?

After installing ROS2 Kilted, OpenCV and Bridge libraries, make sure to grant permission to python archives in the 'src' folder. Run the standard colcon build and source install/setup.bash commands and then:

ros2 launch remake_hover simu_seca.launch.py