Illustrations generated with the same algorithms and parameters as the code below (Gaussian 5x5 + Canny 50/150, 4-connected A with a Manhattan heuristic, PID 1.0 / 0.1 / 0.01 at 10 Hz, DICT_4X4_50
- solvePnP, constant-velocity Kalman filter) on example data.*
Five C++ / ROS 2 / OpenCV exercises, each built around one robotics software topic: multithreading and memory safety, a planner exposed as a ROS 2 service, an asynchronous control loop as a ROS 2 action, 3D pose estimation from a camera, and a Kalman filter in composed nodes with zero-copy intra-process transport.
# Create a workspace
mkdir -p ~/ros2_challenges_ws/src
cd ~/ros2_challenges_ws/src
# Clone the repository
git clone https://github.com/NicoC54/Robotics-Vision-Control-Challenges.git
# Build the workspace
cd ..
colcon build --symlink-install
source install/setup.bash
Objective: Design the software core of a camera node. It continuously captures images in one thread while applying a "Plug & Play" sequence of filters (e.g., Gaussian Blur, Canny) in another.
- Engineering Focus: Multithreading, Mutexes, and memory safety.
- Source Code: View perception_pipeline/src
Objective: A complete ROS 2 Service/Client architecture. The node exposes a GetPath Service taking a 2D occupancy grid, a start point, and a goal, computing the shortest path using the A* algorithm. The Client node sends the request and processes the trajectory.
- Engineering Focus: Algorithmic optimization and strict C++ Standard Template Library (STL) container selection.
- Source Code: View astar_service/src
Objective: A ROS 2 Action Server for a mobile robot. Given a "Follow Object X" goal, the server continuously listens to the TF2 tree to compute the distance to the target, applies a PID controller to the distance error, and streams the resulting speed command as action feedback.
- Engineering Focus: Asynchronous control loops and automatic control theory.
- Source Code: View action_pid/src
Objective: Extract 3D spatial data from a 2D image. The node detects an ArUco marker, uses the camera's intrinsic matrix to compute its 3D pose via solvePnP, and broadcasts it to the ROS 2 network.
- Engineering Focus: Spatial vision and 3D mathematics.
- Source Code: View aruco_node/src
Objective: Two ROS 2 Component Nodes (a Publisher and a Subscriber) running in the same process. Node A publishes camera frames. Node B subscribes, reads the image in Zero-Copy mode, extracts a noisy 2D position, and filters it through a Linear Kalman Filter to predict the true state.
- Engineering Focus: IPC (Inter-Process Communication), ROS 2 Node Composition, and Sensor Fusion.
- Source Code: View tracker_kalman/src
I am currently seeking full-time opportunities as a Robotics Software Engineer or Computer Vision Engineer.
