Mobile Robots: Mapping, Landmarks, Localisation and a ROS 2 Mission

• Autonomous exploration and global point-cloud mapping with a voxel grid filter.
• Cone and container detection from laser scans, published as RViz marker arrays.
• A particle filter with predict, weight and systematic resampling steps.
• A SMACH state machine driven by colour and depth detectors.
• The package, launch file and state-machine design for the YASMIN and YOLO patrol mission.

Practicals from the Mobile Robots course. In ROS 2, one node builds a global point-cloud map from laser scans with a voxel grid filter, and another voxelises in 3D with Open3D and TF2. Landmark nodes cluster each scan by range continuity, fit circles to find cones and lines to find containers, and publish RViz markers.

A particle filter in Python localises a TurtleBot on an occupancy map, and a ROS 1 SMACH state machine reacts to colour and depth obstacle detection.

The final project is a ROS 2 patrol mission: select a waypoint, navigate, detect, spin as a gesture, with abort and end outcomes. The package, launch file, map and state diagram are in the repo.