Autonomous Navigation Rover on ROS 2

Problem

Autonomous systems need reliable localisation, mapping and obstacle-aware navigation before higher-level behaviour matters.

Architecture

LiDAR and IMU inputs feed modular ROS 2 nodes for mapping and EKF state estimation. Nav2 consumes the resulting map and fused pose for costmaps, planning, control and recovery behaviour, with Gazebo Fortress and RViz providing repeatable inspection.

Design decisions

1. 01Kept perception, estimation, planning and control modular so each layer could be tuned and validated independently.
2. 02Regression-tested planning, costmaps, controller gains and recovery behaviour in simulation.

Verification

Gazebo Fortress regression runs and RViz inspection checked maps, transforms, fused pose, planned paths and recovery behaviour in simulation.

Engineering case study

Read the design rationale, verification scope and proposed next experiments.