BIRUNI: A ROS 2-BASED AUTONOMOUS MECANUM-WHEELED MOBILE ROBOT WITH MULTI-SENSOR FUSION, REAL-TIME SLAM, AND NAVIGATION IN GAZEBO FORTRESS -2025

Abstract

This document presents Biruni, a ROS 2-based autonomous mecanum-wheeled mobile robot designed and evaluated in a high-fidelity Gazebo Fortress simulation environment. The robot integrates a multi-sensor perception architecture consisting of a 2D LiDAR, an RGB camera, an Inertial Measurement Unit (IMU), and wheel encoders. A unified Extended Kalman Filter (EKF) is implemented to fuse inertial and odometry data, generating a stable and drift-minimized state estimate for real-time navigation. Cartographer SLAM is used to produce consistent 2D occupancy maps, while the Navigation2 (Nav2) framework provides global and local planning, obstacle avoidance, and autonomous waypoint following. The complete system is structured around a modular ROS 2 package architecture with a fully validated TF tree, URDF/Xacro robot model, and synchronized sensor frames. Simulation results demonstrate accurate localization, reliable mapping, and smooth path execution with omnidirectional control using mecanum kinematics. The Biruni platform offers a reproducible foundation for education, research, and future deployment on physical hardware.