Autonomous Navigation and Object Localization in Unstructured Environments Using SLAM and Path Planning Algorithms on Turtlebot4

This paper aims to perform autonomous navigation and object localization in unstructured, unknown environments. We widely use the SLAM algorithm to map an environment and estimate the robot's position, particularly in dynamic and unstructured environments. This paper proposes an open-source ROS2 Slam Toolbox to solve this problem. Nav2 integrates with Slam Toolbox to enable real-time positioning in dynamic environments for autonomous navigation. The robot localizes and tracks its position in the known environment of the SLAM map using the AMCL algorithm. This research also investigates various path-planning algorithms, such as Dijkstra's algorithm, A*, D*, and the Dynamic Window Approach, to plan collision-free paths in 2D navigation. It also uses the Nav2 plugin of the ROS2 framework to determine optimal path planning and navigate the robot from one place to another with appropriate speed. In this research, the Nav2 vector pursuit controller shows secure mobility, precise positioning, and high-speed, accurate path tracking. We have implemented and validated this research approach on the Clearpath Turtlebot4, demonstrating large-scale autonomous navigation to facilitate optimal, safe, and effective mobile robot navigation and localization in various unknown environments.