Research of Autonomous Navigation for Mobile Robots Using Karto SLAM Algorithm Under ROS

Researching navigation for mobile robots to follow the desired trajectory is crucial to ensure their swift and accurate movement while avoiding obstacles. Most mapping techniques rely on Simultaneous Localization and Mapping (SLAM), which involves creating a map and determining the robot's position by collecting data from sensors like Lidar and cameras. The Karto SLAM algorithm is a graph optimization method that utilizes ghosts. It optimizes the Cholesky factorization and does not require iterating to solve sparse systems. The map is represented by the mean value of the histogram, with each node denoting a point on the robot's trajectory and sensor data. When a new node is added, the map is recalculated and updated based on the node's spatial constraints. In real-world scenarios, Karto SLAM exhibits minimal error (1.03 cm), making it the preferred choice for mobile robots. The accuracy of this navigation solution is evident. The correctness of this navigation solution is demonstrated through ROS_Gazebo simulation.