LIDAR based Mapping and Path Planning for Autonomous Vehicles

This paper presents a low-cost autonomous navigation system for structured indoor spaces based on 2D LiDAR for real-time mapping, localization, and dynamic path planning. The system utilizes Hector SLAM for odometry-free localization and mapping, generating precise pose estimates and occupancy grid maps. Environmental information is preprocessed with DBSCAN for noise filtering, and obstacle detection is improved through temporal filtering to remove static objects. An A*-based algorithm for path planning guarantees optimal and safe navigation by constantly adapting to environmental variations. The entire navigation stack is programmed in ROS and verified using Gazebo simulations of dynamic indoor environments. Experiments show high mapping accuracy, a 92% success rate for detecting obstacles, and small spatial drift of 1.8 cm in a 12×12-meter space. The system proposed provides a scalable, effective solution for indoor automation and GPS-denied navigation operations.