Indoor Autonomous Navigation System of Quadrotor UAV Based on laser SLAM and Dynamic Window Method

Abstract

To address key challenges in autonomous navigation of quadrotor drones under GPS-denied environments — such as positioning failure, complex surroundings, and difficulties in real-time obstacle avoidance —this paper designs and implements an indoor autonomous navigation system for quadrotor drones based on LiDAR-based Simultaneous Localization and Mapping (SLAM) and the Dynamic Window Approach (DWA). The system adopts a layered architecture, using a TIVA microcontroller as the core flight control unit at the lower level and a Raspberry Pi 4B as the onboard decision-making and data processing module. It employs the Hector LiDAR SLAM algorithm to achieve real-time mapping of indoor environments and drone pose estimation, while utilizing the DWA method for local path planning and dynamic obstacle avoidance. A 25 × 25 m simulation environment is built using Python, with static and moving obstacles set up to validate the algorithms. System integration simulation is conducted on the ROS-Gazebo platform, and comparative indoor localization and navigation tests are performed on a custom-built physical drone platform. Experimental results show that the proposed Hector-SLAM achieves a positioning error within ±0.08 m, generates maps closely matching the actual environment, and significantly reduces drift. The optimized DWA algorithm enables stable flight at 1.2 m/s, responds to sudden dynamic obstacles in less than 0.3 s, and maintains a mean global trajectory deviation below 0.12 m. The entire system demonstrates stable autonomous localization, high-precision map building, and safe real-time navigation in unknown and complex indoor environments. It produces smooth flight trajectories, reliable and responsive obstacle avoidance, and high positioning accuracy, offering strong practicality and robustness. This system provides a reliable technical solution and engineering reference for drone applications such as indoor inspection, detection, and reconnaissance in GPS-denied scenarios.