Implementation of Hector SLAM Algorithm for Mapping Indoor Environments with Obstacles

Abstract

Autonomous navigation in obstacle-rich indoor environments is crucial for both industrial and domestic robotic applications. A central aspect of this process is Simultaneous Localization and Mapping (SLAM). This paper presents a comprehensive methodology for implementing Hector SLAM, a widely used SLAM algorithm that operates without relying on external odometry and pose data from wheel encoders or Inertial Measurement Unit (IMU) sensors. The experimental setup involves a custom-built two-wheeled mobile robotic platform, equipped with essential components including a 2D Light Detection and Ranging (LiDAR) sensor, wheel encoders, IMU, motor drivers, an Arduino UNO, and a Raspberry Pi 4B, with Robot Operating System (ROS) serving as the primary software framework. The robot navigates within a predefined rectangular indoor environment, where real-time mapping results are captured at different time intervals to assess mapping accuracy. Additionally, a randomly placed object is introduced to test the obstacle detection capability during operation. Results indicate that Hector SLAM demonstrates high accuracy, achieving precision within a few centimeters, particularly in environments with distinct structural features.