Indoor Localization and Measurement Object in Field for TurtleBot Using Combined 2D-LiDAR and Monocular Camera

Abstract Robotic technology has developed in various ways to challenge problems in complex and modern situations. Working in indoor spaces creates many extra complications due to intricacy in spaces and due to signal instability, that tends to make navigation efficiency difficult for mobile robots. Localizing objects correctly and Measurement them is fundamental. This was achieved using a SLAM system with 2D LiDAR to map the environment and added object detection from a monocular camera that identifies and tracks specific items within the view of the robot. It improves the view of the robot by fusing LiDAR spatial accuracy with camera capability of object recognition. After this data fusion within the ROS framework, the system coordinates outputs of both sensors using a publisher-subscriber model for real-time integration of distance measurements and visual detections. Further, the LiDAR data is filtered using a Kalman filter to get rid of noise, thus achieving stable measurements of distances and eventually enhancing localization reliability. Equipped with the proper field of view calibration between the camera and LiDAR, it is enabled to track and position an object steadily and precisely, which is vitally useful when moving about dynamic indoor spaces. The TurtleBot platform is used to implement this system, enabling the robot to autonomously explore and create maps of indoor environments. The YOLO and Deep SORT algorithms applied to the camera data allow for efficient real-time object detection and Measurement, while the SLAM-based mapping from LiDAR ensures a comprehensive view of the environment. This multi-sensor fusion technique thus addresses the challenges of indoor navigation, facilitating reliable obstacle detection, accurate localization, and robust path planning, ultimately allowing the robot to operate autonomously and respond to environmental changes in real time.