Investigating the Fiducial Marker Network Characteristics for Autonomous Mobile Indoor Robot Navigation Using ROS and Gazebo | Journal of Construction Engineering and Management | Vol 148, No 10

AbstractBuilding service robots rely on dense instrumentation of the building (e.g., Bluetooth beacons) or require high computational capabilities [e.g., simultaneous localization and mapping (SLAM)]. To overcome these limitations, studies explored a landmark-based localization and navigation approach based on inexpensive, computationally efficient, and easily configurable fiducial markers (e.g., AprilTags). However, context-specific assumptions were made regarding the fiducial marker characteristics and sensor configurations. Taking this forward, this study develops a generalized simulation-based approach to determine the optimal design characteristics of the fiducial marker network to achieve successful autonomous mobile indoor robot navigation with low instrumentation and computation. Different marker and camera parameters such as marker size and marker placement to optimize the density of fiducial markers (i.e., the ideal distance between subsequent markers) were investigated using Robot Operating Systems (ROS) and Gazebo (simulator) software. The simulation experiments focused on a specific robotic platform (i.e., TurtleBot3) and marker type (i.e., AprilTag). Results from the simulations suggested that with an increase in marker distance, the navigation success rate does not necessarily decrease. In addition, a marker size of 0.1 m performed the best in terms of navigation success rate (as high as 100% for a specific combination of marker height, marker size, and marker-to-marker distance). Future work aims to test this in a real-world setting to compare and analyze the simulated and actual performance of the robot. The proposed methodology is generic and can be applied to any mobile robotic platform, marker type, building type (e.g., residential), and application (e.g., construction progress monitoring).

Source link