Improving Sensor Adaptability and Functionality in Cartographer Simultaneous Localization and Mapping

Area of Science:

• Robotics
• Computer Vision
• Sensor Fusion

Background:

• Simultaneous Localization and Mapping (SLAM) systems face challenges in adaptability across diverse robotic applications.
• The open-source Google Cartographer project, a graph-based SLAM implementation, requires enhancements for broader functionality.

Purpose of the Study:

• To develop a novel SLAM framework to address sensor-related challenges and improve the adaptability of Google Cartographer.
• To integrate advanced functionalities enhancing SLAM system utility in varied robotic scenarios.

Main Methods:

• Developed a novel SLAM framework integrating five functionalities into Google Cartographer and Robot Operating System (ROS).
• Implemented an inertial data generation system and a sensor data preprocessing system.
• Introduced real-time 3D topographic mapping, multi-node SLAM, and elliptical sensor data filtering.

Main Results:

• The new framework demonstrated low computational overhead with average execution times of 0.55 s for preprocessing and 0.15 ms for virtual inertial data generation.
• Elliptical sensor data filtering achieved performance comparable to existing schemes.
• Enhanced system utility through added functionalities like 3D mapping and multi-node capabilities.

Conclusions:

• The proposed SLAM framework effectively addresses sensor-related challenges, enhancing the adaptability and functionality of Google Cartographer.
• The integrated features provide a robust solution for diverse robotic applications, improving real-time mapping and localization accuracy.