Computationally Efficient Cooperative Dynamic Range-Only SLAM Based on Sum of Gaussian Filter.

Area of Science:

• Robotics and Artificial Intelligence
• Simultaneous Localization and Mapping (SLAM)
• Probabilistic Robotics

Background:

• Cooperative dynamic range-only simultaneous localization and mapping (CDRO-SLAM) algorithms enhance localization accuracy and convergence using inter-node ranges.
• CDRO-SLAM utilizes sum of Gaussian (SoG) filters to track moving nodes in dynamic environments.
• High computational burden associated with inter-node measurements in CDRO-SLAM poses a challenge for real-time applications.

Discussion:

• This paper introduces an efficient implementation of CDRO-SLAM (eCDRO-SLAM) to address the computational demands of the original algorithm.
• A detailed computational analysis demonstrates significant improvements in efficiency for eCDRO-SLAM compared to CDRO-SLAM.
• The performance of eCDRO-SLAM is validated against conventional range-only SLAM algorithms, highlighting its advantages.

Key Insights:

• eCDRO-SLAM significantly reduces computational load while maintaining or improving localization accuracy.
• The proposed algorithm exhibits a faster convergence rate than existing CDRO-SLAM methods.
• Map estimation error remains comparable to other RO-SLAM algorithms, irrespective of map size.

Outlook:

• The eCDRO-SLAM algorithm presents a computationally efficient solution for range-only SLAM applications.
• Its ability to handle dynamic environments and moving nodes makes it suitable for complex robotic systems.
• Further research could explore its integration into multi-robot systems and large-scale mapping scenarios.