An adaptive scheme for robot localization and mapping with dynamically configurable inter-beacon range measurements

Area of Science:

• Robotics
• Sensor Networks
• Simultaneous Localization and Mapping (SLAM)

Background:

• Robot-sensor network cooperation is crucial for localization and mapping.
• Range-only (RO) SLAM faces challenges with landmark (beacon) measurement accuracy and resource management.

Purpose of the Study:

• To develop a dynamic RO-SLAM scheme that optimizes measurement gathering for enhanced accuracy and efficiency.
• To balance SLAM performance (accuracy) with resource consumption (computational burden, energy).

Main Methods:

• Implemented a RO-SLAM scheme with a dynamic measurement gathering module and a supervision module.
• Configured measurement collection for robot-beacon and inter-beacon data at varying rates and depths.
• Utilized an Extended Kalman Filter SLAM with auxiliary Particle Filters for beacon initialization (PF-EKF SLAM).

Main Results:

• Achieved 34% lower map errors and 14% lower robot errors compared to traditional methods.
• Reduced computational burden by 16% while maintaining similar beacon energy consumption.
• Validated through experiments on the CONET Integrated Testbed.

Conclusions:

• The proposed adaptive RO-SLAM scheme effectively improves localization and mapping accuracy.
• Dynamic measurement management offers a significant reduction in computational requirements for SLAM systems.
• This approach is suitable for resource-constrained robot-sensor network applications.