Simultaneous Localization and Mapping Methods for Snake-like Robots Based on Gait Adjustment.

Area of Science:

• Robotics
• Computer Vision
• Sensor Fusion

Background:

• Snake robots need accurate autonomous localization and mapping (SLAM) for field operations.
• Their complex motion, including large turns and fast rotations, causes SLAM errors like drift and failure.
• Head-shaking in snake robots particularly hinders SLAM algorithm convergence.

Purpose of the Study:

• To develop improved gait motion and sensor fusion methods for robust snake robot SLAM.
• To address the challenges posed by snake robot locomotion on positioning and map-building accuracy.
• To enhance the reliability and performance of SLAM systems in challenging robotic applications.

Main Methods:

• Evaluated the suitability of various snake robot gaits for SLAM.
• Developed a head stability control method for the serpentine gait to minimize sensor interference.
• Proposed a tightly coupled visual-inertial SLAM approach integrating serpentine and Arc-Rolling gaits.

Main Results:

• The improved serpentine gait and head stability control reduced motion-induced sensor interference.
• The visual-inertial SLAM method demonstrated enhanced robustness and accuracy for snake robots.
• Experimental validation confirmed significant improvements in positioning and map-building capabilities.

Conclusions:

• The proposed gait improvement and tightly coupled visual-inertial SLAM effectively enhance snake robot autonomous navigation.
• These methods mitigate motion artifacts, leading to more reliable and precise localization and mapping.
• The findings contribute to advancing the field applicability of snake robots.