A Novel Fault-Tolerant Navigation and Positioning Method with Stereo-Camera/Micro Electro Mechanical Systems Inertial...

Area of Science:

• Robotics
• Computer Vision
• Navigation Systems

Background:

• Visual odometry (VO) estimates vehicle motion from images but struggles in adverse conditions like low features, rapid motion, or lighting changes.
• Enhancing VO robustness in hostile environments is a critical research area for reliable navigation.

Purpose of the Study:

• To develop a novel fault-tolerant visual-inertial odometry (VIO) framework for robust navigation and positioning.
• To improve pose estimation accuracy and reliability in challenging environmental conditions.

Main Methods:

• Integration of micro electro mechanical systems inertial measurement unit (MEMS-IMU) with stereo-cameras for sensor fusion.
• Implementation of MEMS-IMU pre-integration to enhance motion estimation with sparse features.
• Introduction of a dramatic change detector and adaptive observation noise factor to mitigate errors from large angular motion and incorrect feature matching.

Main Results:

• The proposed VIO method demonstrates superior position estimation compared to traditional VO and VIO approaches in experiments.
• The system effectively handles environments with limited features, fast angular motions, and illumination variations.

Conclusions:

• The presented fault-tolerant VIO framework significantly enhances navigation robustness in hostile environments.
• The fusion of MEMS-IMU data and advanced algorithms provides a more reliable solution for vehicle ego-motion estimation.