Landmark-Based Scale Estimation and Correction of Visual Inertial Odometry for VTOL UAVs in a GPS-Denied Environment

Area of Science:

• Robotics
• Computer Vision
• Navigation Systems

Background:

• Unmanned Aerial Vehicles (UAVs) are increasingly used in various applications.
• Global Positioning System (GPS) navigation is unreliable in environments like indoors or during bridge inspections.
• Visual-Inertial Odometry (VIO) offers a solution for GPS-denied navigation but suffers from scale errors and long-term drift.

Purpose of the Study:

• To propose and validate a novel method for correcting VIO position errors in VTOL UAVs without GPS.
• To enhance the accuracy and reliability of VIO for UAV navigation in GPS-unavailable environments.

Main Methods:

• Utilizing artificial landmarks and their known information to improve initial VIO positioning.
• Employing an Extended Kalman Filter (EKF) with landmark-based measurements for scale correction via the least squares method.
• Integrating Inertial Measurement Unit (IMU) data for time-update processes and using both visual odometry (VO) and scale-corrected VIO for EKF updates.

Main Results:

• The proposed method effectively corrects scale errors and mitigates long-term drift in VIO.
• Experimental validation using a custom UAV and high-precision RTK demonstrated the method's efficacy.
• The trajectory estimated by landmarks during takeoff and landing phases was used to update scale correction.

Conclusions:

• The developed GPS-free VIO error correction method significantly improves UAV positioning accuracy.
• This approach provides a robust solution for VIO navigation challenges in GPS-denied scenarios.
• The findings contribute to the advancement of autonomous navigation for UAVs in complex environments.