Enhanced Autonomous Vehicle Positioning Using a Loosely Coupled INS/GNSS-Based Invariant-EKF Integration

Area of Science:

• Robotics and Autonomous Systems
• Navigation and Positioning
• Signal Processing

Background:

• High-precision navigation is crucial for autonomous vehicles (AVs).
• Global Navigation Satellite Systems (GNSS) offer primary navigation but suffer signal degradation in certain environments.
• Inertial Navigation Systems (INS) provide continuous navigation but drift over time.

Purpose of the Study:

• To introduce and validate a loosely coupled INS/GNSS integration scheme using the invariant extended Kalman filter (IEKF).
• To address the limitations of traditional Extended Kalman Filter (EKF) based INS/GNSS integration, particularly in GNSS-deprived scenarios.

Main Methods:

• Implemented a loosely coupled INS/GNSS integration framework.
• Utilized the invariant extended Kalman filter (IEKF), leveraging matrix Lie group for state estimation.
• Tested the proposed system on a real-world road trajectory.

Main Results:

• The IEKF-based INS/GNSS integration demonstrated significant performance enhancement over traditional EKF methods.
• Achieved an 82.98% improvement in overall trajectory 2D-position Root Mean Square (RMS) error (19.4m to 3.3m).
• Reduced the 2D-position maximum error by 80.78% (73.9m to 14.2m).

Conclusions:

• The IEKF-based INS/GNSS integration offers superior navigation accuracy and robustness compared to EKF methods.
• This approach is effective in both GNSS signal presence and blockage conditions, crucial for AV reliability.