Attitude and heading measurement based on adaptive complementary Kalman filter for PS/MIMU integrated system

Area of Science:

• Robotics and Autonomous Systems
• Sensor Fusion and Navigation

Background:

• Unmanned vehicles require reliable attitude and heading information, especially when Global Navigation Satellite System (GNSS) is unavailable.
• Existing bionic polarization sensor (PS)/MEMS inertial measurement unit (MIMU) systems struggle in harsh environments (inclining, sheltering) due to underutilization of sensor complementarity.
• Current methods limit system performance by not fully leveraging the synergistic characteristics of gyroscopes, accelerometers, and PS.

Purpose of the Study:

• To propose an improved attitude and heading measurement method for unmanned vehicles.
• To enhance system adaptability and accuracy in challenging environmental conditions.
• To fully exploit the complementary nature of gyroscopes, accelerometers, and polarization sensors.

Main Methods:

• Development of an adaptive complementary Kalman filter (ACKF) for attitude and heading estimation.
• Correction of gyroscope data using accelerometer-measured gravity for improved attitude accuracy.
• Fusion of IMU heading and tilt-compensated polarization heading via Kalman optimal estimation.
• Construction of an adaptive factor using maximum correlation entropy between measured and theoretical gravity for adaptive sensor complementarity.

Main Results:

• The proposed ACKF method demonstrated effectiveness in both outdoor rotation and vehicle occlusion tests.
• Significant reductions in Root Mean Square Error (RMSE) were observed for pitch (89.3%), roll (93.2%), and heading (9.6%) compared to traditional Kalman filters during vehicle testing.
• The method shows great advantages in improving the accuracy and reliability of attitude and heading information.

Conclusions:

• The adaptive complementary Kalman filter (ACKF) effectively addresses the limitations of existing methods for attitude and heading determination in GNSS-denied environments.
• The ACKF enhances system performance by adaptively fusing data from PS, gyroscopes, and accelerometers, improving robustness in harsh conditions.
• This approach offers a significant advancement for navigation systems in unmanned vehicles, particularly under challenging operational scenarios.