Multi-Sensor Information Fusion Positioning of AUKF Maglev Trains Based on Self-Corrected Weighting

Area of Science:

• Transportation Engineering
• Control Systems
• Signal Processing

Background:

• Accurate positioning is critical for maglev train safety and scheduling.
• Existing methods suffer from noise interference, single-sensor reliance, and historical data bias.
• Traditional weighting schemes are susceptible to historical data, leading to positioning deviations.

Purpose of the Study:

• To propose an improved multi-sensor information fusion and positioning method for maglev trains.
• To address limitations of single-sensor methods and noise interference in maglev positioning.
• To enhance the accuracy and reliability of maglev train positioning systems.

Main Methods:

• Developed an AUKF (Adaptive Unscented Kalman Filter) algorithm integrating Sage-Husa noise estimation and self-corrected weighting.
• Utilized multi-sensor data fusion from cross-sensor lines, Inertial Navigation Systems (INS), Doppler radar, and Global Navigation Satellite Systems (GNSS).
• Implemented adaptive statistical feature estimation for measurement noise to overcome single-function limitations.

Main Results:

• The proposed self-correction-based AUKF algorithm demonstrated trajectories closer to real values.
• Achieved reduced Mean Error (ME) and Root Mean Square Error (RMSE) compared to traditional methods.
• The AUKF algorithm effectively eliminated single-function and low-integration shortcomings of individual modules.

Conclusions:

• The self-correction-weighted AUKF algorithm offers significant advantages in stability, accuracy, and simplicity for maglev train positioning.
• This multi-sensor fusion approach enhances precise positioning capabilities for maglev trains.
• The method provides a robust solution for overcoming noise and data integration challenges in maglev systems.