Multi-source positioning information fusion method based on improved robust Kalman filter

Area of Science:

• Geomatics Engineering
• Robotics
• Signal Processing

Background:

• Accurate positioning is crucial for rolling machinery in construction.
• Deep and narrow valleys present significant challenges due to noise interference.

Purpose of the Study:

• To develop a multi-source positioning fusion method to enhance accuracy in harsh environments.
• To improve the robustness and reliability of real-time monitoring data.

Main Methods:

• An improved robust Kalman filter incorporating a thick tail Laplace distribution.
• Adaptive selection of optimal observations from Global Navigation Satellite System (GNSS), Robotic Total Station (RTS), and Ultra Wide Band (UWB).
• Dynamic adjustment of noise covariance to handle large random errors.

Main Results:

• The method effectively compensates for data offset and loss (over 97.33%).
• Localization offset rates were reduced by 7.72%, and loss rates by 1.64% compared to single-method approaches.
• Demonstrated adaptability in complex, harsh environments like deep river valleys.

Conclusions:

• The proposed fusion method significantly enhances the robustness, accuracy, and completeness of real-time monitoring.
• The improved Kalman filter effectively mitigates noise interference and improves data fusion in challenging terrains.
• This approach offers a reliable solution for precise positioning of rolling machinery in difficult construction settings.