GNSS NLOS Signal Classification Based on Machine Learning and Pseudorange Residual Check

Area of Science:

• Geomatics Engineering
• Satellite Navigation Systems
• Machine Learning Applications

Background:

• Global Navigation Satellite System (GNSS) positioning is crucial for autonomous driving and construction.
• Dense urban environments present challenges due to signal obstruction and multipath effects.
• Distinguishing between line-of-sight (LOS) and non-line-of-sight (NLOS) signals is vital for accuracy.

Purpose of the Study:

• To develop and evaluate a novel method for detecting NLOS signals in GNSS positioning.
• To improve the reliability and accuracy of GNSS positioning in challenging urban settings.
• To leverage machine learning for enhanced multipath signal identification.

Main Methods:

• Utilized a support vector machine (SVM) classifier for NLOS signal detection.
• Developed unique features from receiver independent exchange format data and GNSS pseudorange residuals.
• Conducted static tests in downtown Tokyo with high-rise buildings to simulate urban conditions.

Main Results:

• The proposed SVM classifier combined with GNSS pseudorange residual checks effectively reduced positioning errors caused by NLOS signals.
• Horizontal positioning errors within 10 meters were improved by over 80% in static tests.
• The method demonstrated significant error reduction by identifying and excluding satellites with NLOS signals.

Conclusions:

• The SVM-based NLOS detection method offers a robust solution for improving GNSS accuracy in urban environments.
• Accurate NLOS signal identification is key to overcoming multipath interference in satellite navigation.
• This approach enhances the viability of GNSS for critical applications like autonomous systems.