A Post-Processing Multipath/NLoS Bias Estimation Method Based on DBSCAN

Area of Science:

• * Satellite Navigation Systems
• * Signal Processing
• * Geomatics Engineering

Background:

• * Urban environments present significant challenges for Global Navigation Satellite Systems (GNSS) due to multipath and Non-Line-of-Sight (NLoS) effects.
• * These effects introduce pseudorange biases that degrade the accuracy of GNSS positioning applications.
• * Current methods for identifying and classifying multipath/NLoS events remain challenging.

Purpose of the Study:

• * To propose a post-processing method for estimating pseudorange biases caused by multipath/NLoS effects.
• * To provide accurate pseudorange bias data for training machine learning models for multipath/NLoS detection and mitigation.
• * To establish a benchmark for evaluating new methods designed to detect multipath/NLoS effects.

Main Methods:

• * Extraction of multipath/NLoS biases from pseudorange measurements using the Density-Based Spatial Clustering of Applications with Noise (DBSCAN) algorithm.
• * Validation of the method using real-world data collections in static and dynamic urban multipath/NLoS scenarios.
• * Performance evaluation based on positioning accuracy by comparing solutions derived from bias-corrected pseudoranges against ground truth.

Main Results:

• * Successful estimation of pseudorange biases attributed to multipath/NLoS conditions.
• * Demonstrated effectiveness of the proposed method in both static and dynamic scenarios.
• * Validation through improved positioning performance when estimated biases are applied.

Conclusions:

• * The proposed DBSCAN-based method effectively estimates pseudorange biases in urban GNSS environments.
• * The estimated biases can significantly enhance the training of machine learning algorithms for improved GNSS positioning.
• * This approach provides a reliable validation strategy for multipath/NLoS detection techniques.