Researchers have achieved a major breakthrough in increasing the precision of Global Navigation Satellite Systems (GNSS) in cities by developing a novel technique to detect non-line-of-sight (NLOS) errors. This progress carries significant potential for various applications, including self-driving cars and smart city infrastructure.

The Problem of NLOS Errors

In city environments, GNSS signals frequently face obstructions like tall buildings and vehicles, causing NLOS errors. These errors lead to inaccurate positioning, posing challenges for technology that depends on exact navigation 1.

AI-Driven Solution

Addressing this, researchers from Wuhan University, Southeast University, and Baidu developed an AI-based method using the Light Gradient Boosting Machine (LightGBM) model 2. This approach evaluates multiple GNSS signal features to effectively determine and separate NLOS errors, greatly enhancing the accuracy and reliability of GNSS positioning system.

Core Features of the Method

1. Fisheye Camera: The technique uses a fisheye camera to classify GNSS signals as Line-of-Sight (LOS) or NLOS based on satellite visibility3.

• Signal Feature Analysis: The researchers analyzed various signal features, including: Signal-to-noise ratio
• Elevation angle
• Pseudo range consistency
• Phase consistency 1

1. Machine Learning Model: The LightGBM model discovers correlations between these features and signal types, reaching an impressive 92% precision in distinguishing LOS from NLOS signals.
2. Performance: Compared to traditional methods like XGBoost, this method showed better accuracy and computational efficiency. Read more here.

Practical Experiments

The model’s effectiveness was confirmed through dynamic real-world testing in Wuhan, China, proving its capability in complex urban settings.

Impact and Application

• This research has vital implications for industries relying on GNSS technology, including: Autonomous vehicles
• Drones
• Urban planning
• Smart city infrastructure 3

By improving the recognition and exclusion of NLOS errors, this method enhances GNSS system accuracy, making navigation safer and more effective in heavily populated cities 2. The lead researcher, Dr. Xiaohong Zhang, highlighted the importance of this development, describing it as a major leap in enhancing GNSS accuracy in urban spaces, with profound applications such as driverless cars and smart city infrastructure 13. As cities get more connected and advanced technologically, this breakthrough is set to significantly bolster the next generation of transport and navigation tech, promoting smarter and more efficient urban development.