Machine learning improves GNSS signal stability

Ionospheric scintillation, caused by irregularities in the Earth's ionosphere, can severely affect the integrity of Global Navigation Satellite System (GNSS) signals and cause navigation errors.

Traditional detection methods rely on expensive and specialized Ionospheric Scintillation Monitoring Receivers (ISMRs), but with increasing reliance on GNSS for various applications, there is an urgent need for more accessible and cost-effective detection methods.

Due to these challenges, detailed research is needed to detect ionospheric scintillation events utilizing common GNSS receivers.

The new study, led by a team of researchers from the Hong Kong Polytechnic University, has been published in the journal Satellite navigation Published on June 3, 2024, the research team introduces a new strategy to identify ionospheric amplitude scintillation events with very high accuracy using common geodetic GNSS receivers, potentially transforming GNSS monitoring.

This work focuses on utilizing a vast network of geodetic GNSS receivers to detect ionospheric scintillation events that are usually identified by specialized ISMRs. The proposed method is based on the Carrier to Noise Ratio (C/N₀) and elevation angle data are collected at 1 Hz intervals.

By mitigating the multipath effect through detailed analysis of the multipath pattern, this study effectively reduces noise and false alarms and ensures the accuracy of scintillation detection. In this methodology, a surrogate scintillation index (S_4c) Based on C/N₀ Measurements from a geodetic GNSS receiver.

This index shows a high correlation with the traditional S._Four Although geodetic receivers are susceptible to noise and multipath interference, the index used in ISMR exploits the periodicity of multipath effects that differ from scintillation irregularities to improve detection accuracy. Machine learning algorithms exploit the periodicity of multipath effects that differ from scintillation irregularities to improve detection accuracy.

Experimental results demonstrate that the decision tree algorithm achieves an astounding 99.9% detection accuracy, outperforming traditional hard and semi-hard thresholding methods.

“Our research demonstrates the potential of integrating machine learning with widely available GNSS receivers to revolutionize ionospheric scintillation detection, which is not only a cost-effective alternative to specialized equipment but also improves the accuracy and reliability of space weather monitoring,” said lead researcher Dr. Yiping Jiang.

The impact of this research is far-reaching, providing a scalable solution for GNSS users worldwide: improving the detection of scintillation events will contribute to the development of more accurate navigation algorithms and techniques.

This advancement is crucial for a variety of applications, including aviation, maritime and land transportation, where GNSS reliability is paramount.

Provided by TranSpread