Related Experiment Video
Updated: Jul 2, 2025
06:09
P300-Based Brain-Computer Interface Speller Performance Estimation with Classifier-Based Latency Estimation
Published on: September 8, 2023
578
Improved SSA-Based GRU Neural Network for BDS-3 Satellite Clock Bias Forecasting
Hongjie Liu1, Feng Liu1, Yao Kong2
1College of Computer Science, Xi'an Polytechnic University, Xi'an 710600, China.
Sensors (Basel, Switzerland)
|February 24, 2024
Summary
This study introduces the ITSSA-GRU model for forecasting satellite clock bias in the BeiDou Navigation Satellite System (BDS-3). The novel approach enhances prediction accuracy for high-precision global navigation satellite system (GNSS) positioning.
Area of Science:
- Satellite navigation systems
- Geodesy and geomatics engineering
- Artificial intelligence in navigation
Background:
- Satellite clock errors significantly impact Global Navigation Satellite System (GNSS) positioning accuracy.
- Accurate forecasting of satellite clock bias is crucial for high-precision positioning applications.
- Existing models like Gated Recurrent Unit (GRU) face challenges with hyperparameter sensitivity and local optima.
Purpose of the Study:
- To develop an advanced satellite clock bias forecasting model for the BeiDou Navigation Satellite System (BDS-3).
- To improve the prediction accuracy and stability of existing neural network models.
- To enhance the optimization capabilities for satellite clock bias forecasting.
Main Methods:
- Implementation of a novel Improved Sparrow Search Algorithm (ITSSA) combined with a Gated Recurrent Unit (GRU) neural network (ITSSA-GRU).
- Enhancement of the Sparrow Search Algorithm (SSA) with iterative chaotic mapping for population initialization and t-step optimization for iterative updates.
- Comparative analysis of ITSSA-GRU against GRU, Long Short-Term Memory (LSTM), and GM(1,1) models using BDS-3 satellite clock bias data from MEO, IGSO, and GEO orbits.
Main Results:
- The ITSSA-GRU model demonstrated superior generalization ability and forecasting performance across all tested satellite orbit types (MEO, IGSO, GEO).
- The proposed model significantly outperformed SSA-GRU, GRU, LSTM, and GM(1,1) in predicting satellite clock bias.
- The enhanced optimization strategy within ITSSA effectively addressed GRU's limitations.
Conclusions:
- The ITSSA-GRU model offers a robust and accurate solution for satellite clock bias forecasting in the BDS-3 system.
- This new method provides a valuable tool for enhancing the precision of GNSS positioning.
- The findings highlight the potential of hybrid AI-optimization approaches for improving satellite navigation accuracy.
