BDS-3 Broadcast Ephemeris Orbit Correction Model Based on Improved PSO Combined with BP Neural Network

Area of Science:

• Satellite navigation systems
• Computational intelligence
• Orbit determination

Background:

• Navigation satellite operations are susceptible to broadcast ephemeris orbit errors from various ingress factors.
• Accurate satellite positioning relies heavily on precise orbit determination and ephemeris data.

Purpose of the Study:

• To develop an optimized computational intelligence model for correcting navigation satellite track errors.
• To enhance the accuracy of satellite orbit determination using an improved neural network approach.

Main Methods:

• Utilized an improved particle swarm optimization (IPSO) algorithm to optimize hyperparameters of a BP neural network (BPNN).
• Trained the IPSO-BPNN model using broadcast ephemeris Keplerian root numbers and regression parameters.
• Validated the model with BeiDou-3 (BDS-3) Medium Earth Orbit (MEO) and Inclined Geosynchronous Orbit (IGSO) satellite data.

Main Results:

• The IPSO-BPNN model achieved error correction rates ranging from 70.2% to 84% for BDS-3 satellites.
• Demonstrated a significant improvement in error correction rate, increasing by 14.2% to 56.8% compared to the standard PSO-BPNN model.
• The enhanced optimization of inertia weights and learning factors in PSO improved global optimization and convergence speed.

Conclusions:

• The proposed IPSO-BPNN algorithm effectively corrects navigation satellite orbit errors, enhancing ephemeris accuracy.
• This method offers a valuable reference for improving satellite orbit determination in BDS-3 and other global navigation systems.
• Computational intelligence provides a robust framework for addressing challenges in satellite navigation accuracy.