Deep Learning of GNSS Acquisition.

Area of Science:

• Electrical Engineering
• Computer Science
• Signal Processing

Background:

• Global Navigation Satellite System (GNSS) receivers rely on precise signal acquisition.
• Traditional acquisition methods often involve maximizing the Cross-Ambiguity Function (CAF) as a hypothesis testing problem.
• Existing methods can be computationally intensive and may have limitations in performance.

Purpose of the Study:

• To propose a novel deep learning-based method for GNSS signal acquisition.
• To enhance the accuracy and computational efficiency of the signal acquisition process.
• To demonstrate the superiority of the data-driven approach over conventional CAF maximization strategies.

Main Methods:

• Utilizing deep learning models to classify the Cross-Ambiguity Function (CAF).
• Employing Bayesian hypothesis testing on classifier-derived posteriors for signal presence detection.
• Implementing a parallel classifier bank with optimal fusion for computational affordability and versatility.
• Enabling noncoherent integration schemes through data fusion to boost classifier accuracy.

Main Results:

• The proposed deep learning method effectively performs GNSS signal acquisition.
• The parallel processing and data fusion approach enhances computational efficiency and versatility.
• Simulation results indicate superior performance compared to current CAF maximization techniques.
• The method achieves enhanced acquisition at medium-to-high carrier-to-noise density ratios.

Conclusions:

• Deep learning offers a powerful and efficient alternative for GNSS signal acquisition.
• The proposed data-driven method provides enhanced performance and robustness.
• This approach has the potential to significantly improve GNSS receiver capabilities.