Robust Direction-of-Arrival Estimation Using Zero-Crossing-Based Time Delay Measurement for Navigation in GNSS-Denied...

Area of Science:

• Navigation systems
• Signal processing
• Geophysics

Background:

• Accurate Direction-of-Arrival (DOA) estimation is crucial for navigation systems.
• Long-Range Navigation-C (Loran-C) signals offer a potential alternative or supplement to Global Navigation Satellite Systems (GNSS).
• Ultra-Short Baseline (USBL) arrays provide a compact solution for DOA estimation.

Purpose of the Study:

• To investigate and compare DOA estimation techniques for Loran-C signals using a USBL array.
• To evaluate the performance of Correlation-based Least-Squares (Corr-LS), Zero-Crossing-based Least Squares (ZC-LS), and Multiple Signal Classification (MUSIC) algorithms.
• To determine the most suitable algorithm for compact Loran-C receivers in GNSS-challenged environments.

Main Methods:

• Implementation of two least-squares DOA estimation methods: Corr-LS and ZC-LS, based on inter-element delay measurements within the USBL array.
• Application of the Multiple Signal Classification (MUSIC) algorithm as a covariance-based DOA estimator for comparison.
• Experimental validation using Loran-C signals from three distinct transmission stations, collecting 100 valid pulse groups per station.

Main Results:

• Corr-LS exhibited the highest variance in DOA estimates, particularly with interference.
• ZC-LS demonstrated reduced variance and bias by leveraging the Loran-C waveform's zero-crossing structure.
• MUSIC provided the most concentrated azimuth estimates but incurred higher computational costs and required a well-conditioned covariance matrix.

Conclusions:

• ZC-LS offers a superior balance of angular accuracy, robustness against interference, and real-time feasibility compared to Corr-LS and MUSIC.
• The ZC-LS method is well-suited for compact Loran-C receivers.
• ZC-LS can serve as a valuable complementary navigation technology in environments with limited GNSS availability.