Low Complexity Robust Data Demodulation for GNSS.

Area of Science:

• Signal Processing
• Communications Engineering
• Satellite Navigation Systems

Background:

• Direct sequence spread spectrum (DS-SS) systems are crucial for Global Navigation Satellite System (GNSS) applications.
• Accurate log-likelihood ratio (LLR) estimation is vital for reliable data decoding, especially under varying noise conditions.
• Traditional methods often assume known noise variance, potentially causing demodulation errors in GNSS receivers.

Purpose of the Study:

• To develop closed-form approximations for LLR values in DS-SS systems within GNSS environments.
• To address the challenge of noise uncertainty in LLR estimation for improved demodulation.
• To enhance the performance of decoding clock and ephemeris data (CED) encoded with low-density parity-check (LDPC) codes.

Main Methods:

• Derivation of closed-form LLR expressions for additive white Gaussian and Laplacian noise channels under noise uncertainty.
• Development of LLR approximations for noise variance following an inverse log-normal distribution.
• Application and simulation of these LLR expressions in the context of GPS L1C signal decoding using the belief propagation (BP) algorithm.

Main Results:

• The study provides novel closed-form approximations for LLR values applicable to common GNSS scenarios.
• The proposed methods effectively handle noise uncertainty, unlike classical approaches.
• Simulations demonstrate significant frame error rate (FER) improvements when decoding GPS L1C CED with LDPC codes.

Conclusions:

• The derived closed-form LLR approximations offer a robust solution for DS-SS systems in GNSS, particularly under noise uncertainty.
• Accounting for noise variance uncertainty leads to enhanced decoding performance and reduced FER.
• These findings are relevant for improving the reliability and accuracy of satellite navigation systems.