Direction of Arrival Estimation Based on Piecewise Likelihood Accumulation for GNSS Signals Under Time-Varying Jamming Directions

This paper presents a Piecewise Likelihood Accumulation Estimation (PLAE) algorithm to address the challenge of direction-of-arrival (DOA) estimation for GNSS signals under time-varying jamming directions. When jamming directions vary over time, conventional GNSS DOA estimation methods-typically based on a single covariance matrix over the entire signal duration-suffer from a jamming averaging effect, which misinterprets time-varying jammers as a larger number of simultaneously present sources. This leads to an overestimation of the number of jammers, resulting in degraded DOA performance, especially when jamming DOAs are close to those of GNSS signals or when the array experiences degrees-of-freedom (DOF) deficiency. The proposed PLAE algorithm mitigates these limitations by modeling the received signal as a piecewise stationary process. It divides the pre-correlation samples into short intervals, estimates a local covariance matrix for each segment, and accumulates the likelihood functions across segments using maximum likelihood estimation (MLE). This strategy allows PLAE to adapt to the time-varying characteristics of jamming, yielding DOA estimation accuracy close to the Cramér-Rao Lower Bound (CRLB). Simulation results demonstrate the robustness and superior performance of PLAE compared to existing approaches, making it a reliable solution for GNSS applications in complex jamming scenarios with varying jamming directions.