Decentralized snapshot selection via trust game for robust DOA estimation with FMCW radar in static scenarios

Direction Of Arrival (DOA) estimation in Frequency Modulated Continuous Wave (FMCW) radar systems is critical for structural health monitoring, but is often compromised by phase-corrupted snapshots caused by clock jitter or Phase-Locked Loop (PLL) drift. Existing robust DOA methods suffer from high computational complexity or sensitivity to outliers, limiting their real-time applicability. In this paper, we propose the first decentralized Trust Game (TG) mechanism for snapshot selection, which leverages inter-snapshot coherence to identify and exclude anomalies with provable complexity bounds. Combined with a Toeplitz-accelerated reiterative super-resolution estimator that exploits uniform linear array structure for near-linear computational cost, our approach achieves robust and high-resolution DOA estimation. Throughout this work, the anomalous snapshots are not caused by target motion but by hardware-induced data corruption (e.g., clock jitter and PLL drift) in an otherwise static scene. Experimental results demonstrate that the proposed method outperforms state-of-the-art techniques in both outlier robustness and execution speed, offering a practical solution for real-time monitoring under hardware constraints.