With the growing application of Global Navigation Satellite Systems (GNSS), ensuring the security of these systems has become a critical priority. Among the various threats targeting GNSS, spoofing attacks pose the greatest danger, as they exploit the inability of GPS receivers to distinguish between authentic and spoofed signals. This vulnerability can lead to incorrect navigation equation solutions, ultimately resulting in erroneous position estimations. Due to the lack of a universal method for detecting spoofing attacks, various approaches have been proposed, tailored to specific requirements. One prominent category involves the use of unsupervised machine learning clustering algorithms. While density-based clustering algorithms have demonstrated promising performance, their effectiveness is often limited by a strong dependence on input parameters. In this study, spoofing signals are detected using the HDBSCAN algorithm, a hierarchical density-based clustering approach. The algorithm distinguishes spoofing signals by analyzing signal phase, cross-correlation norm, and the power differential between spoofed and genuine signals. The algorithm's performance is assessed using the Silhouette index, Dunn index, and Confusion Matrix metrics. Field tests validated the proposed algorithm, which was implemented on a software-defined receiver. The results demonstrated a spoofing signal detection accuracy of 98.43%.

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -