Secure 3D Drone Localization using Active Watermarked Ranging and Adaptive Swarm Optimization Against AI-Enabled Anchor Spoofing

Consumer drones and smart IoT devices are increasingly deployed in applications such as autonomous delivery, surveillance, environmental monitoring, indoor asset tracking, and smart-home services. Reliable localization is essential for autonomous navigation, geofencing, return-to-home operation, context-aware services, and safety-critical positioning in these resource-constrained connected devices. However, localization accuracy can be severely degraded by adversarial interference, particularly AI-enabled distance-manipulation and spoofing attacks. Traditional swarm-based optimizers generally perform well under benign conditions but may exhibit significant performance degradation when ranging measurements are intentionally manipulated. This paper proposes a secure localization framework that integrates the Adaptive Superb Fairy-wren Optimization Algorithm (ASFWO) with an Active Watermarked Ranging (AWR) trust mechanism. The proposed ASFWO-AWR framework is designed to enhance attack-resilient localization for consumer autonomous drones, smart-home and edge-IoT devices, indoor location-aware electronics, and delivery or inspection platforms. AWR employs dual-power RSSI watermarking and time-of-arrival dithering to detect anchor-level spoofing and authenticate ranging signals. The resulting trust weights are dynamically incorporated into an adaptive Iteratively Reweighted Least Squares (IRLS)-based position estimator, forming a robust hybrid optimization framework capable of maintaining stable convergence even when a subset of anchors is compromised. Experimental results demonstrate that ASFWO-AWR attains an RMSE of 2.07 m under the low-noise condition and provides improved localization accuracy and robustness compared with conventional PSO and SFOA under challenging measurement conditions.