PPIoDT : GSO-FL based privacy preserving IoDT guided ocean-wind aware ship trajectory recommendation

Abstract

Recently, the Federated Learning (FL) has gained the utmost popularity due to its privacy-preserving intelligent decentralized frame-work as an alternative to traditional centralized machine learning approaches. The wind disturbances especially in the ocean highly affect the ship trajectories. A drone-guided ship can mitigate wind effects by predicting local weather data without sharing it with the central server in the ocean environment. The drone flights often have privacy-sensitive data that needs a robust framework, especially in disaster, rescue, surveil-lance, or other various welfare applications. In our work, we have applied the nature-inspired Glowworm Swarm Optimization (GSO) for privacy-preserving Internet of Drone Things (IoDT). We proposed a GSO-FL-based privacy-preserving IoDT-guided ocean-wind aware ship trajectory recommendation model to mitigate real-time wind gust disturbances. The outcome shows the convergence time of the glowworms reduces approximately 17% as compared to other GSO-based models.