AMLB-FL: A Privacy-Preserving Blockchain and Federated Learning Architecture for UAV Networks

There are critical security concern along with issues pertaining to privacy and scalability towards 
the open-air mode exchange of data in Unmanned Aerial Vehicle (UAV) networks. The current solutions 
are noted to adopt centralized security measure, isolated blockchain, and even federated learning and yet 
they cannot mitigate issues pertaining to unreliable fusion of model, communication overhead, and single
point failures. These problems are addressed in proposed study model named as Adaptive Multi-Layer 
Blockchain and Federated Learning (AMLB-FL) meant for securing data exchange in UAV networks. The 
framework is structured with three layers meant for trust management, blockchain operation, and 
anomaly detection with privacy preservation. Experimental outcome shows proposed model to offer 
notably higher detection accuracy, lower latency, increased scalability, and increased energy efficiency in 
contrast to most relevant baseline models.