Artificial intelligence and the Internet of Things (IoT) have brought great convenience to people’s everyday lives. With the emergence of edge computing, IoT devices such as unmanned aerial vehicles (UAVs) can process data instantly at the point of generation, which significantly decreases the requirement for on-board processing power and minimises the data transfer time to enable real-time applications. Meanwhile, with federated learning (FL), UAVs can enhance their intelligent decision-making capabilities by learning from other UAVs without directly accessing their data. This facilitates rapid model iteration and improvement while safeguarding data privacy. However, in many UAV applications such as UAV logistics, different UAVs may perform different tasks and cover different areas, which can result in heterogeneous data and add to the problem of non-independent and identically distributed (Non-IID) data for model training. To address such a problem, we introduce a novel cloud–edge–end collaborative FL framework, which organises and combines local clients through clustering and aggregation. By employing the cosine similarity, we identified and integrated the most appropriate local model into the global model, which can effectively address the issue of Non-IID data in UAV logistics. The experimental results showed that our approach outperformed traditional FL algorithms on two real-world datasets, CIFAR-10 and MNIST.

In recent years, edge-based smart unmanned aerial vehicle (UAV) delivery systems have attracted a lot of attention by both academia and industry given its promising business value and also as an ideal testbed for many emerging technologies such as edge computing, blockchain and machine learning. At the moment, one of the critical challenges for smart UAV delivery systems is data privacy since the massive amount of data is being generated by both users and UAVs and the data is used for training machine learning models to support various smart applications such as autonomous navigation, facial recognition, and person re-identification (ReID). To tackle such a challenge, federated learning (FL) has been widely used as a promising solution since it only needs to share and update model parameters with the centralised server without transmitting the raw data. However, conventional FL still faces the issue of the single-point-of-failure. To address these issues, in this paper, we propose BDFL, a Blockchain-enabled decentralised FL framework for edge-based smart UAV delivery systems. In our framework, Blockchain provides a decentralised network for FL to eliminate the need for a centralised server and store private data in the decentralised permissioned Blockchain to avoid the single-point-of-failure. To motivate our study and analyse the privacy concerns, we employ the person ReID application in smart UAV delivery systems as a typical example. In addition, we also provide the customised proof of quality factor (cPoQF) consensus protocol to address the scalability issue of the Blockchain in order to support the increasing number of smart applications in UAV delivery system. The effectiveness of the framework is demonstrated through experiments on energy efficiency, confirmation time and throughput, with further discussion on the impact of the incentive mechanism, and the analysis of its resiliency under various security attacks.