Trusted Routing for Blockchain-Enabled Low-Altitude Intelligent Networks

TL;DR

This paper presents a blockchain-enabled, secure, and adaptive routing algorithm for low-altitude intelligent networks with UAVs, reducing end-to-end delay by over 22% through a decentralized deep reinforcement learning approach.

Contribution

It introduces a novel zero-trust blockchain architecture and a multi-agent deep reinforcement learning algorithm for stable, secure, and efficient routing in dynamic UAV networks.

Findings

E2E delay decreased by 22.38% on average.

Proposed method outperforms benchmark algorithms.

Enhanced security and stability in UAV networks.

Abstract

Due to the scalability and portability, the low-altitude intelligent networks (LAINs) are essential in various fields such as surveillance and disaster rescue. However, in LAINs, unmanned aerial vehicles (UAVs) are characterized by the distributed topology and high dynamic mobility, and vulnerable to security threats, which may degrade the routing performance for data transmission. Hence, how to ensure the routing stability and security of LAINs is a challenge. In this paper, we focus on the routing process in LAINs with multiple UAV clusters and propose the blockchain-enabled zero-trust architecture to manage the joining and exiting of UAVs. Furthermore, we formulate the routing problem to minimize the end-to-end (E2E) delay, which is an integer linear programming and intractable to solve. Therefore, considering the distribution of LAINs, we reformulate the routing problem into a decentralized partially observable Markov decision process. With the proposed soft hierarchical experience replay buffer, the multi-agent double deep Q-network based adaptive routing algorithm is designed. Finally, simulations are conducted and numerical results show that the total E2E delay of the proposed mechanism decreases by 22.38\% than the benchmark on average.