Search

Unmanned aerial vehicles (UAVs) have been widely used in military and civilian fields by carrying out intelligent applications with deep learning technologies, such as battlefield search and rescue; intelligence, surveillance, and reconnaissance (ISR); and smart city monitoring. Due to deeper and more complex models of deep neural networks (DNNs) and limited on-board resources in UAVs, most existing works either rely on cloud-based intelligence by transmitting original data to remote powerful cloud servers and running DNNs inference therein, or involve edge intelligence by executing lightweight DNN models on-board. Unfortunately, the former is faced with unacceptable long transmission latency and possible network instability over air-to-ground links, while the latter is restricted to relatively low accuracy via lightweight models. Although a few recent works propose to collaboratively run complex DNNs inference within a swarm of UAVs to achieve both high accuracy and low latency, they seldom consider the strong confrontation environment of the battle-field, the unreliable air-to-air links among UAVs as well as likely hardware/software breakdowns in UAV airborne computers, which could result in the failure of collaborative inference. This article proposes a novel architecture based on an observation, orientation, decision, and action (OODA) loop, called elastic collaborative edge intelligence (eCoEI), which could fully utilize the resources of computation and storage in a UAV swarm, and enable the swarm to collaboratively perform DNN inference tasks in an elastic manner in the face of any lost links. This article also designs a prototype system of the proposed eCoEI and conducts a proof-of-concept evaluation to validate its feasibility as well as effectiveness. Experimental results show that the eCoEI has an outstanding ability to deal with single point of failure and network fluctuations within a UAV swarm for collaborative edge intelligence. Finally, the main technical challenges are pointed out in the eCoEI, and an outlook on future research directions is provided.

With the advantages of high-speed parallel processing, quantum computers can efficiently solve large-scale complex optimization problems in future networks. However, due to the uncertain qubit fidelity and quantum channel noise, distributed quantum computing, which relies on quantum networks connected through entanglement, faces many challenges in exchanging information across quantum computers. In this article, we propose an adaptive distributed quantum computing approach, called DQC2O, to manage quantum computers and quantum networks for solving optimization tasks in future networks. Firstly, we describe the fundamentals of quantum computing and its distributed concept in quantum networks. Secondly, to address the uncertainty of future demands of collaborative optimization tasks and instability over quantum networks, we propose a quantum resource allocation scheme based on stochastic programming for minimizing quantum resource consumption. Finally, based on the proposed approach, we discuss the potential military applications of collaborative optimization in future networks, such as smart grid management, IoT cooperation, and semantic communications. Promising research directions that can lead to the design and implementation of future distributed quantum computing frameworks are also highlighted.