The past few years have witnessed unprecedented proliferation of Unmanned Aerial Vehicles (UAVs). They are employed in a growing number of scenarios, from parcel delivery to search and rescue operations, requiring coordinated missions of a fleet of drones. Recently, there has been growing interest in optimized techniques to assign tasks and related trajectories to drones. While these techniques promise high coverage of inspected area, their applicability in real scenarios is precluded by unconsidered constraints. Among these, the limited amount of power of UAVs, and the consequent need of performing multiple trips to provide complete monitoring coverage, with battery replacement/charging and data offloading in between. To address this problem we develop Gen-Path, a genetic algorithm for efficient scheduling of multi-round UAV missions, under several objective functions. By means of simulations we show that Gen-Path fits various scenarios, improving existing solutions in terms of covered points, and energetic cost.

Cooperative UAV networks are becoming increasingly popular in military and civilian applications. Alas, the typical ad-hoc routing protocols, which aim at finding the shortest path, lead to significant performance degradation because of the 3-dimension highly-dynamic nature of UAV networks and the uneven distribution of nodes across the network. This paper proposes OPAR, an optimized predictive and adaptive routing protocol, to face this challenging problem. We model the routing problem with linear programming (LP), where the goal is to maximize network performance, considering the path lifetime and path-length together. This model relies on a precise link lifetime prediction mechanism. We support the LP problem with a lightweight algorithm to find the optimized solution with a computation complexity of O(|E| 2 ), where |E| is the number of network links. We evaluate the OPAR performance and compare it with the well-known routing algorithms AODV, DSDV, and OLSR to cover a wide range of proactive and reactive protocols as well as distance vector and link-state techniques. We performed extensive simulations for different network densities and mobility patterns using the ns-3 simulator. Results show that OPAR prevents a high volume of routing traffic, increases the successful delivery by more than 30%, improves the throughput 25% on average, and decreases the flow completion time by an average of 35% 1 .

Due to the on-demand deployable and flexible-observation features, unmanned aerial vehicles (UAVs) is considered as one of the key enabling techniques in next-generation sensing systems. In many sensing applications (e.g., disaster rescue and environment monitoring), a single UAV is not sufficient to fulfill the requirements of transmission delay and sensing quality. Considering this, a multi-UAV collaborative compression and relaying scheme is proposed in this work. The sensory data is packed according to the sensing quality requirement and the sensory data packet rate can be maximized to improve the freshness of the sensory data. To tackle the non-convex problem of finding the optimal compression ratios and UAVs locations with packet rate maximization, the considered problem is converted into a monotonic optimization (MO) by deriving the closed-form expression of the optimal compression ratio for given UAV locations. Then, by exploiting the inherent structure of the optimal UAV locations, a novel region-elimination-based fast location search algorithm is proposed, which can effectively avoid unnecessary search and achieve a substantially faster convergence as compared to the standard MO algorithm. Besides, numerical simulations are conducted to validate the effectiveness of the proposed scheme.

Radio-frequency (RF) sensing is a key technology for designing intelligent and secure wireless networks with high spectral efficiency and environment-aware adaptation capabilities. However, existing sensing techniques can extract only limited information from RF signals or assume that the RF signals are generated by certain known protocols. As a result, their applications are limited if proprietary protocols or encryption methods are adopted, or in environments subject to errors such as unintended interference. To address this challenge, we study protocol-agnostic cross-layer sensing to extract high-layer protocol information from raw RF samples without any a priori knowledge of the protocols. First, we present a framework for protocol-agnostic sensing for over-the-air (OTA) RF signals, by taking packet boundary recognition (PBR) as an example. The framework consists of three major components: OTA Signal Generator, Agnostic RF Sink, and Ground Truth Generator. Then, we develop a software-defined testbed using USRP SDRs, with eleven benchmark statistical algorithms implemented in the Agnostic RF Sink, including Kullback-Leibler divergence and cross-power spectral density, among others. Finally, we test the effectiveness of these statistical algorithms in PBR on OTA RF samples, considering a wide variety of transmission parameters, including modulation type, transmission distance, and packet length. It is found that none of these benchmark statistical algorithms can achieve consistently high PBR rate, and new algorithms are required particularly in next-generation low-latency wireless systems.

Due to spectrum scarcity and increasing wireless capacity demands, terahertz (THz) communications at 0.1-10THz and the corresponding spectrum characterization have emerged to meet diverse service requirements in future 5G and 6G wireless systems. However, conventional compressed sensing techniques to reconstruct the original wideband spectrum with under-sampled measurements become inefficient as local spectral correlation is deliberately omitted. Recent works extend communication methods with deep learning-based algorithms but lack strong ties to THz channel properties. This paper introduces novel THz channel-aware spectrum learning solutions that fully disclose the uniqueness of THz channels when performing such ultra-broadband sensing in vehicular environments. Specifically, a joint design of spectrum compression and reconstruction is proposed through a structured sensing matrix and two-phase reconstruction based on high spreading loss and molecular absorption at THz frequencies. An end-to-end learning framework, namely compression and reconstruction network (CRNet), is further developed with the mean-square-error loss function to improve sensing accuracy while significantly reducing computational complexity. Numerical results show that the CRNet solutions outperform the latest generative adversarial network (GAN) realization with a much higher cosine and structure similarity measures, smaller learning errors, and 56\% less required training overheads. This THz Ultra-broadband Learning Vehicular Channel-Aware Networking (TULVCAN) work successfully achieves effective THz spectrum learning and hence allows frequency-agile access.

Unmanned aerial vehicles (UAVs), or drones, are emerging as a promising technology for a variety of monitoring and surveillance-based applications. Smart UAVs are not limited only to image capturing, but also to real-time decision making using artificial intelligence. Moreover, it is important to consider the data security of captured images. In this paper, we propose a novel Energy-aware Secure Internet of Drone (ESIoD) architecture. A crucial research problem addressed by this work is how to accomplish faster onboard processing and reduce battery usage for a UAV to prolong the flight time while retaining data security of UAV captured images. Specifically, drone-captured real-time images are encrypted using either AES or RSA algorithms and offloaded by the onboard computer to a cloud server for the processing of cognitive actions using both a standard Haar cascade classifier and an advanced faster RCNN classifier. The focus of this study is to conserve the drone battery life by secure computational offloading to optimize drone flight time. Two sets of experiments were performed using drone-captured sample images and videos. Results show that the ESIoD architecture can conserve 80% onboard processing time and 3X drone battery charge usage as compared to conventional real-time onboard processing for the considered application.

Stochastic geometry has been widely applied to planar networking analysis. With the emergence of unmanned aerial vehicles, it is imperative to extend the SG facility to three-dimensional space. In this paper, an effort is made to combine secrecy analysis with UAV and 3D SG. One of the new features is the relaxation of integer path-loss exponents or their ratio. This is an inherent challenging issue. With the aid of Meijer's G-function, a closed-form expression is derived for the probability of strictly positive secrecy capacity. Several numerical results are presented and discussed.

Motion-controlled robotic systems would become more and more popular in the future since they allow humans to easily control robots to carry out various tasks. However, current authentication methods rely on static credentials, such as passwords, fingerprints, and faces, which are independent of the robot control. Thus, they cannot guarantee that a robot is always under the control of its enrolled user. In this paper, we build a motion-controlled robotic arm system and show that a robotic arm's motion inherits much of its user's behavioral information in interactive control scenarios. Based on that, we propose a novel user authentication approach to verify the robotic arm user. In particular, we log the angle readings of the robotic arm's joints to reconstruct the 3D movement trajectory of its end effector. We then develop a learning-based algorithm to identify the user. Extensive experiments show that our system achieves 95% accuracy to verify users while preventing various impersonation attacks.

This talk does not have an abstract.