Convolutional Neural Network-Based Disaster Assessment Using Unmanned Aerial Vehicles

In Unmanned Aerial Vehicle (UAV) networks, multi-class aerial image classification (AIC) is crucial in various applications, from environmental monitoring to infrastructure inspection. Deep Learning (DL), a powerful tool in artificial intelligence (AI), proves significant in this context, enabling the model to analyze and classify complex aerial images effectually. By utilizing advanced neural network architectures, such as convolutional neural networks (CNN), DL models outperform at identifying complex features and patterns within the aerial imagery. These models can extract spectral and spatial information from the captured data, classifying diverse terrains, structures, and objects precisely. Furthermore, the integration of Snake Optimization algorithms assists in fine-tuning the classification process, improving accuracy. As UAV networks continue to expand, DL-powered multi-class AIC significantly enhances the performance of surveillance, reconnaissance, and remote sensing tasks, contributing to the advancement of autonomous aerial systems. This study proposes a Snake Optimization Algorithm with Deep Learning for Multi-Class Aerial Image Classification (SOADL-MCAIC) methodology on UAV Networks. The main purpose of SOADL-MCAIC methodology is to recognize the presence of multiple classes of aerial images on the UAV networks. To accomplish this, the SOADL-MCAIC technique utilizes Gaussian filtering (GF) for pre-processing. In addition, the SOADL-MCAIC technique employs the Efficient DenseNet model to learn difficult and intrinsic features in the image. The SOA-based hyperparameter tuning process is used to enhance the performance of the Efficient DenseNet technique. Finally, the kernel extreme learning machine (KELM)-based classification algorithm is implemented to identify and classify the presence of various classes in aerial images. The simulation outcomes of the SOADL-MCAIC method are examined under the UCM land use dataset. The experimental analysis of the SOADL-MCAIC method portrayed a superior accuracy value of 99.75% over existing models.

In this paper, cyberattack detection and isolation is studied for a network of formation-flying unmanned aerial vehicles. As the unmanned aerial vehicles communicate to reach consensus on their states while making the formation, the communication network among the unmanned aerial vehicles makes them vulnerable to a potential attack from malicious adversaries. Two types of attacks pertinent to a network of unmanned aerial vehicles have been considered: a node attack on the unmanned aerial vehicles and a deception attack on the communication between the unmanned aerial vehicles. Unmanned aerial vehicle formation control involves using a consensus algorithm to maintain a prespecified formation. Two kinds of cyberattacks in the unmanned aerial vehicle network (node and communication path deception) are considered with their respective models in the formation setup. A bank of unknown input observer-based distributed fault detection schemes as well as with a rule based on residual generated using the bank of unknown input observers are proposed to detect the attacks and to identify the compromised unmanned aerial vehicle in the formation. Furthermore, an algorithm is developed to remove the faulty unmanned aerial vehicle from the network and to keep the compromised unmanned aerial vehicle isolated once an attack is detected, while maintaining the flight formation with a missing unmanned aerial vehicle node.

Non-orthogonal multiple access (NOMA) is indeed an important technique of 5G and beyond 5G (B5G) communication systems, as it can increase the spectrum efficiency and system capacity under the same bandwidth. Long-distance unmanned aerial vehicle (UAV) networks, comprising a base station and multiple UAVs, are commonly used for reconnaissance purposes due to the extended flight range of the UAVs, which can provide broad coverage and satisfy the requirements for real-time communication. To alleviate spectrum underutilization of single channel orthogonal multiple access, we investigate resource allocation in UAV networks by utilizing NOMA and then formulate an optimization problem that maximizes the network capacity. The formulated problem can be decomposed into two sub-problems, namely UAV grouping and power allocation optimization. Accordingly, we present a UAV grouping method with low complexity according to the difference of channel gain between UAVs and base station, in which the UAVs with larger difference will be in the same cluster such that they can transmit signals over the same channel. Then, a power allocation optimization algorithm based on Lagrange multiplier method is then used to optimize the transmit power of UAVs utilizing the same channel. This algorithm considers the quality of service requirements for each UAV, and assigns power levels that maximize the overall network capacity. Simulation results have shown that the presented algorithms improve network capacity by approximately 19% compared with that of orthogonal multiple access, which demonstrates the effectiveness of utilizing NOMA in long-distance UAV networks.

Interference modeling and analysis in 3-dimensional directional UAV networks based on stochastic geometry

Among the many deep learning methods, the convolutional neural network (CNN) model has an excellent performance in image recognition. Research on identifying and classifying image datasets using CNN is ongoing. Animal species recognition and classification with CNN is expected to be helpful for various applications. However, sophisticated feature recognition is essential to classify quasi-species with similar features, such as the quasi-species of parrots that have a high color similarity. The purpose of this study is to develop a vision-based mobile application to classify endangered parrot species using an advanced CNN model based on transfer learning (some parrots have quite similar colors and shapes). We acquired the images in two ways: collecting them directly from the Seoul Grand Park Zoo and crawling them using the Google search. Subsequently, we have built advanced CNN models with transfer learning and trained them using the data. Next, we converted one of the fully trained models into a file for execution on mobile devices and created the Android package files. The accuracy was measured for each of the eight CNN models. The overall accuracy for the camera of the mobile device was 94.125%. For certain species, the accuracy of recognition was 100%, with the required time of only 455 ms. Our approach helps to recognize the species in real time using the camera of the mobile device. Applications will be helpful for the prevention of smuggling of endangered species in the customs clearance area.

As unmanned aerial vehicle (UAV) networks become increasingly critical for diverse applications such as disaster response and commercial services, the demand for situation-aware UAV network operation continues to rise. Traditional UAV network operations are typically context- unaware, resulting in limited adaptability to rapidly changing environments. To cope with this challenge, we propose a cognitively autonomous framework for UAV networks empowered by multi-modal large language models (MMLLMs). The proposed framework integrates environmental cognition, intelligent decision- making, feedback optimization, and domain knowledge comprehension enhanced through retrieval-augmented generation (RAG). By leveraging heterogeneous onboard sensors, UAVs collect multi-modal data, enabling real-time environment cognition and achieving semantic- and intent-level comprehension via MMLLMs. Furthermore, advanced deep reinforcement learning (DRL) techniques are incorporated, facilitating adaptive decision-making and dynamic network control under varying conditions. The embedded RAG capability allows seamless integration of the latest domain-specific knowledge, including recent research findings, technical standards, and environmental prior knowledge. Feedback optimization continually refines network policies based on real-time evaluations, ensuring adaptability to evolving network dynamics. Simulation results show that our proposed framework substantially enhances UAV network performance and autonomy.

We propose a distributed unmanned aerial vehicle (UAV) network performing collaborative radar sensing for multi-target localization and motion parameter estimation. Two UAV network topologies are considered for data propagation and information fusion. In the former, we form a sequential UAV node chain, whereas in the latter, the UAV nodes are grouped into clusters and the information among different clusters is propagated through the cluster master nodes. Sparse reconstruction methods are used to fuse the target state information from previous nodes or clusters with the data measured in the underlying nodes or clusters, depending on the adopted topology, to achieve improved target state estimates. In order to minimize the communication traffic in the UAV network, each node transmits the estimated Doppler signatures or sparse target state estimates to the next UAV node in the network or the cluster master node, in lieu of the large volume of raw sampled data. Simulation results verify the effectiveness of the proposed approaches and compare the performance between the two UAV network topologies.

With the development of airborne equipment and integrated avionics technology, the unmanned aerial vehicle (UAV) network replaces human beings in many fields. To improve the durability of the UAV network, we propose a nondisruptive wireless rechargeable UAV network (WRUN) model, in which UAVs can be charged by wireless static chargers (WSCs) without returning back to the charging platform. Under the nondisruptive WRUN model, a baseline algorithm is proposed to solve the nondisruptive charging time schedule problem (nCTSP), in which chargers do not release energy all the time and can ensure UAVs do not run out of energy. Then to improve the energy utilization rate of WSCs, we propose an efficient charging time scheduling algorithm (ECTSA), in which the flight time and paths of UAVs are discretized and nCTSP is transformed into a linear binary integer programming (LBIP) problem to calculate the efficient charging time periods of WSCs. Finally, experiments are conducted to verify that ECTSA can improve the energy utilization of WSCs.

Routing in Unmanned Aerial Vehicle (UAV) networks is critical for effective data transfer and overall network performance. However, current UAV routing algorithms exhibit high latency, poor route selection, excessive energy consumption, and limited flexibility in changing network topologies. To overcome these limitations, this paper proposes a new routing strategy that uses the Shuffled Frog Leaping Algorithm (SFLA) to improve UAV network routing. Using a two-phase optimization approach considering Quality of Service (QoS), our system combines global exploration with local exploitation, unlike previous techniques. This hybrid method enables UAVs to dynamically change their trajectories, helping to choose the best path even in fast-changing surroundings. Our approach’s self-adaptive population-based search mechanism accelerates convergence and removes a common weakness in traditional metaheuristic algorithms—premature standstill elimination—which determines its effectiveness. By constantly adjusting UAV routing patterns depending on energy economy, latency, and throughput characteristics, SFLA guarantees that UAV networks transmit effectively and consistently. Based on experimental data, our method outperforms benchmark alternatives in terms of energy use by 3.11%, latency by 5.14%, and network lifetime by 2.25%. These developments make our approach ideal for real-time applications including aerial surveillance and disaster response that call for high data transfer speeds and great energy economy.

Unmanned aerial vehicles (UAVs) are being used extensively in various fields. UAVs provide various services to users, including monitoring, logistics, and sensing, because of their flexible deployment and dynamic reconfigurability. However, UAV networks have become more susceptible to malicious threats because of their multiconnectivity and openness. A great effort has been made to develop an effective intrusion detection system (IDS) based on machine-learning approaches for UAVs. Unfortunately, existing methods were unable to identify real time and zero-day attacks for UAV networks. This is due to that existing methods have still used obsolete data sets and past knowledge-based detection. Also, the shortcomings of standalone IDS render them unsuitable for defending UAV networks from potential security risks. Further, the lack of precise identification for compromised UAV nodes in UAV networks poses a critical security gap, risking the entire network’s integrity with the compromise of a single node. Therefore, in this work, we propose an autonomous collaborative IDS (UAV-CIDS) with a feedforward convolutional neural network (FFCNN), which accurately identifies zero-day with high accuracy. The proposed solution takes into account encoded Wi-Fi traffic logs of three popular UAVs types: 1) DBPower UDI; 2) parrot Bebop; and 3) DJI spark. Evaluation results indicate that our FFCNN model has produced outstanding results based on the UAVIDS data set with 98.23% accuracy compared to existing models. After the detection of attacks, their mitigation is equally significant. In addition, we also design and implement real-time incident response handling against cyber-attacks on UAV Networks. The incident response handling will assist in minimizing the effects of a security breach, remediate vulnerabilities and systematically secure the entire UAV networks.

To prolong the lifetime of unmanned aerial vehicle (UAV) networks, this paper studies how to efficiently deploy wireless static chargers (WSCs) in UAV networks (WDU). That is, given a set of UAVs, determining the minimum number of WSCs as well as the location of each WSC, so that UAVs do not run out of energy during flight. We first formulate WDU as an optimization problem, which is proved to be NP-hard. Then to solve WDU, we propose the scheme of selecting WSCs' locations by transforming WDU into a binary integer programming (BIP) problem (SLTB). In SLTB, we first propose a novel area discretization method to reduce the solution space of WDU, and then WDU is reformulated as a BIP problem that can be easily solved by existing methods. Finally, experiments are conducted to evaluate the performance of SLTB in terms of reducing the number of WSCs.

In the last few years, a considerable amount of research effort has been put into creating networks of Unmanned Aerial Vehicles (UAVs). Flying Ad Hoc Networks (FANETs) and Internet of Drones (IoDs) are the result of such efforts. FANETs assume that the member UAV nodes belong to different users. However, in UAV networks that would find their applications in areas such as agriculture, environment monitoring, crowd surveillance and natural disaster situations, all UAVs would belong to the same organization, creating a private UAV network. The number of UAVs used in private UAV networks is relatively small compared to that of FANETs due to the cost. If the packet routing approaches designed for FANETs are used in such networks, the quality of video transmitted through the network will suffer. This is because, while there can be many neighbors for each UAV in a FANET, the movements of UAVs in a private UAV network may cause some UAVs not to have any neighbors. To overcome this problem, a UAV in such a network needs to be controlled differently and an appropriate method of selecting the next hop node is required. The new approach presented in this paper uses mathematical analysis to find the optimum next hop node in a private UAV network by reducing a global optimization problem to a local optimization problem at each UAV node along the path from the source to the destination. Once the best next hop node is found, the UAVs can be controlled accordingly.

The article presents the development of a mathematical model of video monitoring based on a self-organizing network of unmanned aerial vehicles. The necessity of developing models and algorithms for providing geoecological monitoring using a wireless self-organizing network based on unmanned aerial vehicles is shown. Models are presented that allow calculating the speed of information transfer in the network and reducing the number of failures in the process of transmitting video data. With the help of models, it is possible to substantiate the power of network transmitting devices, at which the losses of transmitted packets are significantly reduced. The practical use of the model contributes to the achievement of the required quality of video surveillance in a wireless self-organizing network of unmanned aerial vehicles in the process of geoecological monitoring.

A population explosion has resulted in garbage generation on a large scale. The process of proper and automatic garbage collection is a challenging and tedious task for developing countries. This paper proposes a deep learning-based intelligent garbage detection system using an Unmanned Aerial Vehicle (UAV). The main aim of this paper is to provide a low-cost, accurate and easy-to-use solution for handling the garbage effectively. It also helps municipal corporations to detect the garbage areas in remote locations automatically. This automation was derived using two Convolutional Neural Network (CNN) models and images of solid waste were captured by the drone. Both models were trained on the collected image dataset at different learning rates, optimizers and epochs. This research uses symmetry during the sampling of garbage images. Homogeneity regarding resizing of images is generated due to the application of symmetry to extract their characteristics. The performance of two CNN models was evaluated with the state-of-the-art models using different performance evaluation metrics such as precision, recall, F1-score, and accuracy. The CNN1 model achieved better performance for automatic solid waste detection with 94% accuracy.

End-to-end network slicing in unmanned aerial vehicle (UAV) networks has attracted significant attention for delivering differentiated services tailored to the quality of service (QoS) requirements of user terminals (UTs). In a sliced UAV network, the dual mobility of both UTs and UAVs, combined with the three-layer association of UT-UAV-network slice (NS), makes handoff a critical and challenging problem for maintaining the communication performance of UTs. This paper proposes a sequential handoff process where a central controller makes handoff decisions for each UT sequentially according to the UT queue. This process is modeled as a Markov decision process (MDP). In the MDP, the state space includes NS available resources, UTs' QoS requirements, and UT queue information. Additionally, the reward function accounts for UTs' QoS and handoff overhead. The double deep Q-network (DDQN) technique is then employed to design a handoff algorithm. This algorithm achieves the optimization goal of ensuring UTs' QoS requirements while avoiding frequent handoffs by maximizing the total reward of the MDP. Additionally, a state-dependent exploration technique with adaptive intensity adjustment, called noisy network (NoisyNet), is utilized to drive the DDQN for effective exploration, thereby enhancing the performance of the handoff algorithm. Finally, the runtime complexity of the proposed NoisyNet-DDQN-based sequential handoff (NoisyNet-DDQN-SH) algorithm is analyzed. The simulation results show that the NoisyNet-DDQN-SH algorithm converges well. Moreover, it exhibits strong robustness. Additionally, the NoisyNet technique effectively enhances the algorithm's performance, making the NoisyNet-DDQN-SH algorithm superior to the vanilla DDQN (with an ϵ-greedy strategy), greedy, maximum signal-to-interference-plus-noise ratio (MaxSINR), and random algorithms.