Pavements used for construction and repair of airfield surfaces must be rigorously tested before use in the field. This testing is typically accomplished by trafficking simulated aircraft landing gear payloads across an experimental pavement test patch and measuring deflection, cracking, and other effects on the pavement and aggregate subbase. The landing gear payload is heavily weighted to simulate the pressures of landing and taxiing, and a large tractor pulls the landing gear repeatedly over the test patch while executing an intricate trafficking pattern to distribute the load on the patch in the desired manner. In conventional testing, a human drives the test vehicle, called a load cart, forward and backward over the experimental patch up to about 1000 times while carefully following a set of closely spaced lane markings. This is a dull job that is ripe for automation. One year ago, at this conference, we presented results of kitting the load cart, consisting of the tractor from a Caterpillar 621G scraper and a custom trailer carrying the landing gear simulacrum, with a custom vehicle interface and bringing it under tele-operation. In this paper, we describe the results of fully automating the load cart pavements test vehicle using the Robot Operating System 2 Navigation Stack. The solutions works without GPS, line following, or external tracking systems and involves minimal modifications to the vehicle. Using lidar and Adaptive Monte Carlo localization, the team achieved better than 6" cross-track accuracy with a lumbering, 300,000-pound vehicle.

Off-road autonomous navigation remains an ongoing challenge for autonomous ground vehicles (AGV). The challenges of navigating in an unstructured environment include identifying and detecting both positive and negative obstacles, distinguishing navigable from non-navigable vegetation, identifying soft soil, and negotiating rough or sloping terrain. While many recent works have dealt with various aspects of the off-road navigation problem, up to now there has not been a free and open-source autonomy stack for off-road that included integrated modules for perception, planning, and control. Therefore, we have recently developed the NATURE (Navigating All Terrains Using Robotic Exploration) autonomy stack as a publicly available resource to facilitate the advancement of off-road navigation research. The NATURE stack is implemented using the Robotic Operating System (ROS) and can be built to work with both ROS-1 and ROS-2. The modular nature of the NATURE stack makes it an ideal resource for researchers who want to evaluate a particular algorithm for perception, planning, or control without developing an entire navigation stack from scratch. NATURE features several options for both global and local path planning including A*, artificial potential field, and spline-based planning, as well as multiple options for perception including a simple geometrically based obstacle finder and more advanced custom traversability algorithm derived from 3D lidar. In this presentation we give an overview of the NATURE stack and show some past uses of the stack in both simulated and field experiments.

The capability to rapidly augment airbases with bio-concrete infrastructure to support parking, loading, unloading, rearming, and refueling operations is of interest to the Air Force, because it requires transportation of fewer raw materials to remote sites. Automation of the bio-cement delivery further reduces logistical requirements and mitigates hazards to personnel, especially in contested or austere environments. In this paper we discuss the full-stack development and integration of a robotic applique for a commercial tractor and present the test results for autonomous delivery of bio-cement bacteria, feed stock, and water for stabilization of a sandy test area. The tractor autonomously navigates, sprays, and avoids obstacles using robust and economical off-the-shelf components and software. For this first phase of the project, we employ GNSS for localization and automotive lidar for obstacle detection. We report on the design of the robotic applique, including the mechanical, electrical, and software components, which are mostly commercial-off-the-shelf or open source. We discuss the results of testing and calibration including tests of towing capacity, calibration of steering, measurement of liquid spray distribution, measurement of tracking errors, and determination of repeatability of positioning for refilling of the reservoir. We also examine higher order behaviors and chart a path forward for future development, which includes GNSS-denied navigation.

Swimming animals display exceptional ability to move efficiently in aquatic environments and display a rich diversity of mechanisms for generating and controlling propulsive force. The swimming profile and performance of aquatic animals has been of interest to engineers, biologist, and roboticists alike. Among the different swimming modes, undulatory swimming is common across various animals such as fish and eels. In this swimming mode, traveling or stationary waves are generated over the body and the control of which results in the generation of propulsive force. What is the relationship between the waveforms and the fluid mechanic forces? We sought to answer this question by combining experiments on biological swimmers and computational fluid dynamics simulations. We found that various swimming gaits can be generated through modulation of simple parameterized model. A bio-inspired robotic model was developed to demonstrate and test the locomotion dynamics of the various gaits. The findings from this study pave the way for highly maneuverable swimming robots that exploit the interplay between body waves and fluid forces.

Bio-inspired algorithms have been increasingly applied for autonomous robot path planning problems in complex environments. These environments are often restrictive in nature, where robot navigation must succeed with low margins of error. The complexity of the environment is a performance limiting factor based on density of obstacles and navigability of the robot in difficult environments. The scale of the environment to be examined for any given problem also contributes to the performance of solutions for path planning. These performance limitations are especially evident in time sensitive real-world applications, like autonomous off-road vehicles or search and rescue situations, where computation quality and immediacy are highly valued. One method to mitigate the shortcomings of bio-inspired algorithms involves destructing the problem environment into readily solvable segments. This paper proposes a graph-based near optimal path approach leveraging a bio-inspired algorithm for rapid path planning in autonomous environments. The proposed model utilizes centroid cell decomposition to establish a map in complex environments in a graph-based form. In this approach, centroid points are regulated and determined by the bio-inspired optimization as part of generating final robot trajectories. To improve upon the shortcomings of typical graph-based algorithms, ant colony optimization is applied afterwards to determine the near optimal robot traversal path. The model is validated through simulated environments for performance with comparable algorithms.

Robotic path planning and navigation in intricate environments pose significant challenges in various domains, including search and rescue, agriculture, and various defense applications. There have been various methods proposed to solve these problems, such as graph-based methodologies. Although the majority of cell decomposition methods lack the capability to develop a near optimal path, we propose a middle point cell decomposition in combination with a biologically inspired optimization algorithm for robot path planning and mapping. The proposed model leverages vertical cell decomposition in combination with an enhanced biologically inspired particle swarm optimization algorithm (ePSO). Vertical cell decomposition is employed as a spatial partitioning technique, segmenting complex environments into vertical cells, each characterized by a simplified geometric representation. To improve the path finding process, we introduce middle points within these cells. In this research, midpoints in the graph are regulated and slid by the developed biologically inspired optimization approach to generate optimal robot trajectories. This method enables the algorithm to approximate complex geometry more accurately and efficiently, facilitating smoother navigation for robotic systems. The primary objective of this study is to develop a comprehensive model for robotic path planning and navigation in complex environments, with a particular focus on enhancing efficiency, adaptability, and robustness. The proposed model is validated through extensive simulations in diverse complex environments. Comparative studies are performed against existing path planning algorithms, demonstrating the effectiveness of our approach in terms of path quality, computational efficiency, and adaptability to changing conditions.

This paper presents a comprehensive bibliometric analysis of the field of Autonomous Unmanned Systems (AUS), focusing specifically on genetic algorithm-based path planning, from 1997 to 2023. The study aims to map the evolution, trends, and future directions in this dynamic domain, highlighting the growing significance of genetic algorithms in the advancement of AUS. Our methodology integrates advanced bibliometric and text mining techniques, utilizing data from Scopus to provide both quantitative and qualitative insights. The analysis covers a corpus of 504 documents from 326 sources, revealing an increasing trajectory in research publications, particularly from 2010 onwards. This trend reflects the expanding academic and industrial interest in more sophisticated and efficient path planning methods for AUS. The paper identifies key thematic clusters, including optimization algorithms, energy and path efficiency, and communication technologies, emphasizing the necessity of interdisciplinary approaches in the field. Despite significant progress, challenges remain in safety, regulatory compliance, and enhancing the robustness and energy efficiency of path planning algorithms. The findings indicate a shift from foundational research to more applied and specialized areas, with potential new directions focusing on refining algorithms for specific applications and exploring integration with emerging technologies. Our study provides a comprehensive overview of the development of genetic algorithm-based path planning in AUS, offering valuable insights for future research directions. It underscores the importance of this field in various sectors and its potential for significant advancements in operational efficiency and effectiveness of autonomous unmanned systems.

Recently, reinforcement learning has been exhibited as being capable of providing base-level reasoning towards agent-based intelligence. Agents have had applications of reinforcement learning applied from simpler problem spaces (such as learning how to play with virtual cards), to learning how to make a physical robot walk. With reinforcement learning exhibiting capabilities to provide intelligence towards an individual agent, a question becomes how well could a reinforcement learning agent be able to manage multiple individual agents that have their execution of tasks abstracted. This challenge is important to recognize when we consider more advanced applications of reinforcement learning, such as leveraging reinforcement learning to conduct strategic coordination. In our studies, we have developed a system that leveraged reinforcement learning in an abstracted competitive strategic environment (currently, a real-time strategy (RTS) engine) to evaluate the effectiveness of reinforcement learning in automating the strategic approach of individual agents. To do this, we’ve defined several objectives for our reinforcement learning agent to perform (eg. offense, defense, economy). Given each one of these tasks, we can incentivize and disincentivize based on both general and objective-specific metrics. Based on the cultivation of autonomous strategic coordination systems, we believe this process will enable more robust situational responses in the future for autonomous, cooperative systems, and enable automated strategic response systems in operational domains.

We explore the use of transfer learning to reduce the data and computing resources required for training convolutional neural networks used by autonomous vehicles for predicting target behavior and improving target tracking as the scenario/environment changes. We demonstrate the ability to adapt to four different changes to the baseline scenario: a new target behavior, mission, adversary, and environment. The results from all four scenarios demonstrate positive transfer learning with reduced training datasets and show that transfer learning is a robust approach to dealing with changing environments even when the input or output dimensions of the neural network are changed.

Unmanned vehicles have continued to become commonplace in modern society, with the recent adoption of small unmanned aerial systems (sUAS) in the commercial, entertainment and defense industries. Despite encouraging trends in sUAS and unmanned systems (UxS) development, these technologies deployed in the field are still, in large part, limited to teleoperation and/or semi-autonomous behaviors of a single agent. The United States Department of Defense (DoD) is interested in elevating this current state of technology, specifically of aerial swarms for intelligence, surveillance, and reconnaissance (ISR) missions. While methods exist for optimal control of multi-agent systems, there remain novel research gaps related to robust field performance. The Robotics Research Center (RRC) at the United States Military Academy (USMA) is working to develop a collaborative aerial swarming architecture (CASA) that enables decentralized command and control (C2) between unmanned aerial systems (UAS). The main factors that support CASA’s decentralized capabilities are found in the dynamic allocation of tasks and the organization of data among sUAS platforms. This paper outlines CASA and its current capabilities. Task Allocation results are presented showing real-time task updates and allocation to a UAS swarm in a simulated environment.

The rapid growth of robot applications across various sectors, such as agriculture, disaster management, and package delivery, has led to an increased demand for efficient and safe multi-robot path planning methods. Existing solutions face significant challenges in simultaneously planning paths for multiple robots, maintaining safe distances from obstacles and other robots, and dealing with local minima issues. In this research, an efficient potential field-based method with nature-inspired algorithm is proposed to overcome these limitations, resulting in competent, short, and time-saving paths even in complex environments. Our proposed algorithms consider the simultaneous path planning of robots located at different locations while ensuring a safe distance from other robots. Multi-Robot multi-task allocation method is developed to optimize the path distance and time. A potential field-based scheme is proposed to avoid static and dynamic obstacles effectively. Additionally, our approach solves the local minima problem with simulated annealing algorithm, which is a supplement to this navigator. Finally, the proposed approach is efficient in generating short and time-efficient paths, even in complex dynamic environments. By augmenting the efficiency and safety of multi-robot operations, our work will contribute to the reduction of operational expenses, the enhancement of productivity, and the minimization of accident risks.

The growing emergence of UAV swarms in industrial and security applications such as search-and-rescue and emergency response underlines a need for efficient and accurate aerial navigation in potentially cluttered and dynamic environments. However, the introduction of multiple drones presents a two-fold complexity within the collaboration space: 1) the agents must be able to perform some desired task(s) cooperatively or independently, and 2) the agents must avoid colliding into one another (or other obstacles) in the process. In this work, we investigate a conceptual method for path planning and teaming behaviors for UAV swarms. Specifically, we explore the conceptual implementation of a gravitational-inspired spatial organization and path planning method for autonomously controlling a UAV swarm in a simple wayfinding scenario. The method models drones and the environment itself with characteristics of orbital dynamics. Preliminary experiments are demonstrated, showing the potential applicability to individual drone navigation and obstacle avoidance, along with a proposed concept for swarming behaviors. Keywords: drones, teaming, swarm, wayfinding, autonomy

Recent advances in tactical communications allow high bandwidth, directed line-of-sight (LOS) peer-to-peer information sharing between both manned and unmanned assets, yet current high-bandwidth communications infrastructures for ground and air lack the flexibility to support highly dynamic and mobile operations. Current communications solutions often require manned ground stations and established, planned-out infrastructure, yielding a rigidity which can struggle to meet the demands of highly dynamic Joint All Domain Command and Control (JADC2) environments. Drone Hosted Autonomous Radio Mesh Activity (DHARMA) seeks to augment JADC2 networks when needed and aid in network re-construction in the face of large-scale degradation of these networks. Prior work on DHARMA demonstrated the ability to augment tactical networks with a small swarm of radio-equipped drones in a sprawling urban environment. This work expands DHARMA capabilities to function in a multi-domain environment and allow significant scalability of swarm size over substantially larger and more diverse geographic areas.

Intelligent control techniques are increasingly assuming vital roles in contemporary unmanned systems. While Boolean logic, probability theory, and fuzzy logic stand out as widely used reasoning methods for intelligent control, they do come with their own set of limitations. To address these constraints associated with existing reasoning methods, we introduce a novel inference system called Earthling Logic (EL) into the domain of intelligent control. Within the framework of EL, all events are treated as statements with varying degrees of truth, referred to as credibility or truth value, which accommodates both vagueness and randomness as specific attributes. We can establish the existence of this credibility under broad and general conditions. Furthermore, we have developed analytically simple and computationally efficient techniques for deducing the credibility of statements based on the credibility of their premises. To illustrate the practical applicability of EL, we propose the Earthling Logic Control framework for unmanned systems.

The advancement of unmanned systems has set higher standards for making control decisions in the presence of uncertainty. In many unmanned system tasks, stochastic processes influence the stopping time of decision and the control performance index. This paper introduces an asymptotic distribution theory for stochastic processes with independent increments relevant to control systems. We show that, when properly normalized, the stopping time of control decision and the value of the stochastic processes at the stopping time converge asymptotically and independently, with the normalized value of the stochastic processes at the stopping time converging to a Gaussian random variable. Additionally, we derive the limiting distribution for the performance index, which depends on the stopping time and the corresponding value of the stochastic process. To illustrate the practical applications of these asymptotic results, we provide an example related to an integration system, a crucial component in stochastic control systems.

Fielding deep learning artificial intelligence (AI) capabilities in environments that are not well represented within training datasets remains a challenge in deploying reliable perception algorithms. Environments with scarce or no representation within training datasets lend to poor semantic scene segmentation and subsequently result in suboptimal autonomous navigation performance in Unmanned Ground Vehicles (UGVs). This can be attributed to multiple technical variables including intrinsic camera properties, lighting, weather, and seasonal differences, all of which pose significant issues related to a model’s ability to generalize to diverse environments and hardware configurations. Recently, zero-shot generalization capabilities for scene segmentation have been demonstrated with pre-trained foundational models. Combining the capabilities of such models with state-of-the-art semantic segmentation models can result in semantic representations of scenes with less label noise and better object boundaries. If accurate semantic labels can be applied to unlabeled segments, the resulting pseudo-labeled semantic segmentation data could be used to re-train an existing semantic segmentation model for new environments. To achieve this goal, we develop an architecture based on ensembles of semantic segmentation models to improve inferencing results in new environments by strengthening pixel label predictions used to classify unlabeled segmentation outputs. The process of automatically generating pseudo-labeled data can be computationally intensive and lacks the speed required for online inference on embodied systems. By utilizing the capabilities of pre-trained segmentation models in conjunction with an ensemble of semantic models, we can rapidly label data collected from a UGV in an environment that our fielded lightweight online model has never seen. Once the data is labeled, the original field model is retrained using the AI pseudo-labeled dataset and evaluated against the original field model. This work explores the possibility of a continuous learning framework that applies an ensemble of models to rapidly label data for model retraining. We present results showing that the approach can lead to improved algorithm performance with practical effect on the capabilities of UGVs relying on AI models which were trained on data from domains outside of the current operating environment. We show that models trained using our approach improved overall mIoU by an average of 4.75% on two distinct datasets and provide qualitative results for a third dataset.

Unmanned aerial systems (UAS) are a remarkably capable technology, and their untold potential scales greatly when multiple units work towards a common objective. However, current wireless and networking architectures limit the capabilities and applications of using multiple UAS primarily to pre-programmed routines – suppressing the advantages of multiple unmanned and autonomous platforms. Ad-hoc communications can unlock the true potential of unmanned and autonomous systems by enabling efficient and direct peer-to-peer data exchange and removing the system’s reliance on infrastructure – enabling them to operate anywhere. Unfortunately, ad-hoc communication is poorly supported by current wireless hardware and standards which are designed for maximized throughput and minimal latency over point-to-point links at the cost of much worse throughput and latency over multicast channels. Attempts to integrate multi-radio and cognitive radio technologies into the wireless ecosystem to increase device inter-connectivity have faced many difficulties because coordinating network resources and implementing a routing algorithm that can take all the involved variables into consideration is a very difficult task in host-centric networking. Our research uses Named Data Networking (NDN), a data-centric Internet architecture that uniquely identifies and retrieves data by name directly, to grant individual nodes the capability to make practical forwarding decisions without the added overhead of complex and centralized routing and coordination algorithms. Each node receives and gathers network information which then enables it to perform intelligent per-hop forwarding decisions toward names that identify data instead of addresses that identify hosts. By understanding what kind of data the node is handling, the requirements to successfully deliver that data, and its standing among other nodes, it can concurrently send data across multiple data-specific channels and radios to ensure quick, reliable, and non-interfering delivery. This will allow the network to effectively utilize the resources available to it and scale rapidly.

Significant increases in computing capacity may diminish in the near future as current processes become incapable of increasing the number of transistors per chip. As a result, there is an urgent need to determine new ways of computing, and the prospect of DNA technologies may provide the key to future advancement beyond current technology exploration. This is particularly critical given the extent to which U.S. defense mechanisms rely solely on current technologies and standards. DNA computing concepts hold the potential to bolster the United States’ operational advantage well into the future. Leonard Adleman in wrote on the possibility of computing directly with molecules, subsequently demonstrating a proof-of-concept using DNA. Since then, researchers have been exploring the potential of DNA computing for solving optimization problems, cryptography, and other computational tasks. While still in the experimental stage, DNA computing holds promise for addressing computationally intensive problems through biochemical reactions and DNA manipulation. DNA computing utilizes the combinatorial properties of DNA for massively parallel computations. Furthermore, advancements in DNA technology demonstrate a threat to current cryptography standards. In Boneh, Dunworth, and Lipton, they prove one can use DNA computing to break Data Encryption Standard, a block cipher standard. DNA computing is further able to encrypt and decrypt through One-time Password schemes with a potential for more complex cryptography such as Diffie Hellman and Elliptic Curve. The promises continue with DNA concepts through its capacity for high-powered computing and long-term, secure and stable storage. DNA computing may well be the necessary concept for future computing and cryptography implementations.

For rapid civil infrastructure assessment following natural and man-made emergencies, the utilization of minimally invasive and cost-effective drone deployable sensor packages has the potential to become a valuable tool. Although compact sensors with wireless data transfer capabilities have proven effective in monitoring the structural dynamics of infrastructure, these systems require data processing to occur externally, frequently off-site. These extra steps impede the high-speed assessment of a structure’s state. Difficulties can arise when the transmission is unfeasible due to degraded communication links during natural or man-made emergencies. Additionally, off-site data processing can add unneeded interruptions to actions that can be taken by emergency personnel after infrastructure damage. To enhance the effectiveness of sensor packages in expediting infrastructure assessment, incorporating real-time data analysis through embedded edge computing techniques emerges as a promising solution. The objective of this work is to demonstrate on-device data processing for frequency-based structural health monitoring techniques using drone-deployable sensors. This approach advances the effectiveness of drone-deployable sensors in rapid infrastructure assessment by mitigating their susceptibility to errors or delays in data communications. The proposed approach computes the frequency components of vibration measurements taken from a structure of interest, for example, the monitoring of a bridge immediately following a damaging event such as a flood. This work presents contributions in terms of outlining a methodology that emphasizes the hardware-based implementation of edge computing algorithms and examines the required on-device performance and resource utilization for structural health monitoring at the edge. The execution time for the sensor’s edge computing functions was profiled, resulting in an additional 9.77 seconds per test, an advancement over traditional transmit and analyze methods.

In the pursuit of research into autonomous systems, Robot Operating System (ROS) has become a de facto standard for both its functionality and extensibility. ROS has helped to provide standardized systems for common challenges such as communication, localization, and mapping. While ROS has historically been successful on these fronts, due to the prominence of containerization and micro-service architecture, ROS has had its usage shift to have resources run on containers leveraging distributed communication capabilities. While its distributed communication capabilities are present, we seek to better appreciate the efficacy of ROS in a distributed context by evaluating its impact on performance. To evaluate the impact of performance, we’ve leveraged the KITTI dataset to provide reproducible information streams that have both visible and LiDAR where we will evaluate mapping throughput to compare performance. Our metrics evaluate three methods of containerized ROS development: 1) A singular container with all relevant processing on a single machine, 2) Multiple containers on a single machine where processing is separated by logical task, and 3) Multiple containers where each process is in a domain-specific container distributed across a cluster of systems. Based on our comparison of performance across these paradigms, we believe these results will help better inform developers and researchers on the trade-offs given by ROS in a micro-service architecture.

Modern unmanned systems require the capability to make intelligent decisions in the face of uncertain information. Existing logical frameworks often fall short in addressing the inherent vagueness and randomness present in the available data. In this paper, we introduce an innovative reasoning system called Earthling Logic (EL), developed by emulating the mechanism of human reasoning. EL extends classical propositional and predicate logic to accommodate statements that exhibit varying degrees of vagueness and randomness. Within the framework of EL, we establish statemental algebras and truth measures. We introduce the notions of syntactical consequences and semantical consequences. Building on these ideas, we develop a general deduction principle that rigorously manages uncertainty when deriving statements from premises. We are able to establish the soundness and completeness of this deduction method. Additionally, we provide a general framework for decision-making based on EL, which can enhance the reasoning abilities of intelligent systems.

Swarm of Unmanned Aerial Vehicles using Emergence (SUAVE) efficiently creates a map in GPS-denied locations using Simultaneous Localization and Mapping (SLAM). Each drone is outfitted with a Lidar and an Intel Realsense Camera, which will be used to develop a high-fidelity infrared 3D map of an area without the need for GPS. Upon completion of each flight, the point clouds made through photogrammetry on each drone will be recorded and fused, creating a map with higher precision and accuracy. The use of the swarm is to generate this map from multiple points of view so that shadow effects can be negated and a more populated, dense map can be produced in respect to one drone’s map. To address localization with no GPS, the onboard IMU suite will be used to track relative position, while the onboard camera will track the local position. These two forms of localization, when coupled, allow for autonomous flights in lieu of GPS localization. This manuscript demonstrates the differences between one, three, and five fused maps during autonomous test flights on an Unmanned Aerial System (UAS) with a lack of GPS.

The unmanned aerial vehicle (UAV) and unmanned aerial system (UAS) are popular in nowadays applications including military, industry, weather casting, monitoring, and many other applications. According to several research, the system must be controlled in precise way to make sure that the UAV and UAS are moving in the desired trajectories. However, this task is not an easy task in real life due to the presence of disturbances and noise in feedback measurements. To overcome this issue, researchers either developed more stable controllers, i.e. active disturbance rejection control (ADRC), or they improved the measured signals using filters with more accurate/stable performance. This work belongs to the second category, where a newly developed filter, which is referred to as sliding innovation filter (SIF), is used to estimate the states of a UAV system while it is tracking a target at the same height to improve the quality of the controller.

In this study, we introduce a novel approach to the parameter estimation of Unmanned Aerial Vehicles (UAVs) utilizing the dandelion algorithm, a bio-inspired optimization technique that simulates the seed dispersal mechanism of dandelions. With UAVs increasingly becoming integral to various sectors, accurate parameter estimation emerges as a critical factor in ensuring their optimal performance and safety. Traditional parameter estimation methods often fall short, plagued by computational inefficiencies and a propensity for local optima, which can significantly hinder UAV operations. The dandelion algorithm, with its unique global search capabilities and adeptness in navigating multidimensional spaces, presents a solution that markedly enhances the precision and speed of parameter estimation. Through a series of simulations involving diverse UAV models, this study compares the performance of the dandelion algorithm against the conventional technique; the Particle Swarm Optimization (PSO), demonstrating its superior ability in achieving rapid convergence, higher accuracy, and an exceptional aptitude for avoiding local optima. Our findings not only underscore the algorithm's potential to revolutionize UAV parameter estimation but also highlight its applicability in advancing UAV technology and bio-inspired computational algorithms. This research contributes to the aerospace engineering field by offering an innovative, efficient alternative to existing parameter estimation methods, promising significant improvements in the design, operation, and safety of UAV systems across a spectrum of applications.

The use of drone networks in telecommunication especially in emergency situations has been increasing during the last years. The deployment of drones in the absence of network infrastructure or in cooperation with critical ones represents an important aid to guarantee the continuity of the services toward the clients. This paper introduces a modified Link-State routing method capable of regularly updating the link status of drones and dynamically reconfiguring the network topology, as well as re-routing current and incoming calls through paths with greater available bandwidth. To determine the optimal path for transmitting traffic between source and destination drones, we propose a modified version of the Ford-Fulkerson algorithm, considering the path with the highest remaining capacity. Additionally, we integrated a simple mobility prediction mechanism into the drone’s capabilities to accelerate route pre-computation during user movement, thereby preventing and minimizing packet loss.