Keywords: Multirotor Design and Control
Abstract: Standard drones propelled by four rotors are under-actuated systems. They use four control inputs to control four degrees of freedom independently. Hexacopters are driven by two more propellers, but since the direction of the total thrust remains normal to the drone's body, still only four degrees of freedom can be controlled independently. In this study, we describe a new hexacopter type consisting of three seesaws. Each seesaw is driven by two propellers, allowing rotation of the seesaw relative to the drone's body. Then, we develop the drone's unique control system and demonstrate its ability to maneuver with six controlled degrees of freedom while propelled by six motors only.

Keywords: Multirotor Design and Control, Control Architectures, Micro- and Mini- UAS
Abstract: This paper presents a trajectory-tracking approach for quadrotors using a tilt-prioritized attitude controller. The proposed control framework prioritizes tilt angles (the direction of the body z-axis) over yaw orientation to improve the translational trajectory tracking performance. A linear modulation parameter is introduced to enable a smooth transition between the UAV only tilting and the UAV tilting and orientating in yaw. Additionally, since it does not use operations with quaternions, matrix multiplication, or matrix inverses, the proposed controller is computationally efficient and easy to implement, making it well-suited for micro or small aerial vehicles. Real-world experiments validate the proposed method, demonstrating its effectiveness in agile trajectory tracking.

Keywords: Multirotor Design and Control, Energy Efficient UAS, Reliability of UAS
Abstract: The flight time of multirotor unmanned aerial vehicles (UAVs) is typically constrained by their high power consumption. Tethered power systems present a viable solution to extend flight times while maintaining the advantages of multirotor UAVs, such as hover capability and agility. This paper addresses the critical aspect of cable selection for tether-powered multirotor UAVs, considering both hover and forward flight. Existing research often overlooks the trade-offs between cable mass, power losses, and system constraints. We propose a novel methodology to optimize cable selection, accounting for thrust requirements and power efficiency across various flight conditions. The approach combines physics-informed modeling with system identification to combine hover and forward flight dynamics, incorporating factors such as motor efficiency, tether resistance, and aerodynamic drag. This work provides an intuitive and practical framework for optimizing tethered UAV designs, ensuring efficient power transmission and flight performance. Thus allowing for better, safer, and more efficient tethered drones.

Keywords: Multirotor Design and Control, Manned/Unmanned Aviation
Abstract: Reversible morphing wings, which can exchange the leading and trailing edges, expand architectural possibilities for aerial robotics (e.g., stopped-rotor configurations). However, few designs are scaled for uncrewed aerial vehicle (UAV) applications or effectively address the coupled aerodynamic and structural challenges of morphing. As such, we present a novel reversible wing design that uses rigid parallelogram slats mounted on a flexible substrate, creating a compliant yet aerodynamically robust structure. One-way fluid-structure interaction simulations validate the wing’s structural performance under airflow. Compared to other reversed-flow wings, the proposed configuration doubles reverse-flow performance relative to a Clark-Y wing and improves upon the 0 degree angle of attack performance compared to other reversible morphing wings.

Keywords: Multirotor Design and Control, Navigation, Autonomy
Abstract: This paper investigates the application of Deep Reinforcement (DRL) Learning to address motion control challenges in drones for additive manufacturing (AM). Drone-based additive manufacturing promises flexible and autonomous material deposition in large-scale or hazardous environments. However, achieving robust real-time control of a multi-rotor aerial robot under varying payloads and potential disturbances remains challenging. Traditional controllers like PID often require frequent parameter re-tuning, limiting their applicability in dynamic scenarios. We propose a DRL framework that learns adaptable control policies for multi-rotor drones performing waypoint navigation in AM tasks. We compare Deep Deterministic Policy Gradient (DDPG) and Twin Delayed Deep Deterministic Policy Gradient (TD3) within a curriculum learning scheme designed to handle increasing complexity. Our experiments show TD3 consistently balances training stability, accuracy, and success, particularly when mass variability is introduced. These findings provide a scalable path toward robust, autonomous drone control in additive manufacturing.

Keywords: Multirotor Design and Control, Perception and Cognition, Control Architectures
Abstract: The problem of image-based visual servoing (IBVS) of an aerial robot using deep-learning-based keypoint detection is addressed in this article. A monocular RGB camera mounted on the platform is utilized to collect the visual data. A convolutional neural network (CNN) is then employed to extract the features serving as the visual data for the servoing task. This paper contributes to the field by circumventing not only the challenge stemming from the need for man-made marker detection in conventional visual servoing techniques, but also enhancing the robustness against undesirable factors including occlusion, varying illumination, clutter, and background changes, thereby broadening the applicability of perception-guided motion control tasks in aerial robots. Additionally, extensive physics-based ROS Gazebo simulations are conducted to assess the effectiveness of this method, in contrast to many existing studies that rely solely on physics-less simulations. A demonstration video is available at https://youtu.be/Dd2Her8Ly-E.

Keywords: Perception and Cognition
Abstract: Autonomous maritime surveillance and target vessel identification in environments where Global Navigation Satellite Systems (GNSS) are not available is critical for a number of applications such as search and rescue and threat detection. When the target vessel is only described by visual cues and its last known position is not available, unmanned aerial vehicles (UAVs) must rely solely on on-board vision to scan a large search area under strict computational constraints. To address this challenge, we leverage the YOLOv8 object detection model to detect all vessels in the field of view. We then apply feature matching and hue histogram distance analysis to determine whether any detected vessel corresponds to the target. When found, we localize the target using simple geometric principles. We demonstrate the proposed method in real-world experiments during the MBZIRC2023 competition, integrated into a fully autonomous system with GNSS-denied navigation. We also evaluate the impact of perspective on detection accuracy and localization precision and compare it with the oracle approach.

Keywords: Perception and Cognition, Path Planning
Abstract: In this paper, we present a novel visual servoing (VS) approach based on latent Denoising Diffusion Probabilistic Models (DDPMs), that explores the application of generative models for vision-based navigation of UAVs (Uncrewed Aerial Vehicles). Opposite to classical VS methods, the proposed approach allows reaching the desired target view, even when the target is initially not visible. This is possible thanks to the learning of a latent representation that the DDPM uses for planning and a dataset of trajectories encompassing target-invisible initial views. A compact representation is learned from raw images using a Cross-Modal Variational Autoencoder. Given the current image, the DDPM generates trajectories in the latent space driving the robotic platform to the desired visual target. The approach has been validated in simulation using two generic multi-rotor UAVs (a quadrotor and a hexarotor). The results show that we can successfully reach the visual target, even if not visible in the initial view. A video summary with simulations can be found in: https://youtu.be/2Hb3nkkcszE.

Keywords: Perception and Cognition, Payloads, Environmental Issues
Abstract: Unmanned aerial vehicles (UAV) have emerged in recent years as powerful, maneuverable sensors capable of real-time computer vision. Real-time image processing onboard UAV often requires data or model compression, acceleration, or edge offloading and is generally restricted to conventional RGB cameras. In this study, we consider real-time in-situ processing for hyperspectral imaging (HSI). HSI cameras detect many wavelengths of light. Material-specific spectral signatures can be matched to camera outputs to identify materials in a UAV's environment, but HSI cameras produce large amounts of information that generally require offline processing by heavy-weight software. We present REMIX, a real-time hyperspectral processing payload for small UAV. REMIX uses a custom software library, light-weight hyperspectral camera, and small embedded device to process and visualize HSI data in real-time. REMIX processes HSI lines in under 5ms, allowing HSI perception to be visualized in real-time where conventional methods may take hours. We show that, when properly configured, adding real-time processing via REMIX degrades UAV flight time by only 4% and increases HSI processing speeds by up to 6X compared to naive payloads, and further decreases post-processing time by 20.48X compared to conventional methods, even when using significantly less powerful equipment.

Keywords: Perception and Cognition, Payloads, Smart Sensors
Abstract: This article describes an RF Direction Finding (DF) payload developed for UAV systems. DF payloads sense RF signals using an antenna array and process the received signals at each antenna location to estimate the number of transmitting RF sources and their bearing relative to the payload. This article uses an open source Software Defined Radio (SDR) known as the KrakenSDR which senses transmitted RF data with (5) antennas. A new deep learning architecture is proposed for estimating the azimuthal Direction of Arrival (DoA) of RF signals from the sensed KrakenSDR antenna data. The recent availability of the compact and comparatively lightweight KrakenSDR hardware for DF applications make academic investigation of this sensor for UAS possible. DF payloads are used in a wide variety of important applications including search-and-rescue, signal intelligence, RF source geolocation, spectrum monitoring, spectrum enforcement and disaster management contexts. This article describes results for a new DoA estimation algorithm and includes discussion on integration challenges, mechanical and electromagnetic design considerations and the payload Size Weight and Power-Cost (SWaP-C) metrics using the KrakenSDR hardware.

Keywords: Perception and Cognition, Sensor Fusion, Reliability of UAS
Abstract: The rapid advancements in unmanned aerial vehicle (UAV) technology have led to their use in different applications, ranging from critical infrastructure monitoring and search-and-rescue to remote sensing. However, UAV operations are easily affected by environmental conditions and sensor malfunctions that lead to the need for an efficient, accurate, and trustworthy state identification and trajectory prediction framework. This work proposes an innovative real-time UAV system with the two-fold objective of state identification and trajectory prediction, employing a lightweight multi-task learning framework based on reservoir computing (RC) network architecture to achieve reliable and robust UAV operations. Specifically, custom multi-task models are designed and fine-tuned to obtain multi-modal sequential data (related to drone movement) by exploiting the ability of shared feature learning in an RC-based network architecture to accurately achieve and enhance real-time and simultaneous drone state classification and trajectory prediction. A real-world dataset is also created to train and evaluate the proposed multi-task model, encompassing drone movements recorded during numerous outdoor experiments. Finally, a UAV prototype system is implemented and extensively tested in a real-world environment to demonstrate its enhanced performance in trajectory prediction and drone state identification compared to existing methods.

Keywords: Perception and Cognition, UAS Applications, Technology Challenges
Abstract: This study compares the performance of state-of-the-art neural networks including variants of the YOLOv11 and RT-DETR models for detecting marsh deer in UAV imagery, in scenarios where specimens occupy a very small portion of the image and are occluded by vegetation. We extend previous analysis adding precise segmentation masks for our datasets enabling a fine-grained training of a YOLO model with a segmentation head included. Experimental results show the effectiveness of incorporating the segmentation head achieving superior detection performance. This work contributes valuable insights for improving UAV-based wildlife monitoring and conservation strategies through scalable and accurate AI-driven detection systems.

Keywords: Biologically Inspired UAS, Micro- and Mini- UAS, Simulation
Abstract: This work focuses on the design and control of a novel type of Flapping-Wing Micro Aerial Vehicle (FWMAV). The drone, known as the X-Fly, is a new under-actuated robotic platform that also has an inner control loop to stabilize its roll angle thanks to an onboard IMU. Such assistance makes the X-Fly easier to pilot. The under-actuation and the flapping oscillations make the modelling and the control of the X-Fly a challenging task.

A dynamical control model is introduced that is able to take advantage of the stabilized roll dynamics to separate the platform into two almost independent sub-systems, one for the altitude and another for the position on the x-y plane.

A trajectory tracking controller for the altitude and a circular trajectory are then proposed and tested in order to corroborate the validity of the presented model.

Keywords: Biologically Inspired UAS, Micro- and Mini- UAS, Swarms
Abstract: This paper employs a biologically inspired logic for trajectory generation for a swarm of autonomous aerial vehicles, using passive distance estimation from onboard visual cameras. The method is inspired by swarming birds that use the perception of neighboring birds to modify their own motion, based on passive sensory data. Based on birds' spatial proximity, the logic enables stable swarming without explicit inter-agent active distance control and specific neighbor identification. A decentralized technique is used that utilizes optimal guidance and control for trajectory tracking without centralized computations while progressing in a general direction and speed. Each agent, equipped with visual cameras, achieves a cohesive and coordinated contribution to the formation. The approach is validated through simulation using unmanned aircraft models controlled by nonlinear model predictive controllers, and by inferring distance from images between adjacent agents.

Keywords: Biologically Inspired UAS, Smart Sensors, Micro- and Mini- UAS
Abstract: Biological fliers’ remarkable manoeuvrability and robust flight control are aided by information from dense arrays of distributed flow sensors on their wings. Bio-inspired fixed-wing uncrewed aerial vehicles (UAVs) with a “flight-by-feel” control approach could mimic these abilities, allowing safe operation in cluttered urban areas. Existing work has focused on longitudinal parameter estimation and control at low angles of attack. This wind-tunnel study estimates both the longitudinal and lateral-directional aerodynamic states of a bio-inspired distributed pressure sensing UAV at angles of attack and sideslip up to 25° and 45°. Four span-wise strips of pressure sensors were found to show strong, location dependent variation with angle of sideslip across all angles of attack, indicating that distributed pressure sensing arrays can encode lateral-directional flow information. This was supported by the use of the pressure signals in estimator algorithms, which showed angle of sideslip estimation was possible with both a linear partial-least-squares regression-based estimator and a non-linear feed-forward artificial neural network estimator. The non-linear estimator could predict angle of sideslip with a lower error than the linear estimator, with a root-mean-square error (RMSE) of 0.70° for the former compared to 1.23° for the latter. They both showed good estimation of angle of attack, even in the post-stall regime, with an RMSE of 0.58° for the linear estimator and 0.54° for the non-linear estimator. These results show that pressure-based distributed sensing can capture a complete aerodynamic picture of a UAV, unlocking the potential of a “flight-by-feel” control system informed by the aerodynamic states of the vehicle across a wide range of aerodynamic conditions.

Keywords: Autonomy, See-and-avoid Systems, Manned/Unmanned Aviation
Abstract: Despite the exponential and promising growth in urban air mobility, this sector faces multifaceted technological and societal challenges. Among the most critical is the development of safe, scalable collision avoidance systems capable of operating within the highly dynamic and congested airspace of metropolitan environments, where complex flight routes must navigate dense infrastructure, variable weather, and unpredictable traffic patterns. Traditional collision avoidance methods, such as potential field methods and TCAS, have limitations at low altitudes and in spatially congested metropolitan areas. This work presents a safety-critical control design using control barrier functions that not only guarantees safe operation but can also be applied to any existing system with minimal impact. Six degrees of freedom simulations show that the controller maintains the ego vehicle’s safety across multiple scenarios and is capable of running in real-time for real-world implementation.

Keywords: Micro- and Mini- UAS, Airspace Control, Navigation
Abstract: Position-Based Visual Servoing (PBVS) is a widely used technique for UAV control, enabling precise motion based on visual feedback. This paper presents a nonlinear control strategy based on a Barrier Lyapunov Function (BLF) to ensure exponential stability in fully actuated UAVs performing PBVS tasks. Unlike underactuated multirotors, fully actuated drones provide independent control of translation and orientation, making them well-suited for vision-based applications. We propose a velocity-state feedback control law that guarantees stability by leveraging a BLF-based approach. The method ensures that the velocity errors remain bounded while converging exponentially to zero, enhancing robustness in trajectory tracking. The control framework is integrated within the PX4 autopilot system and validated through Software-in-the-Loop (SITL) simulations in Gazebo, demonstrating its effectiveness in real-time UAV operations. Software-in-the-loop simulation results confirm the proposed controller’s capability to track PBVS-generated velocity references accurately while maintaining stability under varying conditions. The integration of homography-based visual control further improves precision in vision-based UAV navigation. This work contributes to developing nonlinear control techniques for fully actuated UAVs, bridging the gap between theoretical control design and real-time implementation.

Keywords: Micro- and Mini- UAS, Energy Efficient UAS
Abstract: An experimental wind tunnel study is performed on an EPPLER-186 airfoil equipped with a Gurney flap. The main goal is to improve the lift coefficient and lift-to-drag ratio of a small Unmanned Aerial Vehicle (UAV) during different flight conditions. This way, the vehicles would perform better aerodynamics, reducing take-off and landing distances. The aerodynamic forces are obtained using an external balance to quantify the effect on the flow with the various sizes of Gurney flaps installed. Adding the device at the trailing edge significantly increases lift values at low angles of attack (up to +0.32 points in lift coefficient). Drag values also increase, but for cruise flight at low angles of attack aerodynamic efficiency increases up to +6 points with respect to the base wing without Gurney flaps.

Keywords: Aerial Robotic Manipulation
Abstract: This survey summarizes key control approaches and architectures that reflect the state-of-the-art in aerial manipulation. The central objective is to provide a thorough resource for researchers exploring multirotor configurations suitable for real-time aerial manipulation applications. The focus is on evaluating and comparing prototype systems and their corresponding controller designs, emphasizing real-time implementation, regardless of the number of DOFs of the attached manipulator(s) and of specific applications. The survey groups control methods in three categories based on the specific architecture that is followed: coupled, partially coupled, and decoupled. The metrics used for the comparative study include system configuration, total weight, modeling approach, control architecture, robustness, implementation complexity, task execution precision, and achieved results (via simulations or experiments).

Keywords: Aerial Robotic Manipulation, Airspace Control, Airspace Management
Abstract: Player tracking is a useful tool for tactical analysis and performance evaluation in soccer, providing valuable insights into player movements and team dynamics. This project investigates the feasibility of tracking players using UAV-captured imagery, employing both YOLO and traditional image processing algorithms (TIPA). Initial validation focuses on robot soccer players due to their predictable and controllable movements. The comparative analysis considers processing time, computational cost, adaptability to environmental changes, sensitivity to lighting variations, ability to handle dynamic conditions, tracking accuracy, and real-time performance. Results indicate that, under equivalent hardware and preparation time conditions, YOLO achieves performance comparable to traditional techniques. Nonetheless, the selection of the most suitable approach should be guided by task-specific demands, available computational resources, and the time allocated for system development and deployment.

Keywords: Aerial Robotic Manipulation, Control Architectures, Simulation
Abstract: This work investigates cooperative dual-arm manipulation between two ornithopters in the post-perching phase. Flapping wing aerial systems are lightweight platforms designed to imitate bird flight, suitable for environmental monitoring tasks. When interacting with their environment, these systems must be able to perch on a branch as an initial step, followed by adjusting their position to achieve the desired pose and workspace. This research explores the application of a Port-Hamiltonian-based control method for designing and analysing controllers in cooperative manipulation by two ornithopters during the post-perching phase. The connection of end effectors while holding an object adds complexity and constraints to the problem. To address this, an energy-based approach using Optimal Port-Hamiltonian control and Optimal Load Distribution (OLD) is employed to evenly distribute the load between the arms. The effectiveness and advantages of this method are demonstrated through the defined scenario in which an optimal control law is implemented to derive an efficient trajectory for cooperative manipulation while tracking the desired elliptical path.

Keywords: Aerial Robotic Manipulation, Control Architectures, Simulation
Abstract: Enabling aerial robots to handle dynamic contacts happening at non-vanishing speeds can enlarge the range of their applications. In this work, we propose an impact- aware strategy to allow aerial multirotor robots to recover from impacts. The method leverages a reactive strategy not requiring low-level changes to the motion controller commonly implemented onboard quadrotors, which might be not viable or not desirable for most users. Extensive simulation tests show that the proposed strategy considerably increases the tolerated velocity at impact in tasks in which the robot either picks an object up or collides against an object to clear its way. Preliminary experimental results using Crazyflie UAVs are also presented.

Keywords: Aerial Robotic Manipulation, Micro- and Mini- UAS, Control Architectures
Abstract: This paper presents an enhanced observer control strategy for multirotor aerial grasping. Unlike previous approaches, which focused solely on translational dynamics, this method incorporates dual observers—one for the translational subsystem and another for the rotational dynamics. By leveraging quaternions, the proposed control framework provides a singularity-free representation of orientation while naturally decoupling rotational and translational dynamics. This allows the system to be treated as fully actuated in both position and orientation, improving disturbance rejection and compensating for torques induced by off-center or asymmetrically shaped objects during grasping. A passive, non-actuated gripper further enhances the drone’s ability to interact with objects in real-world scenarios. Experimental validations confirm the robustness and adaptability of the proposed approach, demonstrating its effectiveness in handling dynamic variations in mass and torque while maintaining stable flight.

Keywords: Aerial Robotic Manipulation, Multirotor Design and Control
Abstract: The task of horizontal aerial screwdriving has been a relatively unexplored area of aerial manipulation, yet has immense potential and use cases in fields such as construction, maintenance and inspection, especially in dangerous or costly scenarios. This paper presents an analysis of a novel screwdriving actuator integrated with a fully-actuated UAV, identifying critical performance limits of the proposed design. As the UAV uses frictional torque generated from a high-friction contact plate to counteract reactional torque from screwdriving, a relationship is derived to identify the minimum contact plate annulus diameter. The robustness of the UAV and screwdriving actuator are also quantified by identifying relationships for preventing pivoting about the contact plate from lateral disturbances. An analysis in screw torque requirements is performed and a flight test is conducted to investigate the performance of the screwdriving actuator in flight, which shows successful mitigation of reactional torques up to 0.03 Nm.

Keywords: Navigation, Path Planning, Autonomy
Abstract: Autonomous drone navigation in dynamic environments remains a critical challenge, especially when dealing with unpredictable scenarios including fast-moving objects with rapidly changing goal positions. While traditional planners and classical optimisation methods have been extensively used to address this dynamic problem, they often face real-time, unpredictable changes that ultimately leads to sub-optimal performance in terms of adaptiveness and real-time decision making. In this work, we propose a novel motion planner, AgilePilot, based on Deep Reinforcement Learning (DRL) that is trained in dynamic conditions, coupled with real-time Computer Vision (CV) for object detections during flight. The training-to-deployment framework bridges the Sim2Real gap, leveraging sophisticated reward structures that promotes both safety and agility depending upon environment conditions. The system can rapidly adapt to changing environments, while achieving a maximum speed of 3.0 m/s in real-world scenarios. In comparison, our approach outperforms classical algorithms such as Artificial Potential Field (APF) based motion planner by 3 times, both in performance and tracking accuracy of dynamic targets by using velocity predictions while exhibiting 90% success rate in 75 conducted experiments. This work highlights the effectiveness of DRL in tackling real-time dynamic navigation challenges, offering intelligent safety and agility.

Keywords: Autonomy, Control Architectures, UAS Testbeds
Abstract: Landing a multirotor unmanned aerial vehicle (UAV) on an uncrewed surface vessel (USV) extends the operational range and offers recharging capabilities for maritime and limnology applications, such as search-and-rescue and environmental monitoring. However, autonomous UAV landings on USVs are challenging due to the unpredictable tilt and motion of the vessel caused by waves. This movement introduces spatial and temporal uncertainties, complicating safe, precise landings. Existing autonomous landing techniques on unmanned ground vehicles (UGVs) rely on shared state information, often causing time delays due to communication limits. This paper introduces a learning-based distributed Model Predictive Control (MPC) framework for autonomous UAV landings on USVs in wave-like conditions. Each vehicle's MPC optimizes for an artificial goal and input, sharing only the goal with the other vehicle. These goals are penalized by coupling and platform tilt costs, learned as a Gaussian Process (GP). We validate our framework in comprehensive indoor experiments using a custom-designed platform attached to a UGV to simulate USV tilting motion. Our approach achieves a 53% increase in landing success compared to an approach that neglects the impact of tilt motion on landing. For accompanying video: https://youtu.be/g4cCmE9Rgxs.

Keywords: Aerial Robotic Manipulation, Navigation, Path Planning
Abstract: Autonomous exploration in unknown areas is a challenge for unmanned aerial vehicles when traditional ranging sensors such as LIDARs or cameras fail due to dust, fog, or lack of illumination. In these situations, contact-informed navigation is leveraged by utilizing the end-effector of an unmanned aerial manipulator (UAM) to detect and exploit obstacle contacts. This work presents a contact-informed online replanning algorithm that updates an obstacle-bounding region using online wrench estimates, enabling a UAM to navigate around an unknown convex polyhedral obstacle. The planner generates joint-space setpoints that guide the tool center point (TCP) to track a reference trajectory while ensuring the multirotor body avoids the obstacle-bounding region. Two simulation cases show that this approach prevents multirotor body collisions and ensures TCP trajectory tracking.

Keywords: Control Architectures, Multirotor Design and Control, Autonomy
Abstract: A novel formulation of model predictive control (MPC) coupled with Koopman operator theory is presented and tested for the trajectory tracking problem of a quadrotor UAV. The analytical derivation of Koopman observables allows for the quadrotor model to be written as a fully-actuated quasi-linear system which enables the control problem to be posed as a linear control problem. In fact, the adopted approach embeds the quadrotor nonlinear dynamics into a quasi-linear form through the evolution of the Koopman operator generalized eigenfunctions, a special kind of Koopman observables. Hence, the linear MPC formulation in Koopman coordinates is equivalent to a nonlinear implementation in the original state space. Moreover, in an enhancement from the standard feedback linearization, the Koopman based quadrotor model does not present underactuation, which drastically simplifies the computational requirement for the solution of the MPC optimization problem. The presented methodology is tested through detailed numerical simulations and results are compared to single-loop nonlinear MPC (NMPC). The satisfactory tracking performance are additionally enhanced by the obtained computational speedup which is crucial for real time implementation of flight controllers.

Keywords: Multirotor Design and Control, Energy Efficient UAS, Micro- and Mini- UAS
Abstract: Drones that can morph their shape are used to sidestep a design trade-off when the traversal of small gaps is required. Typically, small drones that are able to fit through small gaps are less efficient than larger drones that can’t fit through the same gaps. Morphing drones combine both advantages by being big and efficient in their normal configuration, and by temporarily becoming small and inefficient in their morphed configuration. The here presented FLIFO (flip + fold) morphing drone manages to shrink to half its width, while maintaining full controllability. It does so purely passively, without requiring any additional actuators beside the ones needed for flight. This is an unprecedented accomplishment in morphing drones. Concretely, FLIFO's design consists of four simple hinges placed in a particular orientation on each arm, that cause the morphing once the drone flips up-side-down. Test flights of a prototype have successfully shown that this design can transition robustly between configurations while remaining in a tightly confined space, barely larger than the drone itself.

Keywords: Swarms, Path Planning, UAS Applications
Abstract: This paper proposes an online motion planning algorithm for defender drones to protect a High-Value Unit (HVU) against a swarm of attacker drones. We formulate an optimal motion planning problem and approximate its solutions using Bernstein polynomials. The favorable geometric properties of the polynomials allow to compute the cost function and constraints efficiently. Since the attackers' dynamics are generally imperfectly known, we resort to model predictive control (MPC) approach. By predicting future trajectories of the attackers over a short time interval, we calculate optimal trajectories for the defenders to shoot down the attackers and maintain the survival probability of the HVU close to one. This optimization problem is solved recursively with a receding time horizon until the attackers are incapacitated.

Keywords: UAS Applications, Autonomy, Aerial Robotic Manipulation
Abstract: This paper presents ρLiRLo, a LiDAR-based Relative Localization method, designed for reliable robot navigation and control in GNSS-denied environments. ρLiRLo enhances point cloud processing using intensity filtering with a retro-reflective marker. The marker’s position is determined via Euclidean clustering, while a Kalman filter tracks the robot's pose. To improve localization accuracy in dynamic conditions, IMU measurements are integrated, and a robotic manipulator actively tracks the marker, expanding LiDAR’s field of view. The method is demonstrated on an Unmanned Aerial Vehicle (UAV) in both indoor and outdoor experiments. Indoor tests benchmark localization against OptiTrack motion capture, while outdoor experiments are conducted in a maritime environment with the tracking system mounted on an Unmanned Surface Vehicle (USV).To mitigate the challenges of dynamic sea conditions, IMU measurements are used to compensate for disturbances introduced by waves and wind. ρLiRLo demonstrates high accuracy, low-latency feedback, and strong potential for applications in GNSS-denied settings.

Keywords: UAS Applications, Environmental Issues, Air Vehicle Operations
Abstract: Infrastructure inspections using UAVs have surged in recent years thanks to their ability to capture high-resolution imagery in hard-to-reach areas. Their versatility has garnered significant interest in applications such as bridge inspections, offering the potential to substantially reduce both costs and inspection time. However, UAVs are highly sensitive to environmental factors like turbulence and wind gusts, which can compromise their stability and lead to accidents. This is particularly critical in bridge inspections, where structural components such as pillars, decks, and cables generate complex wind patterns, including vortices and turbulence. To address these challenges, this paper presents a wind assessment methodology for UAV-based bridge inspections. To this end, an automated 3D urban geometry modeling methodology was developed using open-source geospatial data, and wind predictions were calculated via the CFD (Computational Fluid Dynamics) software OpenFOAM. A practical case study was carried out in Porto, Portugal, to validate the proposed methodology.

Keywords: UAS Applications, Navigation, Sensor Fusion
Abstract: This paper presents a solution for high accuracy fruit counting in a Controlled Environment Agriculture (CEA) settings using Unmanned Aerial Vehicles (UAV) implemented in simulation and real-world scenarios. Using the fine-tuned YOLOv8 model for precise object detection and classification along with a custom path planning algorithm for simulation. This approach integrates a hybrid A* Traveling Salesman Problem (TSP) algorithm for efficient 3D path planning. The solution is further extended to the real-life scenario using LiDAR based mapping and point-cloud filtering techniques to avoid recounting of fruits. The suggested methodology increases operating efficiency, reduces dependency on human labor, and improves accuracy. Experimental results, derived from both simulations and real-world testing, achieving a fruit-count accuracy of 98% and 90% respectively, demonstrate the effectiveness of this integrated approach. This solution was implemented in the International Conference of Unmanned Aircraft System (ICUAS) UAV Competition 2024, securing 2nd position in the simulation phase (out of 24 teams) and 3rd overall in the real-world phase.

Keywords: UAS Applications, Networked Swarms, Manned/Unmanned Aviation
Abstract: Intrusion detection is a critical application in UAV networks with downward-facing cameras, where the barrier coverage problem entails strategically positioning UAVs to protect a region’s perimeter. For a terrain-like border, achieving optimal UAV placement is challenging due to factors like resolution, overlap constraints, and varying altitudes across the terrain that have not been explored in previous studies. This paper addresses the barrier coverage problem in the context of a terrain-like border using Unmanned Aerial Vehicles (UAVs). We first simplify the 3D problem into an equivalent 2D model and introduce a resolution cost to evaluate the quality of terrain coverage. We also define the overlapping length and formulate an optimization problem to ensure barrier coverage for an initially uncovered belt. Our approach is validated through several example simulations.

Keywords: UAS Applications, Path Planning, Autonomy
Abstract: Autonomous inspection of large and complex structures with a commercial unmanned aerial vehicle (UAV) is a challenging problem that has been addressed in recent years. In this paper, we address the global motion planning problem of creating autonomous inspection missions for UAVs considering photogrammetry constraints. We focus on the inspection of large tailings dams, which are dam structures used to store waste byproducts of mining. Our method uses a prior sparse point cloud of the dam to generate a voxel grid, where paths satisfying photogrammetry constraints are tested for collisions. We then apply the A* algorithm as a local planner to avoid obstacles within the global mission. Moreover, we address the problem of changing routes online by using octree-based multi-resolution grids for efficient and fast pathfinding. Our results, obtained using tridimensional maps of an actual coal mine tailings dam, show that using octrees for multi-resolution motion planning is faster than using a fixed voxel grid in online missions while inspecting large structures.

Keywords: UAS Applications, Sensor Fusion, Simulation
Abstract: The increasing availability of high-resolution, multi-spectral cameras for small Unmanned Aerial Systems (sUAS) has enabled detailed aerial mapping for applications such as precision agriculture and environmental monitoring. However, generating orthomosaics from high-resolution imagery presents significant computational challenges, particularly for real-time processing on resource-constrained edge devices. This paper evaluates the feasibility of SLAM-based orthomosaic generation for high-resolution, multi-band sUAS imagery. We systematically analyze trade-offs in resolution scaling, feature extraction strategies, and incremental bundle adjustment techniques, quantifying their effects on accuracy, computational cost, and scalability. Our results show that while global bundle adjustment improves accuracy, localized selection strategies significantly reduce processing time, improving real-time processing performance. Additionally, we discuss the limitations of existing SLAM-based pipelines in handling high-resolution imagery and highlight opportunities to improve performance. By identifying key computational bottlenecks and accuracy trade-offs, this study provides insights for optimizing SLAM-based aerial mapping pipelines for real time scientific grade data analysis.

Keywords: Path Planning
Abstract: Most UAV path planning methods assume that speed is constant or controllable within certain constraints. However, some applications require UAVs to follow predefined speed profiles. This paper proposes a novel offline path planning algorithm for multi-agent UAV systems with fixed speed profiles that facilitates scheduled arrivals at desired final states in R2- space using uniform B-splines. The B-splines are parameterized by a path variable to decouple the path geometry from the speed profile, and a path extension algorithm is introduced for timely arrival. We present the path planning methods and demon- strate their effectiveness through Monte Carlo simulations of a formation control example. Results show that the proposed algorithm consistently ensures simultaneous arrival within 0.2 seconds in all cases, with an average deviation of only 0.07 seconds, regardless of initial conditions. This approach offers an effective solution for coordinated UAV missions with fixed speed profiles.

Keywords: Path Planning, Autonomy
Abstract: In this paper, we consider the problem of inspecting a moving structure, which could be a train or a convoy, using a UAV. The moving structure is assumed to have gaps on its side to allow the UAV to enter or fly through it. Unlike earlier work in this area, since the structure is moving, the gap also moves along with it. The problem then reduces to one of a UAV trying to enter a moving window. We use the relative velocity framework to define a safe approach cone for the UAV so that its velocity vector, if directed inside this cone, will allow the UAV to pass through the window. We show that several parameters, such as the speeds of the window and the UAV, play an important part in deciding the angular span of the safe approach cone and thus have a bearing on the robustness of the guidance strategy. We establish a few theoretical results and illustrate them via simulations.

Keywords: Path Planning, Navigation
Abstract: This paper addresses the challenge of finding the shortest path in complex environments by integrating machine learning and traditional algorithms to enhance path planning techniques. The goal is to strike a balance between path length and processing time, ensuring reliable trajectories for Unmanned Aerial Vehicles. We explore four methodologies: Reinforcement Learning, Sample-Based, Geometric-Based, and Polynomial-Based Methods. Our main focus is on harnessing Reinforcement Learning for its adaptability and experiential learning capabilities in complex environments, despite its known slow convergence and high computational costs. Our proposed algorithm optimizes each step of the standard Reinforcement Learning method, Q-Learning, using classical techniques to refine its core behavior and overcome limitations. Testing in various simulated complex and unknown environments demonstrates the algorithm's efficacy in enhancing path planning efficiency and accuracy. Our approach successfully reduces path lengths by 11%, decreases flight time by 35%, and lowers processing time by 64% compared to the original Q-Learning approach.

Keywords: Path Planning, Navigation, Autonomy
Abstract: Mapping and Localization in large environments is becoming increasingly important for autonomous UAV swarms. UAV swarms solving problems in disaster response, infrastructure inspection, and agriculture rely on fresh and accurate maps to make navigation decisions. SLAM methods are capable of providing highly accurate maps through visual information, but are computationally heavy and ill-suited for UAV onboard computational profiles. For this reason, UAV swarms often dedicate one or more drones to frequent mapping, while other drones use map information for planning and trajectory generation. UAV swarms also centralize heavy-weight workloads like AI inference and SLAM map combination at the edge to extend UAV battery lives at the cost of network provisioning. Both map sharing and offloading necessitate high network bandwidth, but few SLAM or planning approaches account for this. We present NetSLAM, a network assisted SLAM and planning system that builds environmental maps and UAV trajectories that meet quality of service (QoS) requirements. NetSLAM embeds network information into SLAM maps so planning can compensate for changes in network connectivity across the environment. We also present Net*, a path planning algorithm which utilizes NetSLAM maps to build trajectories that maintain QoS requirements to maximize performance. Through real-world experiments and simulation, we show that NetSLAM maps network environments with limited additional overhead compared to existing SLAM approaches. NetSLAM improves swarm QoS by 2.35x while increasing path length by less than 14.7% compared to naive pathfinding.

Keywords: Path Planning, Navigation, Autonomy
Abstract: The Multi-Travelling Salesman Problem (mTSP) provides a fundamental mathematical framework for modelling the complexities of effective and optimised multi-UAV path planning and for developing solution strategies. Different methodologies have been studied for multi-UAV path planning, such as clustering-based techniques for waypoint allocation. Despite classical Kmeans clustering being commonly employed for its efficiency, centroid instability produces an inefficient distribution of UAVs. Traditional Genetic Algorithms (GA) often encounter difficulties with premature convergence and ineffective crossover operations, leading to suboptimal paths. This paper presents DECK-GA, a hybrid framework that combines Dynamic Centroid Kmeans (DCKmeans) clustering with Distance Efficient Genetic Algorithm (DEGA) to address centroid instability, suboptimal UAV path distribution, and premature convergence. DECK-GA applies DCKmeans to improve centroid initialisation and integration, maintaining stable cluster formations; and DEGA to enhance path planning through fitness-proportionate selection and adaptive crossover mutation, increasing diversity and accelerating convergence. DECK-GA is tested in a simulated environment using 30 and 100 randomly distributed 3D waypoints, minimising travel distances by 56.06% and 69.03%, respectively. Computation times are reduced to 28.17 and 43.21 seconds, correspondingly surpassing classical Kmeans, GA, and other six additional clustering methods combined with traditional GAs and DEGA. The enhancements show the efficiency of DECK-GA in multi-UAV waypoint clustering and path planning for the mTSP, especially in applications that require efficient global path optimisation using GNSS waypoints.

Keywords: Path Planning, Networked Swarms, Navigation
Abstract: With the growing demand for efficient logistics and warehouse management, unmanned aerial vehicles (UAVs) are emerging as a valuable complement to automated guided vehicles (AGVs). UAVs enhance efficiency by navigating dense environments and operating at varying altitudes. However, their limited flight time, battery life, and payload capacity necessitate a supporting ground station. To address these challenges, we propose HetSwarm, a heterogeneous multi-robot system that combines a UAV and a mobile ground robot for collaborative navigation in cluttered and dynamic conditions. Our approach employs an artificial potential field (APF)-based path planner for the UAV, allowing it to dynamically adjust its trajectory in real time. The ground robot follows this path while maintaining connectivity through impedance links, ensuring stable coordination. Additionally, the ground robot establishes temporal impedance links with low-height ground obstacles to avoid local collisions, as these obstacles do not interfere with the UAV's flight.

Experimental validation of HetSwarm in diverse environmental conditions demonstrated a 90% success rate across 30 test cases. The ground robot exhibited an average deviation of 45 cm near obstacles, confirming effective collision avoidance. Compared to the Conflict-Based Search (CBS) algorithm, our approach enables agents to navigate within 25 cm of obstacles, whereas CBS maintains a minimum clearance of 73 cm, highlighting our method’s efficiency in utilizing space in real-time. Extensive simulations in the Gym PyBullet environment further validated the robustness of our system for real-world applications, demonstrating its potential for dynamic, real-time task execution in cluttered environments.

Keywords: Aerial Robotic Manipulation, Multirotor Design and Control, Control Architectures
Abstract: Multicopter uncrewed aircraft systems (UAS) commonly use parallel rotors to create body-fixed thrust and torque for control, leaving these systems underactuated. Underactuation poses a significant challenge in tasks where attitude is critical, such as in collision-based aerial manipulation. Planning and control of system state at collision is required to ensure safe post-collision recovery. In particular, setting up pre-impact states such that impulses do not produce moments about mass centers can ensure recoverable departure velocities. To address the underactuated nature of UAS for collision-based aerial manipulation, this paper presents a UAS with an attached actuated pogostick. While the UAS with actuated pogostick is still underactuated, closing a control loop on the collision variables critical to managing collision response becomes possible with the new system equations. The proposed approach leverages an optimal trajectory planner coupled with a run-time controller based on partial feedback linearization of the UAS with actuated pogostick. Results show that the addition of the actuated pogostick enables setup for recoverable post-collision states when given dynamically feasible trajectories from the optimal trajectory planner.

Keywords: Aerial Robotic Manipulation, Multirotor Design and Control, Control Architectures
Abstract: This paper presents an external wrench estimator that uses a hybrid dynamics model consisting of a first-principles model and a neural network. This framework addresses one of the limitations of the state-of-the-art model-based wrench observers: the wrench estimation of these observers comprises the external wrench (e.g. collision, physical interaction, wind); in addition to residual wrench (e.g. model parameters uncertainty or unmodeled dynamics). This is a problem if these wrench estimations are to be used as wrench feedback to a force controller, for example. In the proposed framework, a neural network is combined with a first-principles model to estimate the residual dynamics arising from unmodeled dynamics and parameters uncertainties, then, the hybrid trained model is used to estimate the external wrench, leading to a wrench estimation that has smaller contributions from the residual dynamics, and affected more by the external wrench. This method is validated with numerical simulations of an aerial robot in different flying scenarios and different types of residual dynamics, and the statistical analysis of the results shows that the wrench estimation error has improved significantly compared to a model-based wrench observer using only a first-principles model.

Keywords: Aerial Robotic Manipulation, Multirotor Design and Control, Control Architectures
Abstract: Performing tasks at high altitudes can be inconvenient and unsafe for humans. Unmanned aerial vehicles (UAVs) with physical interaction capabilities are on hand to address these issues. Our previous work introduced one such UAV: a multi-rotor with a pair of tilting rotors and a novel, cable-driven, front-mounted gripper, which improved position accuracy during interaction tasks. However, the UAV could only interact with vertical surfaces, and its control performance was limited by a simple pilot-assisted position controller during interaction. This manuscript advances that work by developing an improved version of our UAV. The modified design includes two pairs of tilt-rotors, which can tilt simultaneously to angle the drone body, allowing interaction with non-vertical targets at specific angles. We develop a high-fidelity simulation package that accurately replicates our new UAV design's physical characteristics and dynamics, including the cable-driven gripper. This simulation package provides a virtual testbed for designing and evaluating advanced interaction control algorithms, minimizing the risks, costs, and time of physical prototyping. We demonstrate its utility by safely testing autonomous control strategies, including force control, in a tree-grasping scenario. We also give insights into hyperparameter selections, challenges faced, and current limitations while developing such a versatile package that involves simulating elastic components such as cables, springs, and soft finger pads. We open-source our simulation package for the community's benefit at https://tinyurl.com/ypwzje2a.

Keywords: Aerial Robotic Manipulation, Multirotor Design and Control, Perception and Cognition
Abstract: Fully actuated aerial robots have shown superiority in Aerial Physical Interaction (APhI) in recent years. This work presents a minimal setup for aerial telemanipulation, improving accessibility to such technologies. The design and control of a 6-Degrees of Freedoms (DoF) joystick with 4-(DoF) haptic feedback are detailed. It is the first haptic device with standard Remote Controller (RC) form factor for APhI. Miniaturizing the haptic device adds sense of touch to RC, enhancing physical awareness. The goal is to provide operators with an extra sense—beyond vision and sound—to support safe (APhI). To the best of the authors' knowledge, this is the first 6-DoF aerial teleoperation system capable of decoupling single-axis input commands. The proposed robot hardware design reduces number of components, aiming for easier maintenance and improved force and thrust-to-weight ratios. Open-source physics-based simulation and successful early flight tests highlight the tool’s promise for future APhI applications.

Keywords: Aerial Robotic Manipulation, UAS Applications, Multirotor Design and Control
Abstract: As aerial robots gain traction in industrial applications, there is growing interest in enhancing their physical interaction capabilities. Pushing tasks performed by aerial manipulators have been successfully demonstrated in contact-based inspections. However, more complex industrial applications require these systems to support higher-DoF (Degree of Freedom) manipulators and generate larger forces while pushing (e.g., drilling, grinding). This paper builds on our previous work, where we introduced an aerial vehicle that can dynamically vary its CoM (Center of Mass) location to improve force exertion during interactions. We propose a novel approach to further enhance this system’s force generation by optimizing its CoM location during interactions. Additionally, we study the case of this aerial vehicle equipped with a 2-DoF manipulation arm to extend the system’s functionality in tool-based tasks. The effectiveness of the proposed methods is validated through simulations, demonstrating the potential of this system for advanced aerial manipulation in practical settings.

Keywords: Multirotor Design and Control, Aerial Robotic Manipulation, Technology Challenges
Abstract: The cutting-edge contact-aware ability of aerial platforms has opened new frontiers in aerial robotics, enabling applications beyond traditional contact-free operations. Performing in-contact tasks and smoothly transitioning between navigation and interaction phases introduce significant challenges, especially in complex scenarios. This paper focuses on multi-rotors designed for physical interaction tasks, reviewing the most used aerial platforms, interaction tools, and control methodologies tailored for contact-aware applications. Special attention is given to the contact detection phase, which bridges the gap between contact-free and in-contact operational phases, ensuring precise and safe engagement with the target.

Keywords: Integration, Technology Challenges, UAS Testbeds
Abstract: Unmanned multi-domain robotic systems are rapidly advancing, with many innovative platforms designed to operate across multiple environmental domains, typically involving the combination of air and either land, water-surface, or underwater capabilities. These systems are designed to seamlessly transition between and operate effectively in these diverse environments, opening up new possibilities for numerous fields - including military, research, search and rescue, and commercial applications. Categorization of the various forms of technology utilized in each innovative design is challenging due to the varying descriptions and definitions within each source publication. This paper is written to describe the current state of physical design for unmanned multi-domain robotic platforms, as well as standardize the definitions and comparatively categorize the design features, propulsion methods, and domain-transition capabilities of many hybrid systems. Descriptions of the terminology used are provided throughout this article as each bi-domain system category is introduced. A complete comparative table of all findings is provided near the end of this manuscript, complete with respective categories, design features, and domain-transition details.

Keywords: Networked Swarms, Control Architectures, UAS Testbeds
Abstract: This paper investigates the resilience of perception-based multi-robot coordination with wireless communication to online adversarial perception. A systematic study of this problem is essential for many safety-critical robotic applications that rely on the measurements from learned perception modules. We consider a (small) team of quadrotor robots that rely only on an Inertial Measurement Unit (IMU) and the visual data measurements obtained from a learned multi-task perception module (e.g., object detection) for downstream tasks, including relative localization and coordination. We focus on a class of adversarial perception attacks that cause misclassification, mislocalization, and latency. We propose that the effects of adversarial misclassification and mislocalization can be modeled as sporadic (intermittent) and spurious measurement data for the downstream tasks. To address this, we present a framework for resilience analysis of multi-robot coordination with adversarial measurements. The framework integrates data from Visual-Inertial Odometry (VIO) and the learned perception model for robust relative localization and state estimation in the presence of adversarially sporadic and spurious measurements. The framework allows for quantifying the degradation in system observability and stability in relation to the success rate of adversarial perception. Finally, experimental results on a multi-robot platform demonstrate the real-world applicability of our methodology for resource-constrained robotic platforms.

Keywords: Payloads, Simulation, UAS Testbeds
Abstract: As uncrewed aerial systems continue to grow in popularity and importance, the long-term and scalable use of these systems for remote sensing and imagery data collection remains a valuable and achievable goal. To enable these systems at scale, real-time onboard imagery processing is required. To determine the feasibility of real-time remote sensing systems, many factors must be accounted for, including the ability of the sensing and processing algorithms to operate on and collect data from real-world scenes and deliver actionable intelligence to the data consumer. In this paper, a holistic simulation system based on ROS 2 and Gazebo is presented, which allows for real-time processing algorithms to be tested and proven for flight in an accurate and extensible way. By using ROS 2 and USU AggieAir's STARDOS platform, it is possible to show how the remote sensing system and onboard, real-time processing algorithms are applicable to the aerial remote sensing task (i.e. it can demonstrate feasibility for physical deployment based on accurate simulated data processing).

Keywords: Simulation, Integration, Technology Challenges
Abstract: Rapid integration and analysis of dynamic data are essential for effective disaster response, as timely insights can significantly impact decision-making and resource allocation. Unmanned aerial vehicles (UAVs) enhance situational awareness by providing real-time aerial surveillance to first responders. This work proposes the AIDERS platform, a multi-UAV platform designed and developed to support real and simulated UAV operations for disaster management. The platform enables collaborative autonomous UAV operations, allowing multiple UAVs to navigate and survey an area while streaming live video feeds for real-time detection of people, objects, and disaster-related damage. Additionally, its simulation capabilities enable extensive testing and validation before real-world deployment. This work demonstrates the robustness of the AIDERS platform through experiments with simulated swarms of up to eight UAVs.

Keywords: Simulation, Perception and Cognition, UAS Testbeds
Abstract: This article presents an approach for recreating experimental Unmanned Aerial Vehicle (UAV) flight in the state-of-the-art 3D simulation software. Through the use of the Unreal Engine, AirSim simulator, and the Cesium for Unreal plugin with Google Maps, we demonstrate replicating an experiment of a real-world flight in the digital twin environment of the same location. Work investigates the viability of replicating real-world experiments by assessing the similarity between the experimental results of the real-world and digital twin experiments. The experiments involve analyzing the image telemetry and map generated of the real-world and digital twin images using the Direct Sparse Odometry (DSO) algorithm. The results have shown that replicating the real-world experiment in the digital environment produces similar results to those seen in the real-world. This will allow researchers to explore the impact of sensor, vehicle, and algorithm parameters in a controlled, repeatable environment before real-world deployment.

Keywords: UAS Applications
Abstract: Corrosion is one of the most common defects in steel infrastructures. Crewed visual inspection is the conventional method for corrosion detection in civil infrastructure, which can be dangerous, inconsistent, and labor-intensive. These limitations encouraged the researchers to explore the feasibility of using the Uncrewed Aerial System (UAS) for autonomous real-time corrosion detection using artificial intelligence. A human-machine interface was implemented to take inspector input in sequential training of a corrosion detection model on visual imagery. The model was trained to detect corrosion using smartphone images. The model output was validated or corrected by the inspector after each inspection. The model was then retrained on the inspector output and used in the next inspection. Each dataset consisted of 225 x 225 pixels image tiles labeled as with corrosion and without corrosion. Six combinations of UAS inspection datasets were used to evaluate how the deep learning model performance changed in terms of true positive rate (TPR), true negative rate (TNR), false positive rate (FPR) and false negative rate (FNR). Before retraining 84.78% of images with corrosion were correctly predicted by the model, and the TNR value was 82.75%. The result showed that the adapted deep learning model performance improved with more inspection, as expected. In particular, the number of reported false calls made by the model reduced. While the improvement was not always tangible due to corrosion image diversity in texture and color, however, the same amount of training, regardless of their order, led to improved but comparable performance.

Keywords: UAS Applications
Abstract: Bridges are vital infrastructure assets whose maintenance is essential to ensure safety and efficient traffic flow. However, due to their nature, they are often located in hard-to-reach places, which makes their regular inspection challenging and risky for human operators. This paper presents a comprehensive multi-UAV (Unmanned Aerial Vehicle) framework for fast and efficient cooperative bridge inspection, leveraging commercial UAVs with open-source tools and integrating both a custom Ground Control Station and advanced multi-UAV motion planning based on Signal Temporal Logic. The proposed approach enables autonomous and safe data collection while minimizing operational constraints and human intervention, making it a valuable contribution to UAV-based infrastructure monitoring. The presented framework has been validated in a real-world environment, showcasing its effectiveness in coordinating UAV teams for autonomous structural inspections.

Keywords: UAS Applications, Control Architectures
Abstract: In this paper, the synthesis of a robust trajectory tracking controller for a UAV carrying a slung load is presented using hierarchical sliding mode and super-twisting methodologies. A second-order sliding mode disturbance observer is used in conjunction with the sliding mode controller. Since the slung load system is underactuated, the sliding mode closed-loop dynamics is analyzed to derive conditions for the sliding surface parameters that guarantee stability. Lyapunov analysis is used to prove the stability of the system. A controllability analysis is also carried out. Simulation results are presented for different cases to demonstrate the performance of the proposed controller.

Keywords: UAS Applications, Control Architectures, Path Planning
Abstract: This paper presents the design and implementation of a system to automate the task of inspection of a physical structure, surrounded by one of three boundary models provided (plane, box or cylinder), using a DJI quadrotor. First, the system plans the flight trajectory of the quadrotor accordingly, to ensure that the whole surface’s model is covered in the inspection, and then executes the flight autonomously. The system is made up of two main software components: a ROS2 robotics application and Android application. The ROS2-based application is used to plan the flight path and control the quadrotor to follow the path, and the Android application to allow communication with the quadrotor using the DJI Mobile SDK V4. The system was tested using the Gazebo simulator and the DJI Assistant 2 simulator to ensure correct functionality. Finally, it was tested in experimental scenarios by flying a DJI Mavic Pro quadrotor with effective results.

Keywords: UAS Applications, Integration, Navigation
Abstract: Given the pivotal role of the railroad industry in modern transportation and the potential risks associated with track malfunctions, the inspection and maintenance of railroad tracks emerges as a critical concern. Existing solutions rely on large, expensive, and time-consuming platforms that are very accurate, however, they require the line to be blocked during the inspection. The use of Unmanned Aerial Vehicles (UAVs) can significantly reduce track downtime and cost while maintaining inspection capabilities. However, current solutions focus on the inspection task while UAVs are programmed to follow predefined paths on the network. This paper presents an autonomous, vision-based track following system that was developed, implemented, and tested onboard a UAV. Notably, this system operates independently of external sensors, such as GPS, thanks to its utilization of advanced computer vision techniques. Two approaches were developed utilizing a forward-facing camera and a downward-facing camera. The experimental results of several field trials show the efficiency of the developed system.

Keywords: Aerial Robotic Manipulation, Autonomy, Integration
Abstract: The integration of Large Language Models (LLMs) into robotic control, including drones, has the potential to revolutionize autonomous systems. Research studies have demonstrated that LLMs can be leveraged to support robotic operations. However, when facing tasks with complex reasoning, concerns and challenges are raised about the reliability of solutions produced by LLMs. In this paper, we propose a prompt framework with enhanced reasoning to enable reliable LLM-driven control for drones. Our framework consists of novel technical components designed using Guidelines, Skill APIs, Constraints, and Examples, namely GSCE. GSCE is featured by its reliable and constraint-compliant code generation. We performed thorough experiments using GSCE for the control of drones with a wide range of task complexities. Our experiment results demonstrate that GSCE can significantly improve task success rates and completeness compared to baseline approaches, highlighting its potential for reliable LLM-driven autonomous drone systems.

Keywords: Autonomy, Navigation, Integration
Abstract: This paper presents a decentralized graph-based exploration and inspection framework for multi-robot systems, designed to address challenges in subterranean and large-scale environments. Unlike prior works that focus solely on exploration or inspection, this framework integrates volumetric exploration, semantic inspection, and dynamic task allocation into a unified decentralized system. A key novelty of this work is the seamless integration of these modules in a multi-robot setting, allowing UAVs to autonomously coordinate their tasks without relying on centralized control. The framework employs a hierarchical graph structure, utilizing a dense local graph for immediate navigation and a sparse global graph for long-term planning and repositioning. Extensive simulations in large-scale complex-shaped environments demonstrate that the proposed approach improves the completeness of the generated maps, reduces inconsistencies in the constructed mesh, and accelerates the overall exploration-inspection process compared to existing decentralized strategies.

Keywords: Autonomy, Reliability of UAS, Integration
Abstract: The rapid growth of the drone market, driven by cost-effective computing and sensor advancements, has expanded applications in safety, security, agriculture, logistics, and infrastructure inspection. The introduction of EU drone regulations in 2020 has enabled Beyond Visual Line-of-Sight (BVLOS) and autonomous operations, opening opportunities for long-term unmanned missions. However, commercial drone systems remain reliant on Remote Piloted Aircraft Systems (RPAS) and predefined waypoint navigation, while autonomous operations are often confined to research settings.

In this work, we present The BEAST framework, a modular open-source framework for long-term, robust, and autonomous BVLOS operations. This framework is designed to integrate intelligent drones into real-world environments while ensuring regulatory compliance. The BEAST framework includes essential components such as obstacle avoidance, intelligent fail-safe mechanisms, reliable communication, and a weather-proof docking station. Unlike previous approaches that address isolated BVLOS challenges, The BEAST provides a comprehensive, end-to-end solution encompassing mission planning, autonomous navigation, and operational safety. The framework has been implemented and validated across multiple drone platforms in safety and security applications.

Keywords: Control Architectures, Training, Autonomy
Abstract: In this research, Koopman operator theory is employed to achieve faster training time and improved performance of a reinforcement learning (RL) based linear quadratic controller (LQ). The proposed methodology, called K-RL-LQ, is implemented for the trajectory tracking problem of a quadrotor UAV. Using the evolution of analytically derived Koopman generalized eigenfunctions allows for the embedding of quadrotor nonlinear dynamics into a quasi-linear model. Specifically, the resulting Koopman based quadrotor dynamics has linear state matrix and state dependent control matrix. Additionally, the obtained formulation is fully actuated, hence, compared to traditional model based hierarchical control the advantages are twofold: i) the controller can be formulated using linear control strategies in Koopman formulation which will result in a nonlinear control law in the original state space; ii) the trajectory tracking task can be achieved through a single control loop. Using this formulation, an RL agent is trained to estimate the controller parameters of a linear quadratic control law. Notably, it is shown that, using a reward function and observation space based on Koopman generalized eigenfunctions over the state space, leads to a considerably faster training time and improved overall performances.

Keywords: Security, Simulation, UAS Testbeds
Abstract: Unmanned aerial vehicles (UAVs) are increasingly adopted in various applications due to their high mobility and advanced sensing capabilities. However, they also face significant security threats and reliability concerns arising from external adversarial attacks and internal system failures. AI and machine learning techniques have shown promise in detecting security threats and anomalies in UAVs, but their effectiveness heavily depends on high-quality UAV security datasets for training. In this paper, we present an open-source simulation platform designed to model diverse UAV security scenarios. Our platform offers flexible customization of attacking effects on major UAV components, including onboard sensors, communication systems, vision modules, and flight control. Additionally, it provides rapid generation and collection of UAV system data under adversarial conditions, facilitating intelligent cybersecurity and reliability analysis. Our experiments successfully simulated over 30 attacking effects toward UAVs, demonstrating our platform’s capability to support extensive UAV security research.

Keywords: Perception and Cognition, UAS Communications, Autonomy
Abstract: Gesture recognition presents a promising avenue for interfacing with unmanned aerial vehicles (UAVs) due to its intuitive nature and potential for precise interaction. This research conducts a comprehensive comparative analysis of vision-based hand gesture detection methodologies tailored for Edge-Assisted UAV Control. The existing gesture recognition approaches involving cropping, zooming, and color-based segmentation, do not work well for this kind of applications in dynamic conditions and suffer in performance with increasing distance and environmental noises. We propose to use a novel approach leveraging hand landmarks drawing and classification for gesture recognition based UAV control. With experimental results we show that our proposed method outperforms the other existing methods in terms of accuracy, noise resilience, and efficacy across varying distances, thus providing robust control decisions. However, implementing the deep learning based compute intensive gesture recognition algorithms on the UAV’s onboard computer is significantly challenging in terms of performance. Hence, we propose to use a edge-computing based framework to offload the heavier computing tasks, thus achieving closed-loop real-time performance. With implementation over AirSim simulator as well as over a real-world UAV, we show the advantage of our end-to-end gesture recognition based UAV control system.

Keywords: UAS Applications, Swarms, UAS Communications
Abstract: Multi-UAV systems tend to require complex infrastructure to deploy in real-world scenarios, limiting their accessibility and scalability. In addition, current research often relies on custom solutions or proprietary hardware to facilitate inter-UAV communication. In this paper, we propose an open-source, low-cost, plug-and-play solution for enabling decentralized UAV-to-UAV communication over 2.4GHz Wi-Fi using a connection-less protocol. Our approach simplifies the deployment of decentralized systems by allowing UAVs to easily exchange any type of binary data, seamlessly interfacing with ROS2. The solution uses an ad-hoc style network that allows UAVs to join or leave dynamically without requiring centralized governance or a priori configuration. We describe the architecture of the system, assess the network performance in an outdoor environment using UAVs, and the system's ability to share information as a swarm through Hardware-in-the-loop (HITL) and experiments using UAVs. Our results show that the proposed system facilitates connectivity and is able to transmit mission-critical data for real-world UAV operations. HITL experiments show that a decentralized planning algorithm running on three simulated UAVs can effectively reach consensus on decentralized task allocation. We have made the code public and thus provide a viable solution for researchers seeking to implement decentralized UAV swarms using cost-effective Commercial Off The Shelf (COTS) hardware and minimal infrastructure.

Keywords: UAS Communications, Networked Swarms, Swarms
Abstract: Unmanned Aerial Vehicles (UAVs) have been increasingly applied in different scenarios, requiring efficient communication networks to handle swarm operations that rely on dynamic ad-hoc infrastructures without fixed support. Mesh architecture, with its ability to offer multipath communications, emerges as a solution to overcome the challenges imposed by high mobility and frequent topology changes. This study, carried out in a real environment with Raspberry Pi devices, compared the routing protocols OLSR, BATMAN-ADV, and Babel, justifying their choices based on the observed performance: OLSR demonstrated greater stability and efficiency for variable traffic loads due to its rapid route adaptation. Babel stood out in highly-mobility scenarios because it presented lower latency, attributed to its agile information update. At the same time, BATMAN-ADV, although efficient in certain conditions, showed greater resource consumption and instability under heavy traffic. The main contribution of this article lies in the detailed comparative analysis of the three protocols in a real environment, analyzing the practical characteristics for selecting the most appropriate routing protocol according to the specific requirements of each application scenario.

Keywords: UAS Communications, UAS Applications, Energy Efficient UAS
Abstract: In numerous applications, such as multi-drone surveillance and search-and-rescue missions, the deployment of multiple robots is essential to accomplish several tasks simultaneously. As the vehicles have limited communication range, it is essential to have a decentralized task allocation algorithm to allocate tasks to robots effectively. One such algorithm is the consensus-based bundle adjustment (CBBA) algorithm, which has shown promise in working with multi-robots and has theoretical guarantees. However, CBBA requires communication at every instance, which can cause communication congestion and packet dropouts that lead towards performance degradation. In this paper, we propose an event-driven communication mechanism to overcome communication issues while retaining the theoretical properties of CBBA in terms of convergence and performance bounds. Theoretically, we show that the solution quality remains the same as that of CBBA and validate through Monte-Carlo simulations for varying numbers of targets, agents and bundles. The results show that the proposed algorithm (ED-CBBA) achieves up to 52% reduction in the number of messages transmitted.

Keywords: Navigation, Path Planning
Abstract: This paper presents a framework that integrates a distance-aware 3D local path planning algorithm based on Euclidean Signed Distance Fields (ESDFs) with a system that trains a Sinusoidal Representation neural network (SIREN) using the HIO-SDF network structure. The main contribution of the paper is a software framework that incorporates online generated ESDF into local planners for efficient and safe 3D path planning by leveraging the ESDF properties. The framework includes a neural network that can be used by the local planner as an up-to-date representation of the environment. Experimental validation shows favorable results in exploiting the intrinsic characteristics of online ESDFs and acknowledges this framework as a feasible method to perform local path computation. Planner code is available at: https://github.com/robotics-upo/neural_esdf_local