Autonomous Unmanned Aerial Vehicle System Research Articles

Deep learning-based obstacle-avoiding autonomous UAVs with fiducial marker-based localization for structural health monitoring.

This paper proposes a framework for obstacle-avoiding autonomous unmanned aerial vehicle (UAV) systems with a new obstacle avoidance method (OAM) and localization method for autonomous UAVs for structural health monitoring (SHM) in GPS-denied areas. There are high possibilities of obstacles in the planned trajectory of autonomous UAVs used for monitoring purposes. A traditional UAV localization method with an ultrasonic beacon is limited to the scope of the monitoring and vulnerable to both depleted battery and environmental electromagnetic fields. To overcome these critical problems, a deep learning-based OAM with the integration of You Only Look Once version 3 (YOLOv3) and a fiducial marker-based UAV localization method are proposed. These new obstacle avoidance and localization methods are integrated with a real-time damage segmentation method as an autonomous UAV system for SHM. In indoor testing and outdoor tests in a large parking structure, the proposed methods showed superior performances in obstacle avoidance and UAV localization compared to traditional approaches.

A novel metaheuristics with adaptive neuro-fuzzy inference system for decision making on autonomous unmanned aerial vehicle systems

Recently, autonomous systems have received considerable attention amongst research communities and academicians. Unmanned aerial vehicles (UAVs) find useful in several applications like transportation, surveillance, disaster management, and wildlife monitoring. One of the important issues in the UAV system is energy efficiency, which can be resolved by the use of clustering approaches. In addition, high resolution remote sensing images need to be classified for effective decision making using deep learning (DL) models. Though several models are available in the literature, only few approaches have focused on the clustering and classification processes in UAV networks. In this aspect, this paper designs a novel metaheuristic with an adaptive neuro-fuzzy inference system for decision making named MANFIS-DM technique on autonomous UAV systems. The proposed MANFIS-DM technique intends to effectively organize the UAV networks into clusters and then classify the images into appropriate class labels. The proposed MANFIS-DM technique encompasses two major stages namely quantum different evolution based clustering (QDE-C) technique and ANFIS based classification technique. Primarily, the QDE-C technique involves the design of a fitness function involving three parameters namely average distance, distance to UAVs, and UAV degree. Besides, the image classification model involves a set of subprocesses namely DenseNet based feature extraction, Adadelta based hyperparameter optimization, and ANFIS based classification. The design of QDE-C algorithm with classification model for autonomous UAV systems show the novelty of the work. The experimental result analysis of the MANFIS-DM method is carried out against benchmark dataset and the results ensured the enhanced performance of the MANFIS-DM technique over the other methods with the maximum accuy of 99.13%.

An autonomous unmanned aerial vehicle system for fast exploration of large complex indoor environments

AbstractThis paper introduces an autonomous system employing multirotor unmanned aerial vehicles for fast 3D exploration and inspection of vast, unknown, dynamic, and complex environments containing large open spaces as well as narrow passages. The system exploits the advantage of small‐size aerial vehicles capable of carrying all necessary sensors and computational power while providing full autonomy and mobility in constrained unknown environments. Particular emphasis is put on the robustness of the algorithms with respect to challenging real‐world conditions and the real‐time performance of all algorithms that enable fast reactions to changes in environment and thus also provide effective use of limited flight time. The system presented here was employed as a part of a heterogeneous ground and aerial system in the modeled Search & Rescue scenario in an unfinished nuclear power plant during the Urban Circuit of the Subterranean Challenge (SubT Challenge) organized by the Defense Advanced Research Projects Agency. The main goal of this simulated disastrous scenario is to autonomously explore and precisely localize specified objects in a completely unknown environment and to report their position before the end of the mission. The proposed system was part of the multirobot team that finished in third place overall and in first place among the self‐funded teams. The proposed unmanned aerial vehicle system outperformed all aerial systems participating in the SubT Challenge with respect to versatility, and it was also the self‐deployable autonomous aerial system that explored the largest part of the environment.

Implementation and evaluation of an autonomous airborne ultrasound inspection system

ABSTRACT The mobility of an Unmanned Aerial Vehicle (UAV) offers significant benefits when deploying remote Non-Destructive Testing (NDT) inspections of large-scale assets. Ultrasonic inspection is primarily a contact-based NDT method, that grants the opportunity to remotely monitor the structural health of an industrial asset with enhanced internal integrity information. Presented in this paper is an implementation of an autonomous UAV system, equipped with an ultrasonic thickness measurement payload. This system is designed to conduct ultrasonic inspections of non-magnetic facilities and industrial infrastructure where surface adhesion cannot be achieved magnetically. Operating within a laboratory environment, this system autonomously positioned the transducer on a vertically mounted, unpainted, aluminium sample and completed an ultrasonic thickness measurement without manual intervention. An onboard laser scanner provided instantaneous UAV alignment and standoff error measurements versus the sample’s surface normal vector. While inspecting a region of the aluminium sample with 12.92 mm nominal thickness, the UAV system demonstrated a measurement error of 0.03 mm. During this process, the standard deviation of the craft’s positional error was recorded to be below 63.26 mm, accompanied by an angular alignment error versus the surface normal vector of below 2.71°. The accuracy of the UAV deployed inspection, including thickness measurement accuracy and positional accuracy, depends on many factors. As such, transducer alignment constraints, electrical noise and UAV stability are investigated and discussed. Findings from this paper may be taken to inform future research regarding autonomous airborne ultrasonic inspection of constructed infrastructure and industrial facilities.

Fluorescence‐based detection of field targets using an autonomous unmanned aerial vehicle system

Abstract Here we describe a proof‐of‐concept autonomous unmanned aerial vehicle (UAV) system that utilizes the fluorescence characteristics unique to different materials to scan and acquire targets in the field that includes fossils, rocks and minerals, organisms and archaeological artefacts. This is possible because these targets are often highly fluorescent against lower fluorescence backgrounds and may exhibit different colours. To detect these targets from a moving UAV, we utilize laser‐stimulated fluorescence. This involves an intense laser beam that—unlike regular UV light—can project greater distances and generate sufficient fluorescence for targets to be detected on the UAV camera many metres above the ground. The system involves a lightweight UAV programmed to fly a waypoint pattern autonomously at night over the area of interest. LIDAR maintains its height above the terrain. A near‐UV laser is projected across the ground as a horizontal line directly below the UAV. Co‐mounted with the laser is a small highly sensitive video camera stabilized on a motion‐controlled gimbal that records the laser line during the flight. An intermittent, powerful white light strobe flashes during the flight to record the UAV's ground position in the scanned area. The UAV returns with the laser video at the end of each autonomous mission. This video is post‐processed, extracting the laser line data into a long continuous scan image showing the fluorescing ground targets. Initial analysis determines what colour the targets fluoresce so that a specific colour range can be extracted from the image to identify the locations of the detected targets. The white light strobe images are then used to quickly follow‐up on the detections. This system holds the promise of becoming the lowest ‘ground truth’ layer in the mix of high‐altitude map data produced by satellite‐ and airplane‐based Geographic Information Systems. With centimetre resolution and the geochemical differences shown via fluorescence, this system will increase the scale and efficiency of data collection involving fossils, rocks and minerals including mineable materials, fluorescent organisms including biogenic mineral producers like shellfish as well as archaeological artefacts. Thus, the fields of evolution, ecology, Earth and planetary science, archaeology and other subjects involving fluorescent targets would all benefit from this new system.

Quantifying impacts on remote photogrammetric inspection using unmanned aerial vehicles

Remote photogrammetric inspection is a Non-Destructive Testing method used to quantify surface integrity and detect external discontinuities. The mobility and size of an unmanned aerial vehicle (UAV) offer the flexibility to quickly deploy remote photogrammetric inspections for large-scale assets. In this paper, the results of a photogrammetric inspection are presented as a 3D profile, reconstructed from UAV captured images. Experiments were conducted indoors using a wind turbine blade section obtained from a recently decommissioned asset. The naturally occurring surface features representative of environmental wear were augmented with a small number of artificial features to aid in the visualisation of inspection quality. An autonomous UAV system for photogrammetric inspections is demonstrated and the influence of image parameters such as environmental light levels, motion blur and focal blur quantified in terms of their impact on the inspection accuracy. Over the range of parameter values studied, the poorest scenario was observed to cause a degradation in reconstruction error by a factor of 13 versus the optimal. Reconstruction quality when employing a laser range scanner to maintain standoff distance relative to the object during flight was also investigated. In this schema, the controller automatically generated a real-time adaptive flight path to follow the outer profile of the wind turbine blade and, consequently, demonstrated improved image quality during close-range inspection of an object with complex geometry. Inspection accuracy was quantified using the error of the photogrammetric reconstruction as compared to a model acquired using independent metrology equipment. While utilising the laser-based adaptive path, error in the reconstructed geometry was reduced by a factor of 2.7 versus a precomputed circular path. In the best case, the mean deviation was below 0.25 mm. Instances of wind turbine blade damage such as edge crushing, surface imperfections, early stage leading edge erosion were clearly observed in the textured 3D reconstruction profiles, indicating the utility of the successful inspection process. The results of this paper evaluate the impact of optical environmental effects on photogrammetric inspection accuracy, offering practical insight towards mitigation of negative effects.

A compact and high gain substrate integrated horn array antenna system

AbstractIn this paper, a novel compact, lightweight and high gain array antennae system is demonstrated which is based on substrate integrated dielectric filled waveguide concepts. It is a planar design which can be easily fabricated with a common two‐layer substrate. The overall goal was to have a high gain so that this array antenna could be potentially used in autonomous unmanned aerial vehicle systems for collision avoidance. An initial model was designed, simulated and fabricated and, upon evaluation, was further modified to achieve a higher performance specification. Additionally, the design can be easily and inexpensively fabricated using standard substrates with a rapid prototyping platform. A two element array resulted in a simulated gain of 9.6 dB with a radiation pattern which had a maximum lobe between a broadside and an end‐fire. This was with a low loss substrate such as Rogers RT duroid 5880. For the design, a low dissipation factor material is preferable as it minimizes losses and also increases compactness of the entire antenna array.

Design and Analysis of Modular, Portable, Rapid Deployment Autonomous Surveillance UAV System

Abstract — In open field operations, such as in the isolated mountainous area, the need of small, compact, and portable surveillance system is inevitable. It is to give the team a surveillance capability while reducing carried-load significantly. This system is a portable mini unmanned aerial vehicle (UAV) with its mobile Ground Control Systems included. Within this paper, the general design processes are presented from determining DRO, initial sizing, aerodynamics analysis, structure, mechanism, backpack design until flight testing to validate the design. Some simple simulations by XFLR5 and complex simulation CFD are also conducted to predict the aerodynamics characteristic of UAV. Outdoor flight testing has been conducted and also still on progress for further results.

Autonomous UAV System Development for Payload Dropping Mission

Abstract — In this paper, power and autonomous system in UAV (Unmanned Aerial Vehicle) for payload dropping mission will be presented. The UAV consists of two main parts, airframe and the avionics system. There are two important subsystems of avionics system which will be discussed, power system and autonomous system. There are three Li-Poly batteries applied in this system to power the autonomous system. The autonomous system consists of an Ardupilot Mega v2.4.2 board (autonomous board), a GPS (Global Positioning System) sensor, attitude sensors, ESC (Electronic Speed Controller), servos, a brushless motor, a camera, a data transmitter, an Audio/Video transmitter, and also a Radio RC receiver. This UAV was designed to participate Kontes Robot Terbang Indonesia 2013 (Indonesia Aerial Robot Contest 2013), payload dropping category, at Institut Teknologi Bandung, Indonesia. In this competition, provided a mission which the UAV has to drop payloads in some target points which is only known when UAV in the flight time. The UAV has to fly autonomously point-to-point while it also search the true point to drop the payloads by camera capturing. By integrating UAV’s system with GCS (Ground Control System), the true point will be known by the operators in the ground control station and the mission can be accomplished. Keywords —airframe, power system, autonomous system, Kontes Robot Terbang Indonesia 2013 , dropping payloads.

Design and Structural Analysis for an Autonomous UAV System Consisting of Slave MAVs with Obstacle Detection Capability Guided by a Master UAV Using Swarm Control

An Unmanned Aerial Vehicle (UAV) is an aircraft without a human pilot. It can either be controlled manually by a pilot on the ground using a trans-receiver or it can be programmed to operate autonomously. In this proposed control system, multiple slave Micro Aerial Vehicles(MAV) are dispatched from a master UAV for surveillance. All the MAVs are synchronized with each other through the master UAV which highlights their purpose and position. The master UAV acts as a mobile base for the surveillance, it stores the data collected by the MAVs and transmits them to a remote base. A design of the UAV-MAV system and its performance analysis is presented.

Shared Control : Balancing Autonomy and Human Assistance with a Group of Quadrotor UAVs

Robustness and flexibility constitute the main advantages of multiple-robot systems with respect to single-robot ones as per the recent literature. The use of multiple unmanned aerial vehicles (UAVs) combines these benefits with the agility and pervasiveness of aerial platforms [1], [2]. The degree of autonomy of the multi-UAV system should be tuned according to the specificities of the situation under consideration. For regular missions, fully autonomous UAV systems are often appropriate, but, in general, the use of semiautonomous groups of UAVs, supervised or partially controlled by one or more human operators, is the only viable solution to deal with the complexity and unpredictability of real-world scenarios as in, e.g., the case of search and rescue missions or exploration of large/cluttered environments [3]. In addition, the human presence is also mandatory for taking the responsibility of critical decisions in high-risk situations [4].