No description available.

No description available.

The increased demand for fast deliveries of goods have led to more costly and less environmentally friendly transports since many of the delivering trucks are not being fully loaded. The inefficiencies of deliveries have created a need for development of new freight systems. One alternative vehicle that has gained increasingly interest is usage of UAVs (unmanned arial vehicles), also known as drones. Several drones in varying sizes and configurations are being developed and applications within transports of both people and goods is seen as promising areas for the future. The study aims at investigating the performance of drone deliveries regarding time, cost, and environmental impact and to see what parameters are important for the performance. This have been made by comparing a UAV to two different vans (electric and HVO) for parcel deliveries in four chosen missions in both urban and rural settings. The evaluation takes a future perspective and are based on information received through both literature review and a market investigation.  The result from this study indicates that UAVs are likely to be a competitive future option for parcel deliveries regarding time and cost. This is concluded since the results shows significant savings in both costs and delivery time and these results are not changed by the sensitivity analyses. The result regarding environmental performance shows that the UAVs competitiveness depends on the vehicle of comparison. The drone has a better environmental performance than vans with fossil-based propellants but given the energy intensity of the UAV, it is not favourable compared to an electric van. The energy requirement of the drone is one of the most important factors affecting the performance. The energy requirements per km for the UAV increases when the routes become shorter since different phases of the flight have different energy intensity. The most demanding phase is lifting and when the distance between the stops is reduced this phase becomes more prominent. Another important factor is the possibility to reduce the travelled distance by taking the straight path with the UAV compared to being bound by the road infrastructure. The shorter distance for the UAV contributes both to reduced time but also reduced energy requirements which in its turn affect both environmental and economic performance. The distances and energy requirements are thus not the most important factor for the economic sustainability but rather the cost of staff. Since the drones are unmanned, several UAVs could be controlled by the same operator contributing to reduced cost of staff. The low energy requirements for the UAV in the longer and more rural cases makes this type of applications the most beneficial regarding environmental performance. Urban missions are instead the most preferable regarding cost and time, since a bigger share of the distance can be saved and the difference in speed between the UAV and the van is larger.

As an extension of previous work done by Luke Newmeyer in his master's thesis \cite{newmeyer2018efficient}, this report presents an improved signal processing chain for efficient, real-time processing of radar data for small-scale UAV traffic control systems. The HDL design described is for a 16-channel, 2-dimensional phased array feed processing chain and includes mean subtraction, windowing, FIR filtering, decimation, spectral estimation via FFT, cross-correlation, and averaging, as well as a significant amount of control and configuration logic. The design runs near the the max allowable memory bus frequency at 300MHz, and using AXI DMA engines can achieve throughput of 38.3 Gb/s (~0.25% below theoretical 38.4 Gb/s), transferring 2MB of correlation data in about 440us. This allows for a pulse repetition frequency of nearly 2kHz, in contrast to 454Hz from the previous design. The design targets the Avnet UltraZed-EV MPSoC board, which boots custom PetaLinux images. API code and post-processing algorithms run in this environment to interface with the FPGA control registers and further process frames of data. Primary configuration options include variable sample rate, window coefficients, FIR filter coefficients, chirp length, pulse repetition interval, decimation factor, number of averaged frames, error monitoring, three DMA sampling points, and DMA ring buffer transfers. The result is a dynamic, high-speed, small-scale design which can process 16 parallel channels of data in real time for 3-dimensional detection of local UAV traffic at a range of 1000m.

The use of multirotor UAVs to map GPS-degraded environments is useful for many purposes ranging from routine structural inspections to post-disaster exploration to search for survivors and evaluate structural integrity. Multirotor UAVs are able to reach many areas that humans and other robots cannot safely access. Because of their relatively short operational flight time compared to other robotic applications, using multiple UAVs to collaboratively map these environments can streamline the mapping process significantly. This research focuses on four primary areas regarding autonomous mapping and navigation with multiple UAVs in complex unknown or partially unknown GPS-denied environments: The first area is the high-level coverage path planning necessary to successfully map these environments with multiple agents. The second area is the lower-level reactive path planning that enables autonomous navigation through complex, unknown environments. Third, is the estimation framework that enables autonomous flight without the use of GPS or other global position sensors. Lastly, it focuses on the mapping framework to build a single dense 3D map of these environments with multiple agents flying simultaneously.

With increased mobile hardware capabilities, improved UAVs and modern algorithms, accurate maps can be created in real-time by capturing overlapping photographs of the ground. A method for mapping that can be used is to position photos by relying purely on the GPS position and altitude. However, GPS inaccuracies will be visible in the created map. In this paper, we will instead present a method for aligning the photos correctly with the help of feature matching. Feature matching is a well-known method which analyses two photos to find similar parts. If an overlap exists, feature matching can be used to find and localise those parts, which can be used for positioning one image over the other at the overlap. When repeating the process, a whole map can be created. For this purpose, we have also evaluated a selection of feature detection and matching algorithms. The algorithm found to be the best was SIFT with FLANN, which was then used in a prototype for creating a complete map of a forest. Feature matching is in many cases superior to GPS positioning, although it cannot be fully depended on as failed or incorrect matching is a common occurrence.

The fact that technological revolutions have a large impact on the way we conduct war is a commonly accepted fact, but is it the technological innovations themselves? Or is the way we perceive them? One aspect not commonly considered is the impact of gendered preconceptions. This study aims to explore the theory of underestimation of military technology considered feminine presented by Lauren Wilcox, by studying discursive presentations of the UAVs and their pilots in three articles. The results of this study are twofold. Firstly, UAVs and their crews are associated with masculine attributes to a low degree, although there seem to be a difference between the crew and the UAV itself. Secondly, the discursive approach to UAVs share similar traits with the early opinions of the machine gun and airplane described by Wilcox. Conclusively, this study indicates that there may be a connection between gendered attributes and discursive approaches to new military technology, although more research on the discursive discrepancy between craft and crew is required.

<div>This thesis presents work and simulations containing the use of Artificial Intelligence for Unmanned Aerial Vehicles in search and rescue and/or surveillance operations. The goal is to create a vision system that leverages Artificial Intelligence, mainly Deep Learning techniques to build a pipeline that enables fast and accurate classification of the environment of the robot. Deep Neural Networks are trained and tested on ’emergency situational data. Further, the power of this vision system is leveraged to extend the problem onto a multiagent system to handle fault tolerance. The multi-agent system is also made resilient to Byzantine malicious attacks to help improve the reliability of the system.</div><div><br></div><div>This thesis also shows the use of Artificial Intelligence for effective surveillance for defense related purposes. Tracking the GPS coordinates of a boat using only the video of the boat captured by a camera and the GPS coordinates of the camera itself is demonstrated. The solution was tested by the Department of Defense - Department of the Navy, Naval Information Warfare Center Pacific.</div>

<p>Green turtles (<em>Chelonia mydas</em>) perform ocean-crossing migrations, maintain healthy marine ecosystems, generate income through tourism, and are endangered and declining globally. For these reasons, among others, this species has been a focus of numerous research programs worldwide for almost a century. Most of these sea turtle research programs require some form of animal handling to collect the required data (e.g., tagging information or the collection of biological samples). However, this can cause stress, especially for wild animals, and that raises ethical issues. Here, I describe novel methods for assessing and mitigating the effects of handling stress on green turtles. Specifically: (1) I used a combination of animal-borne cameras and drone footage to determine how handling stress altered the post-release behavior of green turtles and (2) I used a photo-ID software to determine whether flipper scales can provide more accurate identifications than the more conventionally used facial scale patterns. </p> <p><br></p> <p>I found that turtles spent more time swimming and had shortened dive intervals in the first 30 mins after capture and attachment of a camera than in the hours that follow. Instances of socializing, foraging and resting increased over the 3-3.5 h after release. Animals recorded by drone and not captured were less likely to rest, which suggests this behavior may be a recovery response to handling and/or stress. The same animals were also more likely to socialize. When determining the accuracy of flipper or facial images for photo-ID, I found that head scales provided correct identifications 80% of the time, whereas the flipper provided correct identifications 100% of the time. This implies that researchers could use the flipper instead of more invasive tagging techniques, such as metal flipper tags or using lights to photograph the face for photo-ID, which can induce stress.</p>

<p>Multi-object tracking has been well studied in the field of computer vision. Meanwhile, with the advancement of the Unmanned Aerial Vehicles (UAV) technology, the flexibility and accessibility of UAV draws research attention to deploy multi-object tracking on UAV. The conventional solutions usually adapt using the "tracking-by-detection" paradigm. Such a paradigm has the structure where tracking is achieved through detecting objects in consecutive frames and then associating them with re-identification. However, the dynamic background, crowded small objects, and limited computational resources make multi-object tracking on UAV more challenging. Providing energy-efficient multi-object tracking solutions on the drone-captured video is critically demanded by research community. </p> <p>    </p> <p>To stimulate innovation in both industry and academia, we organized the 2021 Low-Power Computer Vision Challenge with a UAV Video track focusing on multi-class multi-object tracking with customized UAV video. This thesis analyzes the qualified submissions of 17 different teams and provides a detailed analysis of the best solution. Methods and future directions for energy-efficient AI and computer vision research are discussed. The solutions and insights presented in this thesis are expected to facilitate future research and applications in the field of low-power vision on UAV.</p> <p>    </p> <p>With the knowledge gathered from the submissions, an optical flow oriented multi-object tracking framework, named OF-MOT, is proposed to address the similar problem with a more realistic drone-captured video dataset. OF-MOT uses the motion information of each detected object of the previous frame to detect the current frame, then applies a customized object tracker using the motion information to associate the detected instances. OF-MOT is evaluated on a drone-captured video dataset and achieves 24 FPS with 17\% accuracy on a modern GPU Titan X, showing that the optical flow can effectively improve the multi-object tracking.</p> <p>    </p> <p>Both competition results analysis and OF-MOT provide insights or experiment results regarding deploying multi-object tracking on UAV. We hope these findings will facilitate future research and applications in the field of UAV vision.</p>