Approved for public release; distribution is unlimited / Reissued 30 May 2017 with correction to student's affiliation on title page. / Autonomous vehicle teams have great potential in a wide range of maritime sensing applications, including mine countermeasures (MCM). A key enabler for successfully employing autonomous vehicles in MCM missions is motion planning, a collection of algo-rithms for designing trajectories that vehicles must follow. For maximum utility, these algorithms must consider the capabilities and limitations of each team member. At a minimum, they should incorporate dynamic and operational constraints to ensure trajectories are feasible. Another goal is maximizing sensor performance in the presence of uncertainty. Optimal control provides a useful frame-work for solving these types of motion planning problems with dynamic constraints and di_x000B_erent performance objectives, but they usually require numerical solutions. Recent advances in numerical methods have produced a general mathematical and computational framework for numerically solving optimal control problems with parameter uncertainty—generalized optimal control (GenOC)— thus making it possible to numerically solve optimal search problems with multiple searcher, sensor, and target models. In this dissertation, we use the GenOC framework to solve motion planning problems for di_x000B_erentMCMsearch missions conducted by autonomous surface and underwater vehicles. Physics-based sonar detection models are developed for operationally relevant MCM sensors, and the resulting optimal search trajectories improve mine detection performance over conventional lawnmower survey patterns—especially under time or resource constraints. Simulation results highlight the flexibility of this approach for optimal mo-tion planning and pre-mission analysis. Finally, a novel application of this framework is presented to address inverse problems relating search performance to sensor design, team composition, and mission planning for MCM CONOPS development.

The interest for marine research and exploration has increased rapidly during the past decades and autonomous underwater vehicles (AUV) have been found useful in an increased amount of applications. The demand for versatile platform AUVs, able to perform a wide range of tasks, has become apparent. A vital part of an AUV is its motion control system, and an emerging problem for multipurpose AUVs is that the control performance is affected when the vehicle is configured with different payloads for each mission. Instead of having to manually re-tune the control system between missions, a method for automatic tuning of the control system has been developed in this master’s thesis. A model-based approach was implemented, where the current vehicle dynamics are identified by performing a sequence of excitation maneuvers, generating informative data. The data is used to estimate model parameters in predetermined model structures, and model-based control design is then used to determine an appropriate tuning of the control system. The performance and potential of the suggested approach were evaluated in simulation examples which show that improved control can be obtained by using the developed auto-tuning method. The results are considered to be sufficiently promising to justify implementation and further testing on a real AUV. The automatic tuning process is performed prior to a mission and is meant to compensate for dynamic changes introduced between separate missions. However, the AUV dynamics might also change during a mission which requires an adaptive control system. By using the developed automatic tuning process as foundation, the first steps towards an indirect adaptive control approach have been suggested. Also, the AUV which was studied in the thesis composed another interesting control problem by being overactuated in yaw control, this because yawing could be achieved by using rudders but also by differential drive of the propellers. As an additional and separate part of the thesis, an approach for using both techniques simultaneously have been proposed.

Neste trabalho, é abordado o problema da navegação de um veículo submarino autônomo. São propostos estimadores de estado que realizam fusão sensorial baseada em Filtro Estendido de Kalman. Esses estimadores de estado empregam as medidas dos seguintes sensores: uma unidade de medição inercial, um sensor de velocidade por efeito Doppler, um profundímetro e uma bússola. Primeiramente foi projetado um estimador de estados para o AUV Pirajuba, onde a estimação da orientação do veículo é realizada de forma desacoplada à estimação da velocidade e posição do veículo. Em seguida, foram desenvolvidos dois estimadores de estado que estimam orientação, velocidade e profundidade do veículo de forma acoplada. Para o projeto e testes dos estimadores mencionados anteriormente, foi empregada uma base de dados contendo um registro de medições reais dos sensores do veículo submarino autônomo Pirajuba, durante testes de campo no lago de uma represa. Os resultados dos testes validaram os estimadores de estado propostos nesse trabalho. Por último, foi realizada uma análise comparativa dos estimadores de estado mencionados. / This work concerns the navigation problem of an autonomous underwater vehicle. Some state estimators using sensorial fusion were proposed, the sensorial fusion is based in an Extended Kalman Filter. The state estimators are fed by measurements of the following sensors: an inertial measurements unit, a velocity sensor by Doppler effect, a depthmeter and a compass. In the first version of the EKF algorithm, the vehicles attitude estimation was decoupled from the vehicle velocity estimation. The second version considers the coupling between linear velocity and the attitude in the vehicle reference frame, taking the velocity reading for correction of the filter estimates. Finally, in the third version, the coupling between position and attitude is also considered, but the correction of the filters estimates is based on the depth readings. Experiments for supporting the design and validation of the navigation algorithms were based on a database constructed with motion measurements during the AUV maneuvers in the north coast of Sao Paulo, and the Guarapiranga lake in the São Paulo city. This work presents a comparative analysis of those algorithms.

This thesis addresses the path planning problem for Autonomous Underwater Vehicles (AUVs) using homotopy classes to provide topological information on how paths avoid obstacles. Looking for a path within a homotopy class constrains the search into a specific area of the search space, speeding up the computation of the path. Given a workspace with obstacles, the method starts by generating the homotopy classes. Those which probably contain lower cost solutions are determined by means of a lower bound criterion before computing a path. Finally, a path planner uses the topological information of homotopy classes to generate a few good solutions very quickly. Three path planners from different approaches have been proposed to generate paths for the homotopy classes obtained. The path planning is performed on Occupancy Grid Maps (OGMs) improved with probabilistic scan matching techniques. The results obtained with synthetic s scenarios and with real datasets show the feasibility of our method. / Aquesta tesi aborda el problema de la planificació de camins per a Vehicles Submarins Autònoms (AUVs) mitjançant la utilització de classes d'homotopia per a proporcionar informació topològica de com els camins eviten els obstacles. Calcular un camí dins d'una classe d'homotopia permet limitar l'espai de cerca accelerant-ne el càlcul de la solució. Donat un workspace amb obstacles, el mètode primer genera les classes homotòpiques. Aquelles classes que probablement contenen les solucions de menor cost s'identifiquen per mitjà d'una heurística sense haver-ne de calcular el camí al workspace. Finalment, un planificador de camins utilitza la informació topològica de les classes d'homotopia per generar solucions segons les classes seleccionades molt ràpidament. El mètode de planificació de camins s’aplica sobre Mapes d’Occupació de Graella (OGMs) millorats amb tècniques de scan matching probabilístic. Els tests i resultats obtinguts tan en escenaris sintètics com en datasets reals mostren la viabilitat del nostre mètode.

In the past decade, the design and manufacturing of intelligent multipurpose underwater vehicles has increased significantly. In the wide range of studies conducted in this field, the flexibility and autonomy of these devices with respect to their intended performance had been widely investigated. This work is related to the design and manufacturing of a small and lightweight autonomous underwater vehicle (AUV) with vision capabilities allowing detecting and contouring obstacles. It is indeed an exciting challenge to build a small and light submarine AUV, while making tradeoffs between performance and minimum available space as well as energy consumption. In fact, due to the ever-increasing in equipment complexity and performance, designers of AUVs are facing the issues of limited size and energy consumption. By using a pair of thrusters capable to rotate 360o on their axis and implementing a mass shifter with a control loop inside the vehicle, this later can efficiently adapt its depth and direction with minimal energy consumption. A prototype was fabricated and successfully tested in real operating conditions (in both pool and ocean). It includes the design and embedding of accurate custom multi-purpose sensors for multi-task operation as well as an enhanced coordinated system between a high-speed processor and accustomed electrical/mechanical parts of the vehicle, to allow automatic controlling its movements. Furthermore, an efficient tracking system was implemented to automatically detect and bypass obstacles. Then, fuzzy-based controllers were coupled to the main AUV processor system to provide the best commands to safely get around obstacles with minimum energy consumption. The fabricated prototype was able to work for a period of three hours with object tracking options and five hours in a safe environment, at a speed of 0.6 m/s at a depth of 8 m.

Neste trabalho, é abordado o problema da navegação de um veículo submarino autônomo. São propostos estimadores de estado que realizam fusão sensorial baseada em Filtro Estendido de Kalman. Esses estimadores de estado empregam as medidas dos seguintes sensores: uma unidade de medição inercial, um sensor de velocidade por efeito Doppler, um profundímetro e uma bússola. Primeiramente foi projetado um estimador de estados para o AUV Pirajuba, onde a estimação da orientação do veículo é realizada de forma desacoplada à estimação da velocidade e posição do veículo. Em seguida, foram desenvolvidos dois estimadores de estado que estimam orientação, velocidade e profundidade do veículo de forma acoplada. Para o projeto e testes dos estimadores mencionados anteriormente, foi empregada uma base de dados contendo um registro de medições reais dos sensores do veículo submarino autônomo Pirajuba, durante testes de campo no lago de uma represa. Os resultados dos testes validaram os estimadores de estado propostos nesse trabalho. Por último, foi realizada uma análise comparativa dos estimadores de estado mencionados. / This work concerns the navigation problem of an autonomous underwater vehicle. Some state estimators using sensorial fusion were proposed, the sensorial fusion is based in an Extended Kalman Filter. The state estimators are fed by measurements of the following sensors: an inertial measurements unit, a velocity sensor by Doppler effect, a depthmeter and a compass. In the first version of the EKF algorithm, the vehicles attitude estimation was decoupled from the vehicle velocity estimation. The second version considers the coupling between linear velocity and the attitude in the vehicle reference frame, taking the velocity reading for correction of the filter estimates. Finally, in the third version, the coupling between position and attitude is also considered, but the correction of the filters estimates is based on the depth readings. Experiments for supporting the design and validation of the navigation algorithms were based on a database constructed with motion measurements during the AUV maneuvers in the north coast of Sao Paulo, and the Guarapiranga lake in the São Paulo city. This work presents a comparative analysis of those algorithms.

Autonomous underwater vehicles have become a popular countermeasure to naval mines. Saab’s AUV62-MR detects, locates and identifies mine-like objects through three phases. By extracting functionality from the AUV62-MR and placing it on a second vehicle, it is suggested that the second and third phases can be performed in parallel. This thesis investigates how to design the second vehicle so that the runtime of the mine reconnaissance process is minimized. A simulation framework is implemented to simulate the second and third phases of the mine reconnaissance process in order to test various design choices. The vehicle design choices in focus are the size and the route planning of the second vehicle. The route-planning algorithms investigated in this thesis are a nearest neighbour algorithm, a simulated annealing algorithm, an alternating algorithm, a genetic algorithm and a proposed Dubins simulated annealing algorithm. The algorithms are evaluated both in a static environment and in the simulation framework. Two different vehicle sizes are investigated, a small and a large, by evaluating their performances in the simulation framework. This thesis takes into account the limited travelling distance of the vehicle and implements a k-means clustering algorithm to help the route planner determine which mine-like objects can be scanned without exceeding the distance limit. The simulation framework is also used to evaluate whether parallel execution of the second and third phases outperforms the current sequential execution. The performance evaluation shows that a major reduction in runtime can be gained by performing the two phases in parallel. The Dubins simulated annealing algorithm on average produces the shortest paths and is considered the preferred route-planning algorithm according to the performance evaluation. It also indicates that a small vehicle size results in a reduced runtime compared to a larger vehicle.

Visual Odometry (VO) is the process of estimating the relative motion of a vehicle by using solely image data gathered from the camera. In underwater environments, VO becomes extremely challenging but valuable since ordinary sensors for on-road localization are usually unpractical in these hostile environments. For years, VO methods have been purely based on Computer Vision (CV) principles. However, the recent advances in Deep Learning (DL) have ushered in a new era for VO approaches. These novel methods have achieved impressive performance with state-of-the-art results on urban datasets. Nevertheless, little effort has been made to push learning-based research towards natural environments, such as underwater. Consequently, this work aims to bridge the research gap by evaluating the effectiveness of the learning-based approach in the navigation of Autonomous Underwater Vehicles (AUVs). We compare two learning-based methods with a traditional feature-based method on the Underwater Caves dataset, a very challenging dataset collected in the unstructured environment of an underwater cave complex. Extensive experiments are thus conducted training the models on this dataset. Moreover, we investigate different aspects and propose several improvements, such as sub-sampling the video clips to emphasize the camera motion between consecutive frames, or training exclusively on images with relevant content discarding those with dark borders and representing solely sandy bottoms. Finally, during the training, we also leverage underwater images from other datasets, hence acquired from different cameras. However, the best improvement is obtained by penalizing rotations around the x-axis of the camera coordinate system. The three methods are evaluated on test sequences that cover different lighting conditions. In the most favorable environments, although learning-based methods are not up to par with the feature-based method, the results show great potential. Furthermore, in extreme lighting conditions, where the feature-based baseline sharply fails to bootstrap, one of the two learning-based methods produces instead qualitatively good trajectory results, revealing the power of the learning-based approach in this peculiar context. / Visuell Odometri (VO) används för att uppskatta den relativa rörelsen för ett fordon med hjälp av enbart bilddata från en eller flera kameror. I undervattensmiljöer blir VO extremt utmanande men värdefullt eftersom vanliga sensorer för lokalisering vanligtvis är opraktiska i dessa svåra miljöer. I åratal har VO-metoder enbart baserats på klassisk datorseende. De senaste framstegen inom djupinlärning har dock inlett en ny era för VO-metoder. Dessa nya metoder har uppnått imponerande prestanda på dataset urbana miljöer. Trots detta har ganska lite gjorts för att driva den inlärningsbaserad forskningen mot naturliga miljöer, till exempel under vattnet. Följaktligen syftar detta arbete till att överbrygga forskningsgapet genom att utvärdera effektiviteten hos det inlärningsbaserade tillvägagångssättet vid navigering av autonoma undervattensfordon (AUV). Vi jämför två inlärningsbaserade metoder med en traditionell nyckelpunktsbaserad metod som referens. Vi gör jämförelsen på Underwater Caves-datasetet, ett mycket utmanande dataset som samlats in i den ostrukturerade miljön i ett undervattensgrottkomplex. Omfattande experiment utförs för att träna modellerna på detta dataset. Vi undersöker också olika aspekter och föreslår flera förbättringar, till exempel, att delsampla videoklippen för att betona kamerarörelsen mellan på varandra följande bildrutor, eller att träna på en delmängd av datasetet bestående uteslutande på bilder med relevant innehåll för att förbättra skattningen av rörelsen. Under träningen utnyttjar vi också undervattensbilder från andra datamängder, och därmed från olika kameror. Den bästa förbättringen uppnås dock genom att straffa skattningar av stora rotationer runt kamerakoordinatsystemets x-axel. De tre metoderna utvärderas på testsekvenser som täcker olika ljusförhållanden. I de mest gynnsamma miljöerna visar resultaten stor potential, även om de inlärningsbaserade metoder inte är i nivå med den traditionella referensmetoden. Vid extrema ljusförhållanden, där referensmetoden misslyckas att ens initialisera, ger en av de två inlärningsbaserade metoderna istället kvalitativt bra resultat, vilket demonstrerar kraften i det inlärningsbaserade tillvägagångssättet i detta specifika sammanhang.

<p>Small, low cost Autonomous underwater vehicles (AUVs) provide ideal platforms for shallow water survey, as they are capable of unmanned navigation and can be programmed to acquire data at constant depth, or constant altitude above the seabed. AUVs can be deployed under most sea states and are unaffected by vessel motions that often degrade sonar and magnetometer data quality. The integration of sonar and magnetometer sensors on AUV’s is challenging, however, due to limited payload and strong magnetic fields produced by the vehicle motor.</p> <p>In this study, a Marine Magnetics Explorer Overhauser magnetometer was mated to a portable AUV (OceanServer Iver2) creating the first practical AUV- deployed magnetic survey system. To eliminate magnetic interference from the AUV, the magnetometer was tethered to the AUV with a 5 m tow cable, as determined by static and dynamic instrument testing. The results of the magnetic tests are presented, along with field data from a shallow water test area in Lake Ontario near Toronto, Canada. AUV-acquired magnetic survey data were compared directly with a conventional boat-towed magnetic survey of the same area. The AUV magnetic data were of superior quality despite being collected in rough weather conditions that would have made conventional survey impossible. The resulting high-resolution total magnetic intensity and analytic signal maps clearly identify several buried and surface ferrometallic targets that were verified in 500-kHz side- scan sonar imaging and visual inspection by divers.</p> / Master of Science (MSc)

Background Efficient underwater operations with autonomous underwater vehicles (AUVs) relying on several factors for a mission to be successful, such as operation time, distance covered, and waiting times. Today’s methods and processes for AUVs often struggle with inefficiencies and lack of route optimization. These challenges can result in increased operational costs and suboptimal performance. Minimizing operation time and utilizing route planning algorithms enables adaptation to operational challenges, potentially resulting in cost savings. Objectives This thesis aims to identify an efficient and practical solutions that will improve the operations for AUVs and the objective is to optimize the diving process through suboptimal routing algorithms in a predefined scenario. The study addresses one primary question to achive the aim. The question were: How can routing algorithms be implemented to improve the efficiency and reduce the operation time of Autonomous Underwater Vehicles? Methods The method describes three heuristic algorithms for optimizing the operations of AUVs. The first algorithm, the nearest neighbor heuristic (NNH) aims to minimize the distance an AUV needs to travel to recover and deploy ocean bottom nodes (OBN) within a cluster. The second algorithm, inspired by railway traffic, tries to prevent overlaps and minimizing the waiting times at the depot station. The third algorithm is highlighted as a local optimization algorithm that prioritizes the shortest waiting time over the nearest distance, adapting dynamically to available depot stations. Results The results in this thesis are derived from numerous simulations from different scenarios. The relationship between operations time and waiting time for different scenarios was obtained. The first algorithm proved to work for this type of situation. The second algorithm demonstrated its ability to yield superior solutions, albeit at the cost of being time-consuming due to a high number of iterations. The third algorithm was examined under conditions with and without delays. Even with delays, the algorithm consistently manages disturbances effectively. Conclusions While achieving an exact optimal solution remains challenging due to complexity, the research showed promising improvements in the endurance of the AUVs through the algorithms. The first algorithm was effective in minimizing the distance the AUVs traveled by selecting the most efficient path from numerous potential solutions. The second algorithm was slow due to a large number iterations, but the algorithm was able to find a solution where the operation and waiting time of the AUV could be reduced. The third algorithm was faster, but generally resulted in longer operation times. Also, increasing the number of AUVs resulted in shorter operation times but led to longer waiting times at the depot station, particularly in scenarios that became saturated with too many AUVs. / Bakgrund Undervattensoperationer med autonoma undervattensfarkoster (AUV:er) är beroende av flera faktorer för framgångsrika uppdrag, såsom driftstid, avstånd och väntetider. Dagens metoder och processer för AUV:er har ofta problem med ineffektivitet och bristande optimering av rutter. Dessa utmaningar kan leda till ökade driftkostnader och suboptimal prestanda. Genom att minimera operationstiden och använda ruttplaneringsalgoritmer möjliggörs anpassning till operativa utmaningar, vilket potentiellt kan resultera i kostnadsbesparingar. Syfte Detta examensarbete syftar till att utveckla effektiva och praktiska lösningar för att förbättra systemets prestanda och målet är att optimera rutterna genom suboptimala algoritmer i ett fördefinierat scenario. Arbetet behandlar en primärfråga för att uppnå målet. Frågan var: Hur kan ruttalgoritmer implementeras för att förbättra effektiviteten och minska drifttiden för autonoma undervattensfarkoster? Metod Metoden beskriver tre heuristiska algoritmer för att optimera driften förAUV:er. Den första algoritmen, närmaste granne heuristiken (NNH), syftar till attminimera avståndet en AUV behöver resa för att hämta och placera ut havsbottennoder (OBN) inom en kluster. Den andra algoritmen, inspirerad av tågtrafiken, syftar till att endast en AUV befinner sig vid depåstationen åt gången för att förhindra konflikter och minimera väntetider. Den tredje algoritmen är en lokal optimeringsalgoritm som prioriterar kortaste väntetiden över närmaste avstånd och anpassar sig dynamiskt till tillgängliga depåstationer. Resultat Resultaten i denna uppsats baseras på ett flertal simuleringar med olika scenarier. Förhållandet mellan drifttid och väntetid för olika scenarier erhölls. Den första algoritmen visade sig fungera bra för denna typ av situation. Den andra algoritmen visade sin förmåga att ge bättre lösningar, trots att den var tidskrävande pågrund av ett högt antal iterationer. Den tredje algoritmen undersöktes under förhållanden med och utan förseningar. Trots förseningar lyckades algoritmen konsekvent hantera störningar effektivt. Slutsats Trots komplexiteten med att tillhandahålla en exakt optimal lösning, visade arbetet förbättringar i AUV:ers uthållighet genom olika algoritmer. Den första algoritmen var effektiv för att minimera det avstånd som AUV:er färdades genom att välja en optimal väg bland många potentiella lösningar. Den andra algoritmen var långsam på grund av många iterationer, men algoritmen kunde hitta en lösning där AUV:ens drift- och väntetid kunde minskas. Den tredje algoritmen var snabbare men resulterade i längre drifttider. Vidare resulterade ökningen av antalet AUV:er i en minskning av drifttider men ökade väntetider, särskilt i scenarier som blev mättade med för många AUV:er.