No description available.

No description available.

No description available.

No description available.

Achieving autonomous in complex natural environments has the potential to transform society by bringing the benefits of automation from the confines of the factory floor to the outdoors. There, it could benefit areas such as environmental monitoring and clean-up, precision agriculture, delivery of goods. A fundamental requirement for achieving these goals are sensors that can provide reliable support for navigation, e.g., a drone, in natural environments. In this thesis, sonar-based navigation has been investigated as an approach to parsimonious autonomous sensing for drones. Bats living in dense vegetation have demonstrated that autonomous navigation in a complex, natural environment based on two one-dimensional ultrasonic echo streams is feasible. Here, a biomimetic sonar head has been used to collect echo data from recreations of natural foliage in the lab under controlled conditions. This data was used to address the research question whether the grazing angle at which the sonar is looking at a surface can be estimated from the echoes -- despite the random three-dimensional nature of the scatter from the foliage. To investigate this, the echoes have been subjected to statistical analysis such as spectral coherence and cross-correlation. Most importantly, the foliage data was compared against predictions made by the Endura method (energy, duration, and range method) that has been devices for two-dimension random scatterers. The results of this analysis shows that -- despite their profoundly random nature -- echoes can be used to estimate the sonar grazing angle directly, i.e., without the need to resort to reconstructions of the foliage geometry. This opens the possibility of developing simple devices for navigation control in natural environments that can control the direction of motion at a very little computational cost. / Master of Science / Autonomously flying drones is a potential technology that could bring benefits to the society and improve the quality of life for humans[22]. Therefore, a study of autonomously flying in a natural environment is necessary, and this thesis will focus on drone that could recognize objects with different grazing angle and acoustic signal by collecting data from near foliage surface. For example, when a bush wall is in front of the drone, a on board computer could inform drone whether the drone airline will collide with the bush wall or the bush wall is safely out of drone’s path[5]. If on board computer reads that there will be a collision with bush wall, then drone needs to make decision (change direction or stop immediately) to avoid crush on to bush wall. A sonar based navigation system has been investigated as an approach to achieve autonomous sensing for drones, which is inspired by bats. Bats use their natural sonar system to navigate in cave or forest, hence, it is hardly to see bats slam into any obstacles while flying. Bats navigation behaviours could be reconstructed as a sonar based autonomy. Hence, this thesis is inspired by bats to determine if there is a computational way to illustrate that sonar based sensor could be a solution to achieve reactive autonomy by using different grazing angle of the surface’s acoustic signals.

The focus of this thesis is to implement factor graphs into the problem of detection, classification, and localization (DCL) of underwater objects using active SOund Navigation And Ranging (SONAR). A factor graph is a bipartite graphical representation of the decomposition of a particular function. Messages are passed along the edges connecting factor and variable nodes, on which, a message passing algorithm is applied to compute the posterior probabilities at a particular node. This thesis addresses two issues. In the first section, the formulation of factor graphs for each section of the DCL chain required followed by their closed-form solutions. For the detector, the factor graph determines if the signal is a detection or simply noise. In the classifier, it outputs the probability for the elements in the class. Last, when using a factor graph for the tracker, it gives the estimated state of the object being tracked. The second part concentrates on the application to Continuous Active SONAR (CAS). When using CAS, a bistatic configuration is used allowing for a more rapid update rate where two unmanned underwater vehicles (UUVs) are used as the receiver and transmitter. The goal is to evaluate CAS's effectiveness to determine if the tracking accuracy improves as the transmit interval decreases. If CAS proves to be more efficient in target tracking, the next objective is to determine which messages sent between the two UUVs are most beneficial. To test this, a particle filter simulation is used. / Master of Science

Miniature Autonomous Underwater Vehicles are a major area of research and development today. Because of their size and agility, they are capable of exploring and operating in smaller bodies of water in addition to areas of the ocean that would be out of reach for a larger vehicle. Being autonomous requires that the system must be capable of performing without the need for human supervision, so use of external sensors such as sonar are needed to ensure the safety of the vehicle during missions. However, since all of the onboard instrumentation and external equipment must also be miniature in size, the implementation of a small sonar system is desirable. This thesis contains a brief introduction to sound and sonar, leading into a description of the design and development of a small, inexpensive sonar altimeter. Piezoelectric material is used for transduction in the sonar system while a PIC microcontroller processes the return signals from the water. This altimeter was made to be implemented on a miniature autonomous underwater vehicle developed by the Autonomous Systems and Controls Laboratory at Virginia Polytechnic Institute. In addition to being capable of reporting ocean depths, sonar systems can be used to aid in the navigation of underwater vehicles. A constant altitude controller based on sonar data has been designed, tested, and implemented on the autonomous underwater vehicle. Possibilities for an obstacle avoidance system involving sonar are also discussed in this thesis. / Master of Science

Object detection and seafloor segmentation for conventional 2D side-scan sonar imagery is a well-investigated problem. However, due to recent advances in sensing technology, the side-scan sonar now produces a true 3D point cloud representation of the seafloor embedded with echo intensity. This creates a need to develop algorithms to process the incoming 3D data for applications such as object detection and segmentation, and an opportunity to leverage advances in 3D point cloud processing developed for terrestrial applications using optical sensors (e.g. LiDAR). A bottleneck in deploying 3D side-scan sonar sensors for online applications is attributed to the complexity in handling large amounts of data which requires higher memory for storing and processing data on embedded computers. The present research aims to improve data processing capabilities on-board autonomous underwater vehicles (AUVs). A supervoxel-based framework for over-segmentation and object detection is proposed which reduces a dense point cloud into clusters of similar points in a neighborhood. Supervoxels extracted from the point cloud are then described using feature vectors which are computed using geometry, echo intensity and depth attributes of the constituent points. Unsupervised density based clustering is applied on the feature space to detect objects which appear as outliers. / Master of Science / Acoustic imaging using side-scan sonar sensors has proven to be useful for tasks like seafloor mapping, mine countermeasures and habitat mapping. Due to advancements in sensing technology, a novel type of side-scan sonar sensor is developed which provides true 3D representation of the seafloor along with the echo intensity image. To improve the usability of the novel sensors on-board the carrying vehicles, efficient algorithms needs to be developed. In underwater robotics, limited computational and data storage capabilities are available which poses additional challenges in online perception applications like object detection and segmentation. In this project, I investigate a clustering based approach followed by an unsupervised machine learning method to perform detection of objects on the seafloor using the novel side scan sonar. I also show the usability of the approach for performing segmentation of the seafloor.

A maioria dos solos tropicais e subtropicais apresenta disponibilidade insuficiente de nitrogênio (N) para atender a demanda da cultura do algodão visando à obtenção de elevados rendimentos. O N nas plantas pode ser avaliado através da utilização de sensores de espectrometria óptica ativa, nos quais se destaca o equipamento comercial utilizado neste trabalho (N-SensorTM ALS), capaz de determinar a dose de N a ser aplicada em cobertura em tempo real. O objetivo geral deste trabalho foi avaliar o desempenho de um sensor óptico ativo, sonares e clorofilômetro para predizer parâmetros de planta e testar estratégias de adubação nitrogenada baseada nas leituras do sensor óptico para aumento da produtividade do algodoeiro em sistemas de cultivo adensado e convencional. Os experimentos foram realizados durante três anos sendo: a) em Chapadão do Céu, GO nas safras 2012/13 e 2013/14; b) e em Campo Verde, MT na safra 2014/15. Os parâmetros de planta avaliados foram: altura, massa seca, massa fresca e nitrogênio acumulado na biomassa. Estes parâmetros foram analisados nas áreas correspondentes à adubação em taxa fixa e correlacionados com o índice de vegetação gerado pelos sensores. Os resultados demonstraram que o sensor ativo de dossel utilizado é altamente eficiente em estimar altura de planta, massa fresca, massa seca e nitrogênio acumulado na cultura do algodoeiro. A utilização de clorofilômetro para monitorar o teor de N foliar no algodoeiro não demonstrou ser uma técnica eficiente. O uso de sensor ultrassônico (sonar) para estimar parâmetros de altura de planta, massa seca, massa fresca e nitrogênio acumulado na planta do algodoeiro é uma estratégia promissora que, embora não tenha a mesma eficiência de sensores de dossel, demonstrou bons resultados na detecção da variabilidade nas lavouras. As diferentes estratégias de adubação nitrogenada no algodoeiro orientadas pelo sensor ativo de dossel não resultaram em efeitos claros na produtividade. A produtividade de cada tratamento mostrou-se mais dependente da localização das repetições dentro da área do que da estratégia de adubação adotada. / Most tropical and subtropical soils has insufficient availability of nitrogen (N) to meet the demand of the cotton crop in order to obtain high yields. The N plants may be evaluated using active optical spectroscopy sensors as the one used in this investigation (N-SensorTM ALS). The equipment is capable of determining in real time the amount of N to be applied spread in the field. The aim of this study was to evaluate the performance of an active optical sensor, sonar and chlorophyll sensor to predict plant parameters and test nitrogen fertilization strategies based on the optical sensor readings in order to increase cotton yield in narrow and conventional farming systems. The experiments were conducted in three years: a) in Chapadão do Céu - GO 2012/13 and 2013/14; b) in Campo Verde, MT - 2014/15. The evaluated parameters were plant height, dry weight, fresh and accumulated nitrogen in the biomass. These parameters were analyzed in the areas corresponding to fertilization in fixed rate and correlated with the vegetation index generated by the sensors. The results showed that the active sensor used is highly efficient in estimating plant height, fresh weight, dry weight and accumulated nitrogen on cotton crop. The use of chlorophyll sensor to monitor the leaf nitrogen content in cotton was not effective. The use of ultrasonic sensor (sonar) to estimate plant height parameters, dry pasta, fresh pasta and accumulated nitrogen on cotton plant is a promising strategy. Although it has not the same efficiency as canopy sensors, it showed good results in the detection of variability in crops. The different nitrogen fertilization strategies in cotton-driven asset canopy sensor resulted in no clear effects on productivity. The productivity of each treatment was more dependent on the location of the repetitions within the area than the adopted fertilization strategy.

A tarefa de operar um veículo submarino durante missões de inspeção de ambientes estruturados como, por exemplo, duto de usinas hidrelétricas, é feita principalmente por meio de referências visuais e uma bússola magnética. Porém alguns ambientes desse tipo podem apresentar uma combinação de baixa visibilidade e anomalias ferromagnéticas que inviabilizaria esse tipo de operação. Este trabalho, motivado pelo desenvolvimento de um veículo submarino operado remotamente (ROV) para ser usado em ambientes com essas restrições, propõe um sistema de navegação que utiliza o conhecimento prévio das dimensões do ambiente para corrigir o estado do veículo por meio da correlação dessas dimensões com os dados de um sonar de imageamento 2D. Para fazer essa correlação é utilizado o ltro de partículas, que é uma implementação não paramétrica do ltro Bayesiano. Esse ltro faz a estimação do estado com base nos métodos sequenciais de Monte Carlo e permite trabalhar de uma maneira simples com modelos não lineares. A desvantagem desse tipo de fusão sensorial é o seu alto custo computacional o que geralmente o impede de ser utilizado em aplicações de tempo real. Para que seja possível utilizar esse ltro em tempo real, será proposto neste trabalho uma implementação paralela utilizando uma unidade de processamento gráco (GPU) da NVIDIA e a arquitetura CUDA. Neste trabalho também será feito um estudo da utilização de duas congurações de sensores no sistema de navegação proposto neste trabalho. / The task of navigating a Remotely Operated underwater Vehicles (ROV) during inspection of man-made structures is performed mostly by visual references and occasionally a magnetic compass. Yet, some environments present a combination of low visibility and ferromagnetic anomalies that negates this approach. This paper, motivated by the development of a ROV designed to work on such environment, proposes a navigation method for this kind of vehicle. As the modeling of the system is nonlinear, the method proposed uses a particle lter to represent the vehicle state that is a nonparametric implementation of the Bayes lter. This method to work needs a priori knowledge of the environment map and to make the data association with this map, a 2D image sonar is used. The drawback of the sensor fusion used in this work is its high computational cost which generally prevents it from being used in real time applications. To be possible for this lter to be used in real time application, in this work is proposed a parallel implementation using a graphics processing unit (GPU) from NVIDIA and CUDA architecture. In this work is also made a study of two types of sensors conguration on the navigation system proposed in this work.