Indiana University-Purdue University Indianapolis (IUPUI) / An Autonomous vehicle depends on the combination of latest technology or the ADAS safety features such as Adaptive cruise control (ACC), Autonomous Emergency Braking (AEB), Automatic Parking, Blind Spot Monitor, Forward Collision Warning or Avoidance (FCW or FCA), Lane Departure Warning. The current trend follows incorporation of these technologies using the Artificial neural network or Deep neural network, as an imitation of the traditionally used algorithms. Recent research in the field of deep learning and development of competent processors for autonomous or self-driving car have shown amplitude of prospect, but there are many complexities for hardware deployment because of limited resources such as memory, computational power, and energy. Deployment of several mentioned ADAS safety feature using multiple sensors and individual processors, increases the integration complexity and also results in the distribution of the system, which is very pivotal for autonomous vehicles. This thesis attempts to tackle two important adas safety feature: Forward collision Warning, and Object Detection using the machine learning and Deep Neural Networks and there deployment in the autonomous embedded platform. 1. A machine learning based approach for the forward collision warning system in an autonomous vehicle. 2. 3-D object detection using Lidar and Camera which is primarily based on Lidar Point Clouds. The proposed forward collision warning model is based on the forward facing automotive radar providing the sensed input values such as acceleration, velocity and separation distance to a classifier algorithm which on the basis of supervised learning model, alerts the driver of possible collision. Decision Tress, Linear Regression, Support Vector Machine, Stochastic Gradient Descent, and a Fully Connected Neural Network is used for the prediction purpose. The second proposed methods uses object detection architecture, which combines the 2D object detectors and a contemporary 3D deep learning techniques. For this approach, the 2D object detectors is used first, which proposes a 2D bounding box on the images or video frames. Additionally a 3D object detection technique is used where the point clouds are instance segmented and based on raw point clouds density a 3D bounding box is predicted across the previously segmented objects.

To teleoperate an unmanned vehicle a rich set of information should be gathered from surroundings.These systems use sensors which sends high amounts of data and processing the data in CPUs can be time consuming. Similarly, the algorithms that use the data may work slow because of the amount of the data. The solution is, preprocessing the data taken from the sensors on the vehicle and transmitting only the necessary parts or the results of the preprocessing. In this thesis a 180 degree laser scanner at the front end of an unmanned ground vehicle (UGV) tilted up and down on a horizontal axis and point clouds constructed from the surroundings. Instead of transmitting this data directly to the path planning or obstacle avoidance algorithms, a preprocessing stage has been run. In this preprocess rst, the points belonging to the ground plane have been detected and a simplied version of ground has been constructed then the obstacles have been detected. At last, a simplied ground plane as ground and simple primitive geometric shapes as obstacles have been sent to the path planning algorithms instead of sending the whole point cloud.

This master thesis is oriented on a problem of creating video and image annotations. This problem is solved by crowdsourcing approach. Crowdsourcing games were designed and implemented to make solution of this problem . It was proven by testing that these games are capable of creating high quality annotations. Launching these games on a larger scale could create large database of annotated videos and images.

The joint European Space Agency and Japan Aerospace Exploration Agency mission Comet Interceptor seeks to perform a flyby of a Small Solar System Body (SSSB), through use of a multi-element spacecraft. It comprises a primary spacecraft and two subspacecraft, the latter of which will encounter the intercepted object at a small enough distance that its end-of-life might occur at an impact of either the object itself or its potential coma. The Optical Periscopic Imager for Comets (OPIC) is an instrument implemented on one of these small probes which will generate monochromatic images during the encounter. Given a limited data budget before the possible impact, there is a need for data prioritization to ensure that only the most scientifically relevant data is collected. To enable this, algorithms for autonomously cropping an object nucleus from an image were developed during this thesis work. As the computational capabilities of OPIC are limited, the algorithms were required to be of low computational complexity. Additionally, given that the close environment of SSSB in general and comets in particular often exhibit considerable quantities of gas and dust which can generate cluttering in images, the algorithms developed were required to be resistant to noise. Three image cropping algorithms were developed with varying computational complexities. These were tested for cropping accuracy and relative execution times on data from both previous space missions as well as simulated photorealistic images. All three algorithms were able to properly find a bounding box of an object nucleus and any of its significant plumes. The accuracy in cropping correctness of the region borders generated increased with the computational complexity of the algorithms.

This work is focused on advanced methods of mobile robot's path planning. The theoretical part describes selected graphical methods, which are useful for speeding up the process of finding the shortest paths, for example through reduction of explored nodes of the state space. In the practical part was created simulate environment in the Python language and in this environment, selected algorithms was implemented.

The thesis deals with the changes of a cable robot to a manipulator. The mechanical changes are mostly about adding an active part to a moving platform with the ability to transfer objects and the effort to exchange the silicon cables for metal ones. The main part of the thesis is the proposed design and implementation of the algorithm for detection of a possible collision of the cable robot with an object in its working space.

Collision detection between deforming models is a difficult problem for collision detection systems to handle. This problem is even more difficult when deformations are unconstrained, objects are in close proximity to one another, and when the entity count is high. We propose a method to perform collision detection between multiple deforming objects with unconstrained deformations that will give good results in close proximities. Currently no systems exist that achieve good performance on both unconstrained triangle level deformations and deformations that preserve edge connectivity. We propose a new system built as a combination of Graphics Processing Unit (GPU) based culling and Axis Aligned Bounding Box (AABB) based culling. Techniques for performing hierarchy-less GPU-based culling are given. We then discuss how and when to switch between GPU-based culling and AABB based techniques.

Technology has revamped every aspect of our life, one of those various facets is fashion industry. Plenty of deep learning architectures are taking shape to augment fashion experiences for everyone. There are numerous possibilities of enhancing the fashion technology with deep learning. One of the key ideas is to generate fashion style and recommendation using artificial intelligence. Likewise, another significant feature is to gather reliable information of fashion trends, which includes analysis of existing fashion related images and data. When specifically dealing with images, localisation and segmentation are well known to address in-depth study relating to pixels, objects and labels present in the image. In this master thesis a complete framework is presented to perform localisation and segmentation on fashionista images. This work is a part of an interesting research work related to Fashion Style detection and Recommendation. Developed solution aims to leverage the possibility of localising fashion items in an image by drawing bounding boxes and labelling them. Along with that, it also provides pixel-wise semantic segmentation functionality which extracts fashion item label-pixel data. Collected data can serve as ground truth as well as training data for the aimed deep learning architecture. A study related to localisation and segmentation of videos has also been presented in this work. The developed system has been evaluated in terms of flexibility, output quality and reliability as compared to similar platforms. It has proven to be fully functional solution capable of providing essential localisation and segmentation services while keeping the core architecture simple and extensible. / Tekniken har förnyat alla aspekter av vårt liv, en av de olika fasetterna är modeindustrin. Massor av djupa inlärningsarkitekturer tar form för att öka modeupplevelser för alla. Det finns många möjligheter att förbättra modetekniken med djup inlärning. En av de viktigaste idéerna är att skapa modestil och rekommendation med hjälp av artificiell intelligens. På samma sätt är en annan viktig egenskap att samla pålitlig information om modetrender, vilket inkluderar analys av befintliga moderelaterade bilder och data. När det specifikt handlar om bilder är lokalisering och segmentering väl kända för att ta itu med en djupgående studie om pixlar, objekt och etiketter som finns i bilden. I denna masterprojekt presenteras en komplett ram för att utföra lokalisering och segmentering på fashionista bilder. Detta arbete är en del av ett intressant forskningsarbete relaterat till Fashion Style detektering och rekommendation. Utvecklad lösning syftar till att utnyttja möjligheten att lokalisera modeartiklar i en bild genom att rita avgränsande lådor och märka dem. Tillsammans med det tillhandahåller det även pixel-wise semantisk segmenteringsfunktionalitet som extraherar dataelementetikett-pixeldata. Samlad data kan fungera som grundsannelse samt träningsdata för den riktade djuplärarkitekturen. En studie relaterad till lokalisering och segmentering av videor har också presenterats i detta arbete. Det utvecklade systemet har utvärderats med avseende på flexibilitet, utskriftskvalitet och tillförlitlighet jämfört med liknande plattformar. Det har visat sig vara en fullt fungerande lösning som kan tillhandahålla viktiga lokaliseringsoch segmenteringstjänster samtidigt som kärnarkitekturen är enkel och utvidgbar.

[pt] As técnicas de aprendizado profundo têm mostrado contribuições significativas em vários campos, incluindo a análise de imagens. A grande maioria dos trabalhos em visão computacional concentra-se em propor e aplicar novos modelos e algoritmos de aprendizado de máquina. Para tarefas de aprendizado supervisionado, o desempenho dessas técnicas depende de uma grande quantidade de dados de treinamento, bem como de dados rotulados. No entanto, a rotulagem é um processo caro e demorado. Uma recente área de exploração são as reduções dos esforços na preparação de dados, deixando-os sem inconsistências, ruídos, para que os modelos atuais possam obter um maior desempenho. Esse novo campo de estudo é chamado de Data-Centric IA. Apresentamos uma nova abordagem baseada em Deep Reinforcement Learning (DRL), cujo trabalho é voltado para a preparação de um conjunto de dados em problemas de detecção de objetos, onde as anotações de caixas delimitadoras são feitas de modo autônomo e econômico. Nossa abordagem consiste na criação de uma metodologia para treinamento de um agente virtual a fim de rotular automaticamente os dados, a partir do auxílio humano como professor desse agente. Implementamos o algoritmo Deep Q-Network para criar o agente virtual e desenvolvemos uma abordagem de aconselhamento para facilitar a comunicação do humano professor com o agente virtual estudante. Para completar nossa implementação, utilizamos o método de aprendizado ativo para selecionar casos onde o agente possui uma maior incerteza, necessitando da intervenção humana no processo de anotação durante o treinamento. Nossa abordagem foi avaliada e comparada com outros métodos de aprendizado por reforço e interação humano-computador, em diversos conjuntos de dados, onde o agente virtual precisou criar novas anotações na forma de caixas delimitadoras. Os resultados mostram que o emprego da nossa metodologia impacta positivamente para obtenção de novas anotações a partir de um conjunto de dados com rótulos escassos, superando métodos existentes. Desse modo, apresentamos a contribuição no campo de Data-Centric IA, com o desenvolvimento de uma metodologia de ensino para criação de uma abordagem autônoma com aconselhamento humano para criar anotações econômicas a partir de anotações escassas. / [en] Deep learning techniques have shown significant contributions in various fields, including image analysis. The vast majority of work in computer vision focuses on proposing and applying new machine learning models and algorithms. For supervised learning tasks, the performance of these techniques depends on a large amount of training data and labeled data. However, labeling is an expensive and time-consuming process. A recent area of exploration is the reduction of efforts in data preparation, leaving it without inconsistencies and noise so that current models can obtain greater performance. This new field of study is called Data-Centric AI. We present a new approach based on Deep Reinforcement Learning (DRL), whose work is focused on preparing a dataset, in object detection problems where the bounding box annotations are done autonomously and economically. Our approach consists of creating a methodology for training a virtual agent in order to automatically label the data, using human assistance as a teacher of this agent. We implemented the Deep Q-Network algorithm to create the virtual agent and developed a counseling approach to facilitate the communication of the human teacher with the virtual agent student. We used the active learning method to select cases where the agent has more significant uncertainty, requiring human intervention in the annotation process during training to complete our implementation. Our approach was evaluated and compared with other reinforcement learning methods and human-computer interaction in different datasets, where the virtual agent had to create new annotations in the form of bounding boxes. The results show that the use of our methodology has a positive impact on obtaining new annotations from a dataset with scarce labels, surpassing existing methods. In this way, we present the contribution in the field of Data-Centric AI, with the development of a teaching methodology to create an autonomous approach with human advice to create economic annotations from scarce annotations.

Multi-GPU machines are being increasingly used in high performance computing. These machines are being used both as standalone work stations to run computations on medium to large data sizes (tens of gigabytes) and as a node in a CPU-Multi GPU cluster handling very large data sizes (hundreds of gigabytes to a few terabytes). Each GPU in such a machine has its own memory and does not share the address space either with the host CPU or other GPUs. Hence, applications utilizing multiple GPUs have to manually allocate and managed at a on each GPU. A significant body of scientific applications that utilize multi-GPU machines contain computations inside affine loop nests, i.e., loop nests that have affine bounds and affine array access functions. These include stencils, linear-algebra kernels, dynamic programming codes and data-mining applications. Data allocation, buffer management, and coherency handling are critical steps that need to be performed to run affine applications on multi-GPU machines. Existing works that propose to automate these steps have limitations and in efficiencies in terms of allocation sizes, exploiting reuse, transfer costs and scalability. An automatic multi-GPU memory manager that can overcome these limitations and enable applications to achieve salable performance is highly desired. One technique that has been used in certain memory management contexts in the literature is that of bounding boxes. The bounding box of an array, for a given tile, is the smallest hyper-rectangle that encapsulates all the array elements accessed by that tile. In this thesis, we exploit the potential of bounding boxes for memory management far beyond their current usage in the literature. In this thesis, we propose a scalable and fully automatic data allocation and buffer management scheme for affine loop nests on multi-GPU machines. We call it the Bounding Box based Memory Manager (BBMM). BBMM is a compiler-assisted runtime memory manager. At compile time, it use static analysis techniques to identify a set of bounding boxes accessed by a computation tile. At run time, it uses the bounding box set operations such as union, intersection, difference, finding subset and superset relation to compute a set of disjoint bounding boxes from the set of bounding boxes identified at compile time. It also exploits the architectural capability provided by GPUs to perform fast transfers of rectangular (strided) regions of memory and hence performs all data transfers in terms of bounding boxes. BBMM uses these techniques to automatically allocate, and manage data required by applications (suitably tiled and parallelized for GPUs). This allows It to (1) allocate only as much data (or close to) as is required by computations running on each GPU, (2) efficiently track buffer allocations and hence, maximize data reuse across tiles and minimize the data transfer overhead, (3) and as a result, enable applications to maximize the utilization of the combined memory on multi-GPU machines. BBMM can work with any choice of parallelizing transformations, computation placement, and scheduling schemes, whether static or dynamic. Experiments run on a system with four GPUs with various scientific programs showed that BBMM is able to reduce data allocations on each GPU by up to 75% compared to current allocation schemes, yield at least 88% of the performance of hand-optimized Open CL codes and allows excellent weak scaling.