A navegação autônoma é um dos problemas fundamentais na área da robótica móvel. Algoritmos capazes de conduzir um robô até o seu destino de maneira segura e eficiente são um pré-requisito para que robôs móveis possam executar as mais diversas tarefas que são atribuídas a eles com sucesso. Dependendo da complexidade do ambiente e da tarefa que deve ser executada, a programação de algoritmos de navegação não é um problema de solução trivial. Esta tese trata do desenvolvimento de sistemas de navegação autônoma baseados em técnicas de aprendizado supervisionado. Mais especificamente, foram abordados dois problemas distintos: a navegação de robôs/- veículos em ambientes urbanos e a navegação de robôs em ambientes não estruturados. No primeiro caso, o robô/veículo deve evitar obstáculos e se manter na via navegável, a partir de exemplos fornecidos por um motorista humano. No segundo caso, o robô deve identificar e evitar áreas irregulares (maior vibração), reduzindo o consumo de energia. Nesse caso, o aprendizado foi realizado a partir de informações obtidas por sensores. Em ambos os casos, algoritmos de aprendizado supervisionado foram capazes de permitir que os robôs navegassem de maneira segura e eficiente durante os testes experimentais realizados / Autonomous navigation is a fundamental problem in the field of mobile robotics. Algorithms capable of driving a robot to its destination safely and efficiently are a prerequisite for mobile robots to successfully perform different tasks that may be assigned to them. Depending on the complexity of the environment and the task to be executed, programming of navigation algorithms is not a trivial problem. This thesis approaches the development of autonomous navigation systems based on supervised learning techniques. More specifically, two distinct problems have been addressed: a robot/vehicle navigation in urban environments and robot navigation in unstructured environments. In the first case, the robot/vehicle must avoid obstacles and keep itself in the road based on examples provided by a human driver. In the second case, the robot should identify and avoid unstructured areas (higher vibration), reducing energy consumption. In this case, learning was based on information obtained by sensors. In either case, supervised learning algorithms have been capable of allowing the robots to navigate in a safe and efficient manner during the experimental tests

O contínuo aumento da complexidade no controle de sistemas robóticos, bem como a aplicação de grupos de robôs auxiliando ou substituindo seres humanos em atividades críticas tem gerado uma importante demanda por soluções mais robustas, flexíveis, e eficientes. O desenvolvimento convencional de algoritmos especializados, constituídos de sistemas baseados em regras e de autômatos usados para coordenar estes conjuntos físicos em um ambiente dinâmico é um desafio extremamente complexo. Diversos modelos de desenvolvimento existem, entretanto, muitos desafios da área da robótica móvel autônoma continuam em aberto. Esta tese se insere no contexto da busca por soluções inteligentes a serem aplicadas em robôs móveis autônomos com o objetivo de permitir a operação destes em ambientes dinâmicos. Buscamos, com a investigação e aplicação de estratégias inteligentes por meio de aprendizado de máquina no funcionamento dos robôs, a proposta de soluções originais que permitam uma nova visão sobre a operação de robôs móveis em três dos desafios da área da robótica móvel autônoma, que são: localização, navegação e operações com grupos de robôs. As pesquisas sobre localização e coordenação de grupos apresentam investigação e propostas originais, buscando estender o estado da arte, onde apresentam resultados inovadores. A parte sobre navegação tem como objetivo principal ser um elo entre os conceitos de localização e coordenação de grupos, sendo o foco o desenvolvimento de um veículo autônomo com maior implicação em avanços técnicos. Relacionado com a coordenação de grupos de robôs, fizemos a escolha de trabalhar sobre uma aplicação modelada como o problema de combate a incêndios florestais. Buscamos desenvolver um ambiente de simulação realístico, onde foram avaliadas quatro técnicas para busca de iii estratégias de formação do grupo: Algoritmos Genéticos, Otimização por Enxame de Partículas, Hill Climbing e (iv) Simulated Annealing. Com base nas diversas avaliações realizadas pudemos mostrar quais das técnicas e conjuntos de parâmetros permitem a obtenção de resultados mais acurados que os demais. Além disso, mostramos como uma heurística baseada em populações anteriores pode auxiliar na tolerância a falhas da operação. Relacionado com a tarefa de navegação, apresentamos o desenvolvimento de um veículo autônomo de grande porte funcional para ambientes externos. Buscamos aperfeiçoar uma arquitetura para navegação autônoma, baseada em visão monocular e com capacidade de seguir pontos esparsos de GPS. Mostramos como a simulação e os usos de robôs de pequeno porte auxiliaram no desenvolvimento do veículo de grande porte e apresentamos como as redes neurais podem ser aplicadas nos modelos de navegação autônoma. Na investigação sobre localização, mostramos um método utilizando informação obtida de redes sem fio para prover informação de localização para robôs móveis. As informações obtidas da rede sem fio são utilizadas para aprendizado da posição de um robô móvel por meio de uma rede neural. Diversas avaliações foram realizadas buscando entender o comportamento do sistema com diferentes números de pontos de acesso, com uso de filtros, com diferentes topologias. Os resultados mostram que o modelo usando redes sem fio pode ser um possível método prático e barato para localização de robôs móveis. Esta tese aborda temas relevantes e propostas originais relacionadas com os objetivos propostos, apresentando métodos que provenham autonomia na coordenação de grupos e nas atividades individuais dos mesmos. A busca por altos graus de eficiência na resolução de tarefas em ambientes dinâmicos ainda é um campo que carece de soluções e de um aprofundamento nas pesquisas. Sendo assim, esta pesquisa buscou agregar diversos avanços científicos na área de pesquisa de robôs móveis autônomos e coordenação de grupos, por meio da aplicação de estratégias inteligentes / The constant increasing of the complexity in the control of robotic systems, as well as the application of groups of robots assisting or replacing human beings in critical activities has generated a significant demand for more robust, flexible and efficient solutions. The conventional development of specialized algorithms consisted of rule-based systems and automatas, used to coordinate these physical sets in a dynamic environment is an extremely complex challenge. Although several models of development of robotic issues are currently in use, many challenges in the area remain open. This thesis is related to the search for intelligent strategies to be applied in autonomous mobile robots in order to allow practical operations in dynamic environments. We seek, with the investigation of intelligent strategies by means of the use of machine learning in the robots, to propose original solutions to allow contributions in three challenges of the robotic research area: localization, navigation and coordination of groups of robots. The investigations about localization and groups of robots show novel and original proposals, where we sought to extend the state of the art. The navigation part has as its major objective to be a link between the subjects of localization and navigation, being its aim to help the deployment of a autonomous vehicle implying in greater technical advances. Related to the robotic group coordination, we have made the choice to work on an application modeled as a wildfire combat operation. We have developed a simulation environment in which we have evaluated four techniques to obtain strategies for the group formation: genetic algorithms, particle swarm optimization, hill climbing and simulated annealing. The v results showed that we can have very different accuracy with different techniques and sets of parameters. Furthermore, we show how a heuristic based on the use of past populations can assist in fault tolerant operation. Related to the autonomous navigation task, we present the development of a large autonomous vehicle capable of operating in outdoor environments. We sought to optimize an architecture for autonomous navigation based on monocular vision and with the ability to follow scattered points of GPS.We show how the use of simulation and small robots could assist in the development of large vehicle. Furthermore, we show how neural networks can be applied as a controller to autonomous navigation systems. In the investigation about localization, we presented a method using wireless networks to provide information about localization to mobile robots. The information gathered by the wireless network is used as input in an artificial neural network which learns the position of the robot. Several evaluations were carried out in order to understand the behavior of the proposed system, as using different topologies, different numbers of access points and the use of filters. Results showed that the proposed system, using wireless networks and neural networks, may be a useful and easy to use solution for localization of mobile robots. This thesis has addressed original and relevant topics related to the proposed objectives, showing methods to allow degrees of autonomy in robotic operations. The search for higher degrees of efficiency in tasks solving in dynamic environments is still a field that lacks solutions. Therefore, this study sought to add several scientific contributions in the autonomous mobile robots research area and coordination of groups, by means of the application of intelligent strategies

Orientadores: Fernando Jose Von Zuben, Leandro Nunes de Castro Silva / Tese (doutorado) - Universidade Estadual de Campinas, Faculdade de Engenharia Eletrica e Computação / Made available in DSpace on 2018-08-06T14:08:06Z (GMT). No. of bitstreams: 1 Vargas_PatriciaAmancio_D.pdf: 4626705 bytes, checksum: b203a58e3f5f1c613db0787b3e396196 (MD5) Previous issue date: 2005 / Resumo: Este trabalho propõe uma classificação circunstancial para sistemas complexos, incluindo uma estrutura unificada de descrição a ser empregada na análise e síntese de sistemas computacionais bio-inspirados. Como um ramo dos sistemas complexos organizados, os sistemas computacionais bio-inspirados admitem uma sub-divisão em sistemas de inteligência computacional e sistemas homeostáticos artificiais. Com base neste formalismo, duas abordagens híbridas são concebidas e aplicadas em problemas de navegação autônoma de robôs. A primeira abordagem envolve sistemas classificadores com aprendizado e sistemas imunológicos artificiais, visando explorar conjuntamente conceitos intrínsecos a sistemas complexos, como auto-organização, evolução e cognição dinâmica. Fundamentada nas interações neuro-imuno-endócrinas do corpo humano, a segunda abordagem propõe um novo modelo de sistema homeostático artificial, explorando mudanças de contexto e efeitos do meio sobre o comportamento autônomo de um robô móvel. Embora preliminares, os resultados obtidos envolvem simulação computacional em ambientes virtuais e alguns experimentos com robôs reais, permitindo extrair conclusões relevantes acerca do potencial das abordagens propostas e abrindo perspectivas para a síntese de sistemas complexos adaptativos de interesse prático / Abstract: This work proposes a circumstantial classification for complex systems, including a unified description structure to be employed in the analysis and synthesis of biologically inspired computing metaphors. Considered as a branch of organized complex systems, these bio-inspired computing frameworks may be subdivided into computation intelligence systems and artificial homeostatic systems. Developed under this formalism, two novel hybrid systems are conceived and applied to robot autonomous navigation problems. The first approach involves learning classifier systems and artificial immune systems, in an attempt to investigate intrinsic concepts of complex systems as self-organization, evolution, and dynamic cognition. Drawn on the principles of the human nervous, immune and endocrine systems, the second approach envisages a new model of an artificial homeostatic system to explore context changes and environmental effects on the behaviour of an autonomous robotic agent. Though preliminary, the obtained results encompass computer simulation on virtual environments in addition to a number of real robot¿s experiments. Relevant conclusions can be invoked, mainly related to the potentiality of the proposed frameworks, thus opening attractive prospects for the synthesis of complex adaptive systems of practical interest / Doutorado / Engenharia de Computação / Doutor em Engenharia Elétrica

Cette thèse s'inscrit dans la problématique de la perception des véhicules autonomes, qui doivent notamment être capables de détecter et de positionner à tout moment les éléments fixes et mobiles de leur environnement. Les besoins sont ensuite multiples, de la détection d'obstacles à la localisation du porteur dans l'espace, et de nombreuses méthodes de la littérature s'y attellent. L'objectif de cette thèse est de reconstituer, à partir de prises de vues de stéréo-vision, une carte en trois dimensions décrivant l'environnement proche ; tout en effectuant une détection, localisation et suivi dans le temps des objets mobiles.La détection et le suivi dans le temps d'un grand nombre de points d'intérêt constitue une première étape. Après avoir effectué une comparaison exhaustive de divers détecteurs de points d'intérêt de la littérature, on propose pour réaliser le suivi de points une implémentation massivement parallélisée de l'algorithme KLT, dans une configuration redondante réalisée pendant cette thèse. Cette implémentation autorise le suivi fiable de milliers de points en temps réel, et se compare favorablement à l'état de l'art.Il s'agit ensuite d'estimer le déplacement du porteur, et de positionner ces points dans l'espace, tâche pour laquelle on propose une évolution robuste d'une procédure bien connue, dite "SVD", suivie d'un filtrage par UKF, qui nous permettent d'estimer très rapidement le mouvement propre du porteur. Les points suivis sont ensuite positionnés dans l'espace, en prenant en compte leur possible mobilité, en estimant continuellement la position la plus probable compte tenu des observations successives.La détection et le suivi des objets mobiles font l'objet d'une dernière partie, dans laquelle on propose une segmentation originale tenant compte des aspects de position et de vitesse. On exploite ainsi une des singularités de notre approche, qui conserve pour chaque point positionné un ensemble cohérent de positions dans le temps. Le filtrage et le suivi des cibles se basent finalement sur un filtre GM-PHD. / This PhD work is to be seen within the context of autonomous vehicle perception, in which the detection and localisation of elements of the surroundings in real time is an obvious requirement. Subsequent perception needs are manyfold, from localisation to obstacle detection, and are the subject of a continued research interest. The goal of this work is to build, in real time and from stereovision acquisition, a 3D map of the surroundings ; while detecting and tracking moving objects.Interest point selection and tracking on picture space are a first step, which we initiate by a thorough comparison of detectors from the literature. As regards tracking, we propose a massively parallel implementation of the standard KLT algorithm, using redundant tracking to provide reliable quality estimation. This allows us to track thousands of points in real-time, which compares favourably to the state of the art.Next step is the ego-motion estimation, along with the positioning of tracked points in 3D space. We first propose an iterative variant of the well known “SVD” process followed by UKF filtering, which allows for a very fast and reliable estimation. Then the position of every followed interest point is filtered on the fly over time, in contrast to most dense approaches from the literature.We finally propose a segmentation of moving objects in the augmented position-speed space, which is made possible by our continuous estimation of feature points position. Target tracking and filtering finally use a GM-PHD approach.

Este projeto consiste em um sistema de navegação autônomo baseado em redes neurais nebulosas modulares capacitando o robô a alcançar alvos, ou pontos metas, em ambientes desconhecidos. Inicialmente, o sistema não tem habilidade para a navegação, após uma fase de experimentos com algumas colisões, o mecanismo de navegação aprimora-se guiando o robô ao alvo de forma eficiente. Uma arquitetura híbrida inteligente é apresentada para este sistema de navegação, baseada em redes neurais artificiais e lógica nebulosa. A arquitetura é hierárquica e costitiui-se de dois módulos responsáveis por gerar comportamentos inatos de desvio de obstáculos e de busca ao alvo. Um mecanismo de aprendizagem por reforço, baseada em uma extensão da lei de Hebb, pondera os comportamentos inatos conflitantes ajustando os pesos sinápticos das redes neurais nos instantes de captura do alvo e de colisão contra obstáculos. A abordagem consolidada em simulação é validada em ambientes reais neste trabalho. Para tanto, este sistema foi implementado e testado no simulador Saphira, ambiente de simulação que acompanha o robô Pioneer I e que denota um estágio anterior aos testes em ambientes reais por apresentar comportamentos do robô similares aos comportamentos do robô móvel. Modificações na arquitetura híbrida foram necessárias para adaptar o sistema de navegação simulado ao sistema incorporado no Pioneer I. Experimentos em ambientes reais demonstraram a eficiência e a capacidade de aprendizagem do sistema de navegação, validando a arquitetura híbrida inteligente para aplicação em robôs móveis / This project consists in a autonomous navigation system based on modular neuro-fuzzy networks that is able to guide the robot in unknown environments from a initial point to the goal. Initially, the system is not able to navigate, but after a trial and error period and some collisions, it improves in guiding the robot to the goal efficiently. A intelligent hybrid architecture is presented for this naviga tion system based on artificial neural networks and fuzzy logic. This architecture is hierarquical and consists in two modules that generate innate behaviors, like obstacles avoiding and target reaching. A reinforcement learning mecanism, based on the extended Hebb law, balances this conflicting innate behaviors adjusting the neural network synaptic weights as obstacle and collision avoidance and target reaching takes place. In this project, the approach is consolidated in simulation and validated in real environments. To this end, this system has been implemented by using Saphira simulator and Pioneer I simulation environment. This simulated evironment is a previous stage of tests performed real time and presents simulated robot behaviors similar to real mobile robot behaviors. The hybrid architecture was modified to adapt the simulated navigation system into Pioneer I software. Experiments in a real environments show the efficiency and learning capabilities of the navigation system, validating the intelligent hybrid architecture for mobile robots applications

<div><div><div><p>Globally, there are more than 727 million people aged 65 years and older in the world, and the elderly population is expected to grow more than double in the next three decades. Families search for affordable and quality care for their senior loved ones will have an effect on the care-giving profession. A personal robot assistant could help with daily tasks such as carrying things for them and keeping track of their routines, relieving the burdens of human caregivers. Performing mentioned tasks usually requires the robot to autonomously navi- gate. An autonomous navigation robot should collect the knowledge of its surroundings by mapping the environment, find its position in the map and calculate trajectories by avoiding obstacles. Furthermore, to assign specific tasks which are in various locations, robot has to categorize the rooms in addition to memorizing the respective coordinates. In this research, methods have been developed to achieve autonomous navigation and room categorization of a mobile robot within indoor environments. A Simultaneously Localization and Map- ping (SLAM) algorithm has been used to build the map and localize the robot. Gmapping, a method of SLAM, was applied by utilizing an odometry and a 2D Light Detection and Ranging (LiDAR) sensor. The trajectory to achieve the goal position by an optimal path is provided by path planning algorithms, which is divided into two parts namely, global and local planners. Global path planning has been produced by DIJKSTRA and local path planning by Dynamic Window Approach (DWA). While exploring new environments with Gmapping and trajectory planning algorithms, rooms in the generated map were classified by a powerful deep learning algorithm called Convolutional Neural Network (CNN). Once the environment is explored, the robots localization in the 2D space is done by Adaptive Monte Carlo Localization (AMCL). To utilize and test the methods above, Gazebo software by The Robotic Operating System (ROS) was used and simulations were performed prior to real life experiments. After the trouble-shooting and feedback acquired from simulations, the robot was able to perform above tasks and later tested in various indoor environments. The environment was mapped successfully by Gmapping and the robot was located within the map by AMCL. Compared to the theoretical maximum efficient path, the robot was able to plan the trajectory with acceptable deviation. In addition, the room names were classified with minimum of 85% accuracy by CNN algorithm. Autonomous navigation results show that the robot can assist elderly people in their home environment by successfully exploring, categorizing and navigating between the rooms.</p></div></div></div>

[pt] Neste trabalho, aborda-se a modelagem e controle de um quadricóptero para navegação autônoma em ambientes agrícolas. Os modelos cinemático e dinâmico do veículo aéreo são computados a partir do formalismo de Newton-Euler, incluindo efeitos aerodinâmicos e características das hélices. O sistema de movimento do quadricóptero pode ser dividido em dois subsistemas, um translacional e outro rotacional, responsáveis pelo controle de posição nos eixos x, y, z, and atitude do veículo no espaço Cartesiano. A primeira abordagem de controle é linear, se presenta dois controladores, um controlador proporcional-derivativo (PD) e o adaptativo baseado no espaço de estados. A segunda abordagem é não-linear e baseada em um controlador adaptativo a fim de lidar com a presença de incertezas nos parâmetros do sistema. Simulações numéricas são executadas em Matlab para ilustrar o desempenho e a viabilidade da metodologia de controle proposta. Simulações computacionais 3D são executadas em Gazebo para verificar a navegação autônoma em um campo agrícola. / [en] In this work, we address the modeling and control design of a quadrotor for autonomous navigation in agricultural environments. The kinematic and dynamic models of the aerial vehicle are derived following the Newton-Euler formalism. The motion system of the quadrotor can be split into two subsystems, that is, translational and rotational subsystems, responsible for controlling the position along the longitudinal, transverse and vertical axes of the Cartesian space as well as its orientation about the corresponding axes. The first linear control approach is based on the proportional-derivative (PD) controller, whereas the second nonlinear control approach is based on an adaptive controller in order to deal with the presence of uncertainties in the system parameters. Numerical simulations are carried out in Matlab to illustrate the performance and feasibility of the proposed control methodology. Gazebo was used to perform the 3D simulations for verifying autonomous navigation in agricultural fields.

Autonomous drone navigation in classical approaches typically involves constructing a map representation and employing path planning and collision checking algorithms within that map. Recently, novel deep learning techniques combined with depth camera observations have emerged as alternative approaches capable of achieving comparable collision-free performance. While these methods have demonstrated effective collision-free performance in dense environments, they rely on low-noise range or visual data, which may not be feasible in extreme degraded environments characterized by factors such as dust, smoke, weak geometries, or low-texture areas. A possible alternative is to leverage recent progress in mmWave radar imaging, which previously has produced data of insufficient resolution for such purposes. Through the use of a Variational Autoencoder and existing collision prediction algorithms, the goal of this study is to prove the use of mmWave radar for navigating difficult environments. The results of the study exhibit successful navigation in simulated scenarios featuring sparse obstacles. Additionally, results of utilizing real-world mmWave radar data in example scenarios is provided to demonstrate the potential for further application of this technology. / Autonom navigation för drönare i klassiska tillvägagångssätt innebär vanligtvis att man konstruerar en kartrepresentation och använder vägplanerings- och kollisionskontrollalgoritmer inom den kartan. Nyligen har nya djupinlärningstekniker kombinerat med djupkameraobservationer framträtt som alternativa tillvägagångssätt som kan uppnå jämförbar prestanda utan kollisioner. Även om dessa metoder har visat effektiv prestanda utan kollisioner i täta miljöer, är de beroende av störningsfria avstånds- eller visuella data, vilket kanske inte är genomförbart i extrema försämrade miljöer som karakteriseras av faktorer som damm, rök, svaga geometrier eller områden med låg textur. Ett möjligt alternativ är att dra nytta av de senaste framstegen inom mmWave-radaravbildning, vilket tidigare har producerat data med otillräcklig upplösning för sådana ändamål. Genom användning av en varieabel autoencoder och befintliga kollisionsprognosalgoritmer syftar denna studie till att bevisa användningen av mmWave-radar för att navigera i svåra miljöer. Resultaten från studien visar framgångsrik navigering i simulerade scenarier med glesa hinder. Dessutom presenteras resultat från användning av verkliga mmWave-radardata i exempelscenarier för att visa potentialen för ytterligare tillämpningar av denna teknik.

<p dir="ltr">The Global Navigation Satellite System (GNSS) is heavily relied on for the navigation of Earth satellites. For satellites in cislunar space and beyond, GNSS is not readily available. As a result, other sources such as NASA's Deep Space Network (DSN) must be relied on for navigation. However, DSN is overburdened and can only support a small number of satellites at a time. Furthermore, communication with external sources can become interrupted or deprived in these environments. Given NASA's current efforts towards cislunar space operations and the expected increase in cislunar satellite traffic, there will be a need for more autonomous navigation options in cislunar space and beyond.</p><p dir="ltr">In this thesis, a navigation system capable of accurate and computationally efficient orbit determination in these communication-deprived environments is proposed and investigated. The emphasis on computational efficiency is in support of cubesats which are constrained in size, cost, and mass; this makes navigation even more challenging when resources such as GNSS signals or ground station tracking become unavailable.</p><p dir="ltr">The proposed navigation system, which is called GRAVNAV in this thesis, involves a two-satellite formation orbiting a planet. The primary satellite hosts an Extended Kalman Filter (EKF) and is capable of measuring the relative position of the secondary satellite; accurate attitude estimates are also available to the primary satellite. The relative position measurements allow the EKF to estimate the absolute position and velocity of both satellites. In this thesis, the proposed navigation system is investigated in the two-body and three-body problems.</p><p dir="ltr">The two-body analysis illuminates the effect of the gravity model error on orbit determination performance. High-fidelity gravity models can be computationally expensive for cubesats; however, celestial bodies such as the Earth and Moon have non-uniform and highly-irregular gravity fields that require complex models to describe the motion of satellites orbiting in their gravity field. Initial results show that when a second-order zonal harmonic gravity model is used, the orbit determination accuracy is poor at low altitudes due to large gravity model errors while high-altitude orbits yield good accuracy due to small gravity model errors. To remedy the poor performance for low-altitude orbits, a Gravity Model Error Compensation (GMEC) technique is proposed and investigated. Along with a special tuning model developed specifically for GRAVNAV, this technique is demonstrated to work well for various geocentric and lunar orbits.</p><p><br></p><p dir="ltr">In addition to the gravity model error, other variables affecting the state estimation accuracy are also explored in the two-body analysis. These variables include the six Keplerian orbital elements, measurement accuracy, intersatellite range, and satellite formation shape. The GRAVNAV analysis shows that a smaller intersatellite range results in increased state estimation error. Despite the intersatellite range bounds, semimajor axis, measurement model, and measurement errors being identical for both orbits, the satellite formation shape also has a strong influence on orbit determination accuracy. Formations that place both satellites in different orbits significantly outperform those that place both satellites in the same orbit.</p><p dir="ltr">The three-body analysis primarily focuses on characterizing the unique behavior of GRAVNAV in Near Rectilinear Halo Orbits (NRHOs). Like the two-body analysis, the effect of the satellite formation shape is also characterized and shown to have a similar impact on the orbit determination performance. Unlike the two-body problem, however, different orbits possess different stability properties which are shown to significantly affect orbit determination performance. The more stable NRHOs yield better GRAVNAV performance and are also less sensitive to factors that negatively impact performance such as measurement error, process noise, and decreased intersatellite range.</p><p dir="ltr">Overall, the analyses in this thesis show that GRAVNAV yields accurate and computationally efficient orbit determination when GMEC is used. This, along with the independence of GRAVNAV from GNSS signals and ground-station tracking, shows that GRAVNAV has good potential for navigation in cislunar space and beyond.</p>

This thesis project is part of an internship at Visual Behavior. The company aims at producing computer vision models for robotics, helping the machine to better understand the world through the camera eye. The image holds many features that deep learning models are able to extract: navigable area, depth inference and object detection. Example of recent advances are the RAFTstereo model [1] to infer or refine depth features from stereo images, or the end-to-end Object detection model DETR [2]. The field of autonomous navigation can then benefit from these advanced features to propose better path planning methods. In particular, to help the deployment of ground robots in human crowded environments, the robots behavior must not only be safe but it must also look smart so as to inspire trust. This thesis proposes a local path planner based on the Dynamic Window Approach [3] using a scoring function inspired from the Velocity Obstacle method [4] so as to benefit from the flexibility of the first and the long-term anticipation of the second. The proposed method can induce a smart behavior by setting the robot on safe tracks from a long time horizon without increasing the time to reach a positional goal, compared to a closer-ranged strategy inspired from the DW4DO method [5]. This improves the robot’s ability to deal with several moving obstacles and to avoid engaging in already occupied corridors. The code produced in this thesis uses ROS and the gazebo simulator and is available in the following git page https://github.com/FloCoic oi/fc_thesis along with the minimal instructions to run the install and get started to quickly run a demo. / Detta examensarbete är en del av en praktik på Visual Behavior. Företaget har som mål att ta fram modeller för datorseende för robotar som hjälper maskinen att bättre förstå världen genom kamerans öga. Bilden innehåller många egenskaper som modeller för djupinlärning kan extrahera: navigerbart område, djupinferens och objektsdetektering. Exempel på nya framsteg är RAFT-stereo-modellen [1] för att härleda eller förfina djupegenskaper från stereobilder, eller ”end-to-end” objektdetektering modellen DETR [2]. Inom området autonom navigering kan man sedan dra nytta av dessa avancerade funktioner för att föreslå bättre metoder för vägplanering. För att underlätta användningen av markrobotar i miljöer med mycket människor måste robotarnas beteende inte bara vara säkert utan också se smart ut så att de väcker förtroende. I den här avhandlingen föreslås en lokal vägplanerare som bygger på Dynamic Window Approach [3] och som använder en poängfunktion inspirerad av Velocity Obstacle metoden [4] för att dra nytta av flexibiliteten hos den första metoden och den långsiktiga förutsebarheten hos den andra. Den föreslagna metoden kan framkalla ett smart beteende genom att sätta roboten på säkra spår på lång sikt utan att öka tiden för att nå ett positionsmål, jämfört med en strategi med närmare avstånd som inspirerats av DW4DOmetoden [5]. Detta förbättrar robotens förmåga att hantera flera rörliga hinder och att undvika att gå in i redan upptagna korridorer. Koden som produceras i denna avhandling använder ROS och gazebosimulatorn och finns tillgänglig på följande git-sida https://github.c om/FloCoicoi/fc_thesis tillsammans med minimala instruktioner för att köra installationen och komma igång för att snabbt köra en demo.