Industrial robots rapidly gained popularity as they can perform tasks quickly, repeatedly and accurately in static environments. However, in modern manufacturing, robots should also be able to safely interact with arbitrary objects and dynamically adapt their behavior to various situations. The large masses and rigid constructions of industrial robots prevent them from easily being re-tasked. In this context, this work proposes an immediate solution to make rigid manipulators compliant and able to efficiently handle object interactions, with only an add-on module (a custom designed instrumented compliant wrist) and an original control framework which can easily be ported to different manipulators. The proposed system utilizes both offline and online trajectory planning to achieve fully automated object interaction and surface following with or without contact where no prior knowledge of the objects is available. To minimize the complexity of the task, the problem is formulated into four interaction motion modes: free, proximity, contact and a blend of those. The free motion mode guides the robot towards the object of interest using information provided by a RGB-D sensor. The RGB-D sensor is used to collect raw 3D information on the environment and construct an approximate 3D model of an object of interest in the scene. In order to completely explore the object, a novel coverage path planning technique is proposed to generate a primary (offline) trajectory. However, RGB-D sensors provide only limited accuracy on the depth measurements and create blind spot when it reaches close to surfaces. Therefore, the offline trajectory is then further refined by applying the proximity motion mode and contact motion mode or a blend of them (blend motion mode) that allow the robot to dynamically interact with arbitrary objects and adapt to the surfaces it approaches or touches using live proximity and contact feedback from the compliant wrist. To achieve seamless and efficient integration of the sensory information and smoothly switch between different interaction modes, an original hybrid switching scheme is proposed that applies a supervisory (decision making) module and a mixture of hard and blend switches to support data fusion from multiple sensing sources by combining pairs of the main motion modes. Experimental results using a CRS-F3 manipulator demonstrate the feasibility and performance of the proposed method.

Um aspecto importante da interação humana é a atenção compartilhada. Ela é um processo de comunicação onde uma pessoa redireciona a sua atenção para um objeto ou evento e a outra pessoa ou pessoas seguem o seu olhar para o mesmo lugar. O processo é finalizado com a pessoa que segue a atenção realizando um apontamento sobre o objeto e um comentário sobre a situação. Esta habilidade importante é aprendida por nós durante o período da infância e hoje, alguns pesquisadores em robótica estão tentando desenvolver arquiteturas robóticas para aprender essa habilidade em robôs. Deste modo, o laboratório de aprendizado de robôs está trabalhando em uma arquitetura robótica para esse fim. Ela é composta por três módulos, percepção de estímulo, controle de consequência e emissão de resposta. Esta arquitetura robótica foi avaliada no controle de uma cabeça robótica e foi capaz de aprender a seguir o olhar e identificar alguns objetos. No entanto, todos esses módulos têm algumas limitações. A fim de ter uma melhor interação entre um robô e um humano e reduzir os efeitos das limitações, algumas melhorias foram desenvolvidas. Entre elas incluem um novo algoritmo de classificação das posições da cabeça através do histograma de gradiente orientado, inserir novas funcionalidades (definidas como reflexos) ao módulo de controle de consequência e novos algoritmos de aprendizado para selecionar a melhor ação. Todas as modificações realizadas reduziram as limitações e pode melhorar as interações entre um robô e um ser humano / One important aspect of human interaction is the shared attention. It is a communication process where one person redirect his or her attention to an object or event and the other person or people follow gaze to the same place. This process end with a pointing and a comment about the situation by the person that follows the attention. This important ability was learned by us during the childhood and some roboticist are trying to develop robotics architectures to learn this ability in robots. In this way, the Learning Lab Robotics has been working on a robotic architecture used with this proposed. It is composed by three modules, stimulus perception, consequence control and response emission. This robotic architecture was evaluated to control a robotic head and it was capable to learn to follow gaze and identify some objects. However, all of these modules have some limitations. In order to take a better interaction between a robot and a human and reduce the effects of limitations, some improvements were developed. They include a new head pose classification algorithm using histogram of oriented gradient, increase the capability of consequence control with new reflexes and new learning algorithms to select the best action. All modification reduce the limitations and it can improve the interactions between a robot and a human being

The diploma thesis deals with a new approach to create a robotic arm trajectory. The first part descibes existing methods of trajectory planning. The second part shows virtual point method, implementation of the trajectory planning into the web environment and trajectory generation. Third part descibes the virtual robot model and designed comunnication protocols. The fourth part shows grafical user interface and his posibilities. The last part presents implemention of web aplication into industrial visualization.

Robotization of industry is very important in these days. It is possible to replace human workers with robots and because of that make manufacturing faster. This master’s thesis includes feasibility study of robotic welding for stainless steel part guide rail. Quality of robotic welding was evaluated on testing pieces. Robotic welding was 88,5 % faster than current welding method. Costs of manufacture of one piece were 40 % cut.

Cette thèse s’intéresse aux effets des technologies d’assistance robotique sur l’expertise individuelle et collective des médecins dans un bloc opératoire de chirurgie gastrique. Notre recherche est fondée sur l’analyse de l’émergence des routines organisationnelles et de leur mise en évidence en mobilisant l’approche naturaliste de la décision. D’un point de vue managérial, nos résultats démontrent que la technologie introduite atténue l’expertise individuelle en dégradant respectivement la détection des signes cliniques et la coordination au sein de l’équipe. Nous établissons des recommandations concernant la mise en œuvre d’une technologie dans le cadre des équipes restreintes. Notre méthodologie consiste à fusionner des sources hétérogènes : entretiens, observation in situ, vidéos, débriefing, auto-confrontation... elle ouvre ainsi la voie à de nouvelles stratégies d’analyse de données jusqu’à lors essentiellement utilisées en psychologie. Enfin, nos contributions théoriques ont permis un enrichissement de l’approche naturaliste en lui adjoignant les rôles tenus par le leader et la mise en évidence du rôle de l’expertise individuelle dans la formation de nouvelles routines fonctionnelles. / The aim of this research is the understand the effect of a new technology, a robotic system, on the individual and collective expertise of practitioners in bariatric surgery. Our analysis is based on the emergence of organizational routines while taking into consideration naturalistic decision making approach. From the managerial perspective, our results demonstrate that the implementation of this technology decrease individual expertise. This is done through the degradation of individual situation awareness and coordination between team members. We established recommendations for an efficient technology implementation by restricted teams. Our methodology consists of mixing heterogenic sources: interviews, observations, videos, debriefing with auto-confrontation… this opens the door for new analysis strategies till now used mainly in psychology. Finally, our theoretical contributions reinforced the naturalistic approach while insisting on the important role of a team leader and the role of individual expertise in the development of new functional organizational routines.

Friction Stir Welding has seen a fast uptake in many industry segments. Mechanical properties superior to fusion welding, the ability to weld "unweldable" aluminium alloys and low distortion are often described as the main reasons for the fast industrial implementation of FSW. Most existing applications consist of long straight welding joints. Applications with complex weld geometries, however, are rarely produced by FSW. These geometries can induce thermal variations during the welding process, thus making it challenging to maintain a consistent weld quality. In-process adaptation of weld parameters to respond to geometrical variations and other environmental variants allow new design opportunities for FSW. Weld quality has been shown to be reliant on the welding temperature. However, the optimal methodology to control the temperature is still under development.The research work presented in this thesis focuses on some steps to take in order to reach the improvement of the FSW temperature controller, thus reach a better and consistent weld quality. In the present work different temperature methods were evaluated. Temperature measurements acquired by the tool-workpiece thermocouple (TWT) method were accurate and fast, and thereby enhanced suitable for the controller. Different environmental conditions influencing the material heat dissipation were imposed in order to verify the controller effect on the joint quality. In comparison with no controlled weld, the use of the controller enabled a fast optimization of welding parameters for the different conditions, leading to an improvement of the mechanical properties of the joint.For short weld lengths, such as stitch welds, the initial plunge and dwell stages occupy a large part of the total process time. In this work temperature control was applied during these stages. This approach makes the plunge and dwell stages more robust by preventing local material overheating, which could lead to a tool meltdown. The TWT method was demonstrated to allow a good process control during plunging and continuous welding. The approach proposed for control offers weld quality consistency and improvement. Also, it allows a reduction of the time required for the development of optimal parameters, providing a fast adaptation to disturbances during welding and, by decreasing the plunge time, provides a significant decrease on the process time for short welds.

Det faktum att RPA är ett så pass nytt fenomen, kan till viss del utmana implementeringsprocessen på grund av otillräcklig erfarenhet hos involverade medarbetare samt bristande utbud av allmän information om ämnet. För att hantera denna utmaning kan ett agilt arbetssätt underlätta processen. Trots att RPA och agilt arbetssätt kombineras i stor utsträckning inom näringslivet, finns det tämligen lite forskning kring kombinationen - vilket har bidragit till vårt forskningsgap. Utifrån detta forskningsgap har vi vidare lyckats formulera följande frågeställning: ​Hur kan en robotisk processautomatisering (RPA) implementeras i ekonomifunktionen, med hjälp av ett agilt arbetssätt? Fortsättningsvis är studiens genomgående syftet att beskriva och analysera hur RPA kan implementeras i en etablerad ekonomifunktion, genom användning av ett agilt tillvägagångssätt. För att lyckas generera ett nyanserat och träffsäkert svar till frågeställningen, har svar inhämtats via åtta intervjuer. En teoretisk grund har även inhämtats i form av ett antal vetenskapliga artiklar, för att skapa en vetenskaplig tyngd i arbetet. Analys har därefter gjorts utifrån en jämförelse mellan teori och empiri i syfte att kunna framställa en allmän slutsats kopplat till problemet. Metoden som använts är således en kvalitativ studie baserad på en induktiv ansats. I slutsatsen konstateras att RPA-implementering och agilt arbetssätt är en hållbar metod att arbeta i. Vidare har ett antal olika implementeringslösningar presenterats; därav ett antal konkreta metoder inom agilt arbetssätt samt ett antal modeller inom RPA. Däremot drogs slutsatsen att nyckeln till framgång är ett öppet mindset från medarbetarnas sida; där medarbetaren förstår värdet av varför agilt arbetssätt används i den specifika kontexten. Avslutningsvis presenterar även vi som författare våra uppfattningar om ämnet baserat på de empiriska och teoretiska intryck vi fått uppleva under studiens gång.

Obstacle avoidance is one of the core problems in the field of autonomous navigation. An obstacle avoidance approach is developed for the navigation task of a reconfigurable multi-robot system named STORM, which stands for Self-configurable and Transformable Omni-Directional Robotic Modules. Various mathematical models have been developed in previous work in this field to avoid collision for such robots. In this work, the proposed collision avoidance algorithm is trained via Deep Reinforcement Learning, which enables the robot to learn by itself from its experiences, and then fit a mathematical model by updating the parameters of a neural network. The trained neural network architecture is capable of choosing an action directly based on the input sensor data using the trained neural network architecture. A virtual STORM locomotion module was trained to explore a Gazebo simulation environment without collision, using the proposed collision avoidance strategies based on DRL. The mathematical model of the avoidance algorithm was derived from the simulation and then applied to the prototype of the locomotion module and validated via experiments. Universal software architecture was also designed for the STORM modules. The software architecture has extensible and reusable features that improve the design efficiency and enable parallel development. / Master of Science / In this thesis, an obstacle avoidance approach is described to enable autonomous navigation of a reconfigurable multi-robot system, STORM. The Self-configurable and Transformable Omni-Directional Robotic Modules (STORM) is a novel approach towards heterogeneous swarm robotics. The system has two types of robotic modules, namely the locomotion module and the manipulation module. Each module is able to navigate and perform tasks independently. In addition, the systems are designed to autonomously dock together to perform tasks that the modules individually are unable to accomplish. The proposed obstacle avoidance approach is designed for the modules of STORM, but can be applied to mobile robots in general. In contrast to the existing collision avoidance approaches, the proposed algorithm was trained via deep reinforcement learning (DRL). This enables the robot to learn by itself from its experiences, and then fit a mathematical model by updating the parameters of a neural network. In order to avoid damage to the real robot during the learning phase, a virtual robot was trained inside a Gazebo simulation environment with obstacles. The mathematical model for the collision avoidance strategy obtained through DRL was then validated on a locomotion module prototype of STORM. This thesis also introduces the overall STORM architecture and provides a brief overview of the generalized software architecture designed for the STORM modules. The software architecture has expandable and reusable features that apply well to the swarm architecture while allowing for design efficiency and parallel development.

The modern painting process in automotive industry is complex and a lot of factors affect the result. The trial and error method is used today to control the quality and introduce new colours. This method takes a lot of time and does not show any clear numbers of how the process is affected by changing the parameters. During this thesis, we have investigated a delimited number of parameters. The work is based on experiments performed on samples that represents a flat surface of the cab, to reduce experimental costs. Our master thesis is done at Scania in Oskarshamn, where all the cabs for the European production is produced. The objectives with this thesis has been to explain how the process parameters of the robotic applicator affect the paint distribution, paint thickness and the colour of the top coat. We also optimised the process by finding which settings gives an even paint distribution, a correct thickness and an accepted colour of the top coat. We have been using Design of Experiments to achieve the goals of this study. Design of Experiments is a statistic method that is used to perform experiments effectively. It also shows the effect of changing the factors from a low to a high level. We have chosen to divide the workflow into three parts: screening, optimisation and confirmation. The experiments are performed during the daily production to replicate the real circumstances. The shape air, paint flow and high rotation is the most important parameters to control. Paint flow also seems to have a linear impact on the thickness of the top coat layer. The Shape air and the high rotation on the other hand mainly affect the distribution of the top coat layer. Different levels are needed for the shape air and high rotation depending on what paint flow is used. The optimal settings of the factors for our colour were found to be paint flow at 82 %, the shape air at 90 %, the high rotation at 90 % and high voltage at 100 %. The optimal settings give a result of 1,535 μm in spread and 40,08 μm in mean thickness. Our settings compared to today’s results contributes to a reduced paint consumption, better quality and therefore less rework.

This thesis describes the design of a robotic inspection tool which is based on a differential track-drive platform. The robotic inspection tool is a one man-portable UGV that has been developed for the purpose of non-destructive evaluation (NDE) and internal inspection of environments where human penetration may be difficult or hazardous. Various NDE and sensing techniques are described in this paper but the focus is on the mechanical and electrical design of the platform itself. The platform is a versatile device for mobile robotics research and development that supports a wide variety of instrumentation and payloads. Variable height control of the payload is achieved with a scissor lift assembly that provides accurate positioning of equipped sensors and instrumentation. The architecture of the platform was designed to support future autonomous implementations. / Master of Science