In the field of robotics and autonomous vehicles, the use of RGB-D data and LiDAR sensors is a popular practice for applications such as SLAM[14], object classification[19] and scene understanding[5]. This thesis explores the problem of semantic segmentation using deep multimodal fusion of LRF and depth data. Two data set consisting of 1080 and 108 data points from two scenes is created and manually labeled in 2D space and transferred to 1D using a proposed label transfer method utilizing hierarchical clustering. The data set is used to train and validate the suggested method for segmentation using a proposed dual encoder-decoder network based on SalsaNet [1] with gradual fusion in the decoder. Applying the suggested method yielded an improvement in the scenario of an unseen circuit when compared to uni-modal segmentation using depth, RGB, laser, and a naive combination of RGB-D data. A suggestion of feature extraction in the form of PCA or stacked auto-encoders is suggested as a further improvement for this type of fusion. The source code and data set are made publicly available at https://github.com/Anguse/salsa_fusion.

Prediction of both big and small decisions is something most companies have to make on a daily basis. The importance of having a highly accurate technique for different decision-making is not something that is new. However, even though the importance of prediction is a fact to most people, current techniques for estimation are still often highly inaccurate. The consequences of an inaccurate prediction can be huge in the differences between the misclassifications. Not just in the industry but for many different areas. Machine learning have in the recent couple of years improved significantly and are now considered a reliable method to use for prediction. The main goal of this research is to predict casting defects with the help of a machine-learning algorithm based on process parameters. In order to achieve the main goal, some sub-objectives have been identified to successfully reach those goals. A problem when dealing with machine learning is an unbalanced dataset. When training a network, it is essential that the dataset is balanced. In this research we have successfully balanced the dataset. Undersampling was the method used in our research to establish our balanced dataset. The research compares and evaluates a couple of different undersample methods in order to see which undersampling is best suited for this project. Three different machine models, “random forest”, “artificial neural network”, and “k-nearest neighbor”, are also compared to each other to see what model performs best. The conlcusion reached was that the best method for both undersampling and machine learning model varied due to many different reasons. So, in order to find the best model with the best method for a specific job, all the models and methods need to be tested. However, the undersampling method that provided best performances most times in our research was the NearMiss version 2 model. Artificial Neural Network was the machine learning model that had most success in our research. It performed best in two out of three evaluations and comparisons.

Millions of people suffer from heart failure worldwide. The need for heart donations started the development of mechanical circulatory support systems. A suction phenomenon can occur in the artificial heart when not enough blood is available. Due to occlusion, suction in the artificial heart can cause the arteries to collapse and have fatal consequences. This thesis is in collaboration with Scandinavian Real Heart AB and follows the implementation of a preemptive atrial suction detection algorithm for the total artificial heart (TAH) developed by Realheart. The main limitation is the number of sensors available to collect data from, restricted to a pressure and current sensor. The data used in the thesis is collected on a mock loop that simulates the pressures in the human body. The implementation follows an iterative process where different Artificial Intelligence algorithms are tested and evaluated. The final algorithm uses a recurrent neural network (RNN) for classification and is evaluated based on the accuracy and the number of seconds before suction occurs. The results show that the RNN can preemptively classify the data one second before it occurs. The algorithm assumes suction to happen one second before it occurs, preemptively detecting suction. The results from this thesis enable a continuation that can improve the development of TAHs. Future work includes an addition of features for a more accurate and robust algorithm, a more diverse dataset, an improved labelling process and the addition of a time axis to the RNN to improve the time before suction is detected.

Developing the greenest battery cell and establishing a European supply of batteries is the main goal for Northvolt. To achieve this, the recycling of batteries is a key enabler towards closing the loop and enabling the future of energy.When it comes to the recycling of electric vehicle battery packs, dismantling is one of of the main process steps.Given the size, weight and high voltage of the battery packs, automatic disassembly using robots is the preferred solution. The work presented in this thesis aims to develop and integrate a vision system able to identify and verify the battery pack dismantling process. To achieve this, two cameras were placed in the robot cell and the object detectors You Only Look Once (YOLO) and template matching were implemented, tested and compared. The results show that YOLO is the best object detector out of the ones implemented. The integration of the vision system with the robot controller was also tested and showed that with the results from the vision system, the robot controller can make informed decisions regarding the disassembly.

The following documents a research through design inquiry into how socialites of a hospital environment are disrupted or improved by implementing a service robot. The robot, support for excessive work, represents a new intermediary between a patient and a caregiver. Robotic work routines appear as better, more efficient, and more affordable. Apart from other ethical and inclusive considerations given to this dialogue, the social values hidden in traditional workflows are of equal importance.  This thesis attempts to generate constructive design research about emergent social norms and social dynamics caused by service robots’ implementation. The lessons learned are presented in a final research discussion. Further applied, the knowledge held common grounds with a rehabilitation robot developed by Blue Ocean Robotics.

Today, drones are used in both a civilian and military context, a swarm being a group of autonomous drones which are partly controlled by an operator. Swarms are a concept that is still in the research stage and previous research has consisted of how a swarm will communicate with the operator and how the swarm will work technically. However, the problem with the research available today is that it largely lacks the military connection. Hence, this study has aimed to investigate how the Armed Forces want to use swarms in the context of reconnaissance missions. The questions posed were, how do personnel within the Armed Forces want to plan swarm missions? What behaviors and capabilities do Armed Forces personnel want a swarm to possess? How do Armed Forces personnel want to receive information from swarms? Empirical data were collected by conducting three interviews with three individuals employed by the Armed Forces. The empirical data was then analyzed using qualitative content analysis where the responses were divided into smaller groups and sub-groups.  The study resulted in tables of desired abilities and behaviors that can be used in later research to create a real swarm system.

This compilation thesis presents an overarching framework on the utilization of nonlinear model predictive control(NMPC) for various applications in the context of Unmanned Aerial Vehicle (UAV) path planning and collision avoidance. Fast and novel optimization algorithms allow for NMPC formulations with high runtime requirement, as those posed by controlling UAVs, to also have sufficiently large prediction horizons as to in an efficient manner integrate collision avoidance in the form of set-exclusion constraints that constrain the available position-space of the robot. This allows for an elegant merging of set-point reference tracking with the collision avoidance problem, all integrated in the control layer of the UAV. The works included in this thesis presents the UAV modeling, cost functions, constraint definitions, as well as the utilized optimization framework. Additional contributions include the use case on multi-agent systems, how to classify and predict trajectories of moving (dynamic) obstacles, as well as obstacle prioritization when an aerial agent is in the precense of more obstacles, or other aerial agents, than can reasonably be defined in the NMPC formulation. For the cases of dynamic obstacles and for multi-agent distributed collision avoidance this thesis offers extensive experimental validation of the overall NMPC framework. These works push the limits of the State-of-the-Art regarding real-time real-life implementations of NMPC-based collision avoidance. The works also include a novel RRT-based exploration framework that combines path planning with exploration behavior. Here, a multi-path RRT * planner plans paths to multiple pseudo-random goals based on a sensor model and evaluates them based on the potential information gain, distance travelled, and the optimimal actuation along the paths.The actuation is solved for as as the solutions to a NMPC problem, implying that the nonlinear actuator-based and dynamically constrained UAV model is considered as part of the combined exploration plus path planning problem. To the authors best knowledge, this is the first time the optimal actuation has been considered in such a planning problem. For all of these applications, the utilized optimization framework is the Optimization Engine: a code-generation framework that generates a custom Rust-based solver from a specified model, cost function, and constraints. The Optimization Engine solves general nonlinear and nonconvex optimization problems, and in this thesis we offer extensive experimental validation of the utilized Proximal-Averaged Newton-type method for Optimal Control (PANOC) algorithm as well as both the integrated Penalty Method and Augmented Lagrangian Method for handling the nonlinear nonconvex constraints that result from collision avoidance problems.

Aim: The study aims to develop a DES model to evaluate the impacts of implementing packaging and palletizing robotics in the goods-receipts process. Research questions: RQ1: “What are the requirements when developing a Discrete Event Simulation (DES) model of the goods receipt process in the internal logistics section of a manufacturing company?”, RQ2: “What is the impact of implementing a new layout with industrial palletizer robotics in manufacturing companies' internal logistics - specifically the goods receipt process in terms of robot Overall Equipment Effectiveness (OEE), throughput, and WIP?” and, RQ3: “What considerations should companies take into account when implementing industrial palletizer robotics in their internal logistics, in terms of process flow?” Methodology: This study applied an inductive approach, letting the empirical findings guide the themes explored in the theoretical framework while having a constant interplay between empirical data collection and the construction of the simulation model. Empirical data collection was collected via observation, time study, and documentation. Literature was collected via Scopus and restricted to peer-reviewed articles and book chapters. The simulation model building followed the steps presented by Banks (2005). Based on the model building and simulation model results, conclusions could be made. Conclusion: The study concluded and presented nine requirements to develop a DES model for the goods-receipt process in manufacturing companies' internal logistics. The impact of robot OEE was determined to be low, 24-27%, WIP increased, and throughput increased with the implementation of a new layout with an industrial palletizer robot. Lastly, to answer RQ3, the study concluded and presented six considerations for companies to take into account when evaluating the implementation of an industrial palletizer robot / Syfte: Studiens syfte är att utveckla en DES modell för att utvärdera effekterna av att implementera omplocknings- och palleteringsrobotar för avdelningen –ankommande gods.  Frågeställningar: "Vilka krav finns vid utvecklingen av en modell för diskret händelsesimulering (DES) av godsmottagningsprocessen inom den interna logistiken hos ett tillverkningsföretag?", "Vilken är effekten av att implementera en ny layout med industriella palleteringsrobotar i tillverkningsföretags interna logistik - specifikt inom godsmottagningsprocessen - i termer av robot (OEE), produktionskapacitet per timme, och WIP?" och "Vilka överväganden bör företag ta i beaktande vid implementeringen av industriella palleteringsrobotar inom sin interna logistik, med avseende på processflöde?" Metod: Denna studie använde en induktiv metod, där de empiriska resultaten styrde de teman som utforskades inom den teoretiska referensramen, samtidigt som det fanns en konstant samverkan mellan insamling av empiriska data och konstruktionen av simulationsmodellen. Empiriska data samlades in genom observation, tidsstudier och dokumentation. Litteratur samlades in via Scopus och begränsades till expertgranskade artiklar och bokkapitel. Byggandet av simulationsmodellen följde de steg som presenterades av Banks (2005). Baserat på modellbyggandet och resultaten från simulationsmodellen kunde slutsatser dras. Slutsats: Studien presenterade nio krav för att utveckla en DES-modell för godsmottagningsprocessen inom tillverkningsföretags interna logistik. Effekten av robotens OEE bedömdes vara låg, 24–27%. Work in Progress (WIP) ökade och produktionskapaciteten per timme ökade med införandet av en ny layout med en industriell palleteringsrobot. Slutligen, för att svara på den tredje forskningsfrågan, presenterade studien sex överväganden som företag bör ta hänsyn till vid utvärderingen av implementeringen av en industriell palleteringsrobot

This thesis report presents a study on the virtualization of an Electric Cabin Scooter used to validate the feasibility of converting it into an autonomous vehicle. The project aimed to design, develop, and test a virtual model of the car that can navigate from points A to B while avoiding obstacles. The report describes the methodology used in the project, which includes setting up the workspace, construction of the virtual model, implementation of ROS2 controllers, and integration of SLAM and Navigation2. The thesis report also describes and discusses related work, as well as the theoretical background of the project. Results show a successfully developed working virtual vehicle model, which provides a solid starting point for future work. / Detta examensarbete presenterar en studie om virtualiseringen av en elektrisk kabinscooter. Den virtuella modellen används för att validera genomförbarheten av att omvandla den till ett autonomt fordon. Projektet syftade till att designa, utveckla och testa en virtuell modell av bilen som kan navigera från punkt A till B medan den undviker hinder. Rapporten beskriver metodiken som används i projektet, vilket inkluderar att sätta upp arbetsytan, konstruktion av den virtuella modellen, implementering av ROS2-kontroller och integration av SLAM och Navigation2. Rapporten diskuterar även relaterat arbete, samt teoretisk bakgrund till arbetet. Resultaten visar en framgångsrikt utvecklad fungerande virtuell fordonsmodell, som ger en solid utgångspunkt för framtida arbete.

En kostnadseffektiv, liten robotbil som är fjärrstyrd via Wi-Fi harutformats och testats för att möjliggöra inspektion och identifiering avpotentiella faror vid olycksplatser. Robotbilenär avsedd att användasför rekognosering och inspektion och kan bidra till att utforma enadekvat insatsplan. Hjulupphängning och hjul konstruerades medhjälp av en 3D-skrivare. Den färdiga robotbilen är utrustad medmikrokontrollern Raspberry Pi som har flera funktioner som gör denanvändbar i olika scenarier. Det finns en kamera som möjliggörfjärrinspektion av bilens omgivning. Utöver kameran finns det tresensorer som är kopplade till Raspberry Pi-enheten, nämligen engassensor för att upptäcka farliga gaser, en ultraljudssensor för attmäta avståndet till närmaste objekt och en temperatursensor för attmäta omgivningstemperaturen. Robotbilen använder enmotorstyrningsmodul för att styra bilens rörelse och två servomotorerför att möjliggöra att rotera kameran i vertikal och horisontell led.Robotbilens strömförsörjning kommer från ett batteri och tvåspänningsomvandlare används för att reglera spänningen tillmotorstyrningsmodulen och Raspberry Pi-enheten. Genom attintegrera dessa komponenter i en enda enhet och programmeraRaspberry Pi-enheten för att styra dem, kan robotbilen effektivt hjälpaatt undersöka och hantera potentiella faror. / A cost-effective small robot car that is remote-controlled via Wi-Fi hasbeen designed and tested for inspecting and identifying potentialhazards at accident sites. The robot car is intended for reconnaissanceand inspection purposes and can contribute to formulating anadequate action plan. Wheel suspension and wheels were constructedusing a 3D printer. The finished robot car is equipped with theRaspberry Pi microcontroller, which has severalfeatures that make ituseful in various scenarios. There is a camera that allows remoteinspection of the car's surroundings. In addition to the camera, thereare three sensors connected to the Raspberry Pi unit: a gas sensor todetect dangerous gases, an ultrasonic sensor to measure the distanceto the nearest object, and a temperature sensor to measure theambient temperature. The robot car uses a motor control module tocontrol its movement and two servo motors to enable the rotation ofthe camera in the vertical and horizontal directions. The robot car ispowered by a battery, and two voltage converters are used to regulatethe voltage to the motor control module and the Raspberry Pi unit. Byintegrating these components into a single unit and programming theRaspberry Pi unit to control them, the robot car can effectively assistin investigating and managing potential hazards.