No description available.

No description available.

Artificial neural networks (ANNs) have been used successfully in various practical problems. Though extensive improvements on different types of ANNs have been made to improve their performance, each ANN design still experiences its own limitations. The existing digital human models are mature enough to provide accurate and useful results for different tasks and scenarios under various conditions. There is, however, a critical need for these models to run in real time, especially those with large-scale problems like motion prediction which can be computationally demanding. For even small changes to the task conditions, the motion simulation needs to run for a relatively long time (minutes to tens of minutes). Thus, there can be a limited number of training cases due to the computational time and cost associated with collecting training data. In addition, the motion problem is relatively large with respect to the number of outputs, where there are hundreds of outputs (between 500-700 outputs) to predict for a single problem. Therefore, the aforementioned necessities in motion problems lead to the use of tools like the ANN in this work. This work introduces new algorithms for the design of the radial-basis network (RBN) for problems with minimal available training data. The new RBN design incorporates new training stages with approaches to facilitate proper setting of necessary network parameters. The use of training algorithms with minimal heuristics allows the new RBN design to produce results with quality that none of the competing methods have achieved. The new RBN design, called Opt_RBN, is tested on experimental and practical problems, and the results outperform those produced from standard regression and ANN models. In general, the Opt_RBN shows stable and robust performance for a given set of training cases. When the Opt_RBN is applied on the large-scale motion prediction application, the network experiences a CPU memory issue when performing the optimization step in the training process. Therefore, new algorithms are introduced to modify some steps of the new Opt_RBN training process to address the memory issue. The modified steps should only be used for large-scale applications similar to the motion problem. The new RBN design proposes an ANN that is capable of improved learning without needing more training data. Although the new design is driven by its use with motion prediction problems, the consequent ANN design can be used with a broad range of large-scale problems in various engineering and industrial fields that experience delay issues when running computational tools that require a massive number of procedures and a great deal of CPU memory. The results of evaluating the modified Opt_RBN design on two motion problems are promising, with relatively small errors obtained when predicting approximately 500-700 outputs. In addition, new methods for constraint implementation within the new RBN design are introduced. Moreover, the new RBN design and its associated parameters are used as a tool for simulated task analysis. This work initiates the idea that output weights (W) can be used to determine the most critical basis functions that cause the greatest reduction in the network test error. Then, the critical basis functions can specify the most significant training cases that are responsible for the proper performance achieved by the network. The inputs with the most change in value can be extracted from the basis function centers (U) in order to determine the dominant inputs. The outputs with the most change in value and their corresponding key body degrees-of-freedom for a motion task can also be specified using the training cases that are used to create the network's basis functions.

Simulating realistic human behavior on a virtual avatar presents a difficult task. Because the simulated environment does not adhere to the same scientific principles that we do in the existent world, the avatar becomes capable of achieving infeasible postures. In an attempt to obtain realistic human simulation, real world constraints are imposed onto the non-sentient being. One such constraint, and the topic of this thesis, is self-collision avoidance. For the purposes of this topic, a posture will be defined solely as a collection of angles formed by each joint on the avatar. The goal of self-collision avoidance is to eliminate the formation of any posture where multiple body parts are attempting to occupy the exact same space. My work necessitates an extension of this definition to also include collision avoidance with objects attached to the body, such as a backpack or armor. In order to prevent these collisions from occurring, I have implemented an effort-based approach for correcting afflicted postures. This technique specifically pertains to postures that are sequenced together with the objective of animating the avatar. As such, the animation's coherence and defining characteristics must be preserved. My approach to this problem is unique in that it strategically blends the concept of keyframe interpolation with an optimization-based strategy for posture prediction. Although there has been considerable work done with methods for keyframe interpolation, there has been minimal progress towards integrating a realistic collision response strategy. Additionally, I will test this optimization-based approach through the use of a complex kinematic human model and investigate the use of the results as input to an existing dynamic motion prediction system.

Cette thèse explore l'utilisation des modèles bayésiens dans les IA de jeux vidéo multi-joueurs, particulièrement l'IA des jeux de stratégie en temps réel (STR). Les jeux vidéo se situent entre la robotique et la simulation totale, car les autres joueurs ne sont pas simulés, et l'IA n'a pas de contrôle sur la simulation. Les jeux de STR demandent simultanément d'effectuer des actions reactives (contrôle d'unités) et de prendre des décisions stratégiques (technologiques, économiques) et tactiques (spatiales, temporelles). Nous avons utilisé la modélisation bayésienne comme une alternative à la logique (booléenne), étant capable de travailler avec des informations incomplètes, et donc incertaines. En effet, la spécification incomplète des comportement "scriptés", ou la spécification incomplète des états possibles dans la recherche de plans, demandent une solution qui peut gérer cette incertitude. L'apprentissage artificiel aide à réduire la complexité de spécifier de tels modèles. Nous montrons que la programmation bayésienne peut intégrer toutes sortes de sources d'incertitudes (états cachés, intentions, stochasticité) par la réalisation d'un joueur de StarCraft complètement robotique. Les distributions de probabilité sont un moyen de transporter, sans perte, l'information que l'on a et qui peut représenter au choix: des contraintes, une connaissance partielle, une estimation de l'espace des états et l'incomplétude du modèle lui-même. Dans la première partie de cette thèse, nous détaillons les solutions actuelles aux problèmes qui se posent lors de la réalisation d'une IA de jeu multi-joueur, en donnant un aperçu des caractéristiques calculatoires et cognitives complexes des principaux types de jeux. En partant de ce constat, nous résumons les catégories transversales de problèmes, et nous introduisons comment elles peuvent être résolues par la modélisation bayésienne. Nous expliquons alors comment construire un programme bayésien en partant de connaissances et d'observations du domaine à travers un exemple simple de jeu de rôle. Dans la deuxième partie de la thèse, nous détaillons l'application de cette approche à l'IA de STR, ainsi que les modèles auxquels nous sommes parvenus. Pour le comportement réactif (micro-management), nous présentons un controleur multi-agent décentralisé et temps réel inspiré de la fusion sensori-motrice. Ensuite, nous accomplissons les adaptation dynamiques de nos stratégies et tactiques à celles de l'adversaire en le modélisant à l'aide de l'apprentissage artificiel (supervisé et non supervisé) depuis des traces de joueurs de haut niveau. Ces modèles probabilistes de joueurs peuvent être utilisés à la fois pour la prédiction des décisions/actions de l'adversaire, mais aussi à nous-même pour la prise de décision si on substitue les entrées par les notres. Enfin, nous expliquons l'architecture de notre joueur robotique de StarCraft, et nous précisions quelques détails techniques d'implémentation. Au delà des modèles et de leurs implémentations, il y a trois contributions principales: la reconnaissance de plan et la modélisation de l'adversaire par apprentissage artificiel, en tirant partie de la structure du jeu, la prise de décision multi-échelles en présence d'informations incertaines, et l'intégration des modèles bayésiens au contrôle temps réel d'un joueur artificiel. / This thesis explores the use of Bayesian models in multi-player video games AI, particularly real-time strategy (RTS) games AI. Video games are an in-between of real world robotics and total simulations, as other players are not simulated, nor do we have control over the simulation. RTS games require having strategic (technological, economical), tactical (spatial, temporal) and reactive (units control) actions and decisions on the go. We used Bayesian modeling as an alternative to (boolean valued) logic, able to cope with incompleteness of information and (thus) uncertainty. Indeed, incomplete specification of the possible behaviors in scripting, or incomplete specification of the possible states in planning/search raise the need to deal with uncertainty. Machine learning helps reducing the complexity of fully specifying such models. We show that Bayesian programming can integrate all kinds of sources of uncertainty (hidden state, intention, stochasticity), through the realization of a fully robotic StarCraft player. Probability distributions are a mean to convey the full extent of the information we have and can represent by turns: constraints, partial knowledge, state space estimation and incompleteness in the model itself. In the first part of this thesis, we review the current solutions to problems raised by multi-player game AI, by outlining the types of computational and cognitive complexities in the main gameplay types. From here, we sum up the transversal categories of prob- lems, introducing how Bayesian modeling can deal with all of them. We then explain how to build a Bayesian program from domain knowledge and observations through a toy role-playing game example. In the second part of the thesis, we detail our application of this approach to RTS AI, and the models that we built up. For reactive behavior (micro-management), we present a real-time multi-agent decentralized controller inspired from sensory motor fusion. We then show how to perform strategic and tactical adaptation to a dynamic opponent through opponent modeling and machine learning (both supervised and unsupervised) from highly skilled players' traces. These probabilistic player-based models can be applied both to the opponent for prediction, or to ourselves for decision-making, through different inputs. Finally, we explain our StarCraft robotic player architecture and precise some technical implementation details. Beyond models and their implementations, our contributions are threefolds: machine learning based plan recognition/opponent modeling by using the structure of the domain knowledge, multi-scale decision-making under uncertainty, and integration of Bayesian models with a real-time control program.

Der Einsatz von digitalen Menschmodellen erlaubt neben einer frühzeitigen ergonomischen Analyse die Gestaltung von Arbeitsprozessen und stellt ein hilfreiches Werkzeug in der Produkt- und Prozessgestaltung dar. Im Rahmen dieses Beitrages soll auf ausgewählte Schwerpunkte der Entwicklung des digitalen Menschmodells „The Smart Virtual Worker“ eingegangen werden. Das Forschungsprojekt soll einen Beitrag zur Lösung, der mit dem demografischen Wandel der Gesellschaft einhergehenden Herausforderungen leisten. Die daraus resultierenden Forschungsschwerpunkte liegen insbesondere in der Einbeziehung von Alterungs- und psychischen Faktoren in die Bewegungsgenerierung des Menschmodells und der Modellierung von Umweltbedingungen. In Umsetzung des Projektes wurde ein erstes Arbeitsszenario erarbeitet, auf dessen Basis die vorgenannten Forschungsaufgaben interdisziplinär gelöst werden sollen.

The present work deals with the development of a computational geometric framework for vision modeling for performing visibility and legibility analyses in Digital Human Modeling (DHM) using the field-of-view (FoV), estimated geometrically from 3D tessellated head scans. DHM is an inter-disciplinary area of research with the prime objective of evaluating a product, job or environment for intended users through computer-based simulations. Vision modeling in the existing DHM’s has been primarily addressed through FoV modeling using right circular cones (RCC). Perimetry literature establishes that the human FoV is asymmetric due to unrestricted zygomatic vision and restrictions on the nasal side of the face. This observation is neither captured by the simplistic RCC models in DHM, nor rigorously studied in vision literature. Thus, the RCC models for FoV are inadequate for rigorous simulations and the accurate modeling of FoV is required in DHM. The computational framework developed in this work considers three broad components namely, the geometric estimation and representation of FoV, visibility and statistical visibility, and legibility of objects in a given environment. A computational geometric method for estimating FoV from 3D laser-scanned models of the human head is presented in this work. The strong one-to-one similarity between computed and clinically perimetry maps establishes that the FoV can be geometrically computed using tessellated head models, without necessarily going through the conventional interaction based clinical procedures. The algorithm for FoV computation is extended to model the effect of gaze-direction on the FoV resulting in binocular FoV. A novel unit-cube scheme is presented for robust, efficient and accurate modeling of FoV. This scheme is subsequently used to determine the visibility of 3D tessellated objects for a given FoV. In order to carry out population based visibility studies, the statistical modeling FoV and generation of percentile-based FoV curves are introduced for a given population of FoV curves. The percentile data thus generated was not available in the current ergonomics or perimetry literature. Advanced vision analysis involving character-legibility is demonstrated using the unit-cube with an improved measure to incorporate the effect of character-thickness on its legibility.

abstract: Daily collaborative tasks like pushing a table or a couch require haptic communication between the people doing the task. To design collaborative motion planning algorithms for such applications, it is important to understand human behavior. Collaborative tasks involve continuous adaptations and intent recognition between the people involved in the task. This thesis explores the coordination between the human-partners through a virtual setup involving continuous visual feedback. The interaction and coordination are modeled as a two-step process: 1) Collecting data for a collaborative couch-pushing task, where both the people doing the task have complete information about the goal but are unaware of each other's cost functions or intentions and 2) processing the emergent behavior from complete information and fitting a model for this behavior to validate a mathematical model of agent-behavior in multi-agent collaborative tasks. The baseline model is updated using different approaches to resemble the trajectories generated by these models to human trajectories. All these models are compared to each other. The action profiles of both the agents and the position and velocity of the manipulated object during a goal-oriented task is recorded and used as expert-demonstrations to fit models resembling human behaviors. Analysis through hypothesis teasing is also performed to identify the difference in behaviors when there are complete information and information asymmetry among agents regarding the goal position. / Dissertation/Thesis / Masters Thesis Electrical Engineering 2020

Bad worker health leads to an unnecessary increase of absent days and loss of money, particularly the more physically demanding jobs as in industry. This can be seen in both a broader perspective and also in the suffering of the individual. In order to stop this trend and lower the work injuries connected to bad ergonomics good, reliable, and preferably digital, methods have to be generated and evaluated. The aim of this master’s thesis is to investigate how Digital Human Modelling (DHM) tools can beused to evaluate physical ergonomics by building a real-life workstation in the DHM tool IPS IMMA. The workstation used here is based on a newly installed station at a large company placed in Skövde. This station was developed in collaboration of both technical specialists but also ergonomist consulting from the company Feelgood. The goal has therefore been to examine where in the process a DHM tool could be used and if it would contribute to the process. The methods chosen to investigate the use of DHM tools were to build a final model in IPS IMMA, which is based of four different versions of the workstation. By building four different versions of the workstation the process has simultaneously been analyzed and documented in order to compare the findings made in the program to the ones made in real life. The results have also been made in combination to finding the opportunities, challenges, and disadvantages with using DHM tools. The needed improvements within the DHM field have also been noted and discussed.

The objective of this project is to develop a driving test using a Digital Human Modeling tool (DHM), specifically IPS IMMA, which will allow the evaluation of the ergonomics of the interior of vehicles as currently demanded by the automotive companies. Thus, improving both the design and the design process. This will involve a study of the driving and the tasks carried out by a real person to end up programming them in the DHM software. Based on this study an interface is suggested that guides engineers or ergonomists to design their own driving tests and enable them to evaluate their own designs without a high specialization in DHM tools and software. Taking into account the already present autonomous cars and their future development, the conceptual design of a two positions steering wheel (autonomous/manual driving) will be introduced as an example to be added in the driving test. This example is intended to show how DHM tools can be used to evaluate different designs solutions in early stages of the product development process. This project will be a contribution to one of the sections of the ADOPTIVE project carried out at the University of Skövde and in collaboration with Swedish automotive companies.