A desire to advance humanity's presence in space prompts the need for improved technology to send crew to places such as the Moon, Mars, and nearby asteroids. Safely placing a crewed vehicle on and in any landing condition requires a design decision regarding the distribution of responsibilities between the crew and automation. In this thesis, a cognitive process model is used to determine the necessary automated functionality to support astronaut decision making. Current literature lacks sufficient detailed knowledge regarding astronaut decision making during this task and observations of astronauts landing on the Moon are not readily available. Therefore, a series of human-in-the-loop experiments, one of which was conducted with the NASA Astronaut Office at Johnson Space Center, have been conducted to examine the changes in performance due to differing function allocations, trajectory profiles, and scenario operations. The data collected in the human-in-the-loop study has provided empirical data that has informed the cognitive process model, the requirements analysis, and provided insight regarding cockpit display usage and information needs. The proposed system requirements include design guidance for assisting astronauts during both nominal and off-nominal landing scenarios.

The manning of a ship is a major driver of life cycle cost. The U.S. Government Accounting Office (GAO) has determined that manpower is the single most influential component in the life cycle cost of a ship. Life cycle cost is largely determined by decisions made during concept design. Consequently, reliable manpower estimates need to be included early in the design process, preferably in concept design. The ship concept exploration process developed at Virginia Tech uses a Multi-Objective Genetic Optimization to search the design space for feasible and non-dominated ship concepts based on cost, risk and effectiveness. This requires assessment of thousands of designs without human intervention. The total ship design problem must be set up before actually running the optimization. If manning is to be included in this process, manning estimate tools must be run seamlessly as part of the overall ship synthesis and optimization. This thesis provides a method of implementing a manning task network analysis tool (ISMAT, Integrated Simulation Manning Analysis Tool, Micro Analysis and Design) in an overall ship synthesis program and design optimization. The inputs to the analysis are ship systems (propulsion, combat systems, communication, etc), maintenance strategy, and level of automation. The output of the manning model is the number of crew required to accomplish a given mission for a particular selection of systems, maintenance and automation. Task network analysis programs are ideal for this problem. They can manage the probabilistic nature of a military mission and equipment maintenance, and can be used to simplify the problem by breaking down the complex functions and tasks of a ship's crew. The program builds large and complex functions from small related tasks. This simplifies the calculation of personnel and time utilization, and allows a more flexible scheme for building complex mission scenarios. In this thesis, ISMAT is run in a pre-optimization step to build a response surface model (RSM) for calculating required manning as a function of systems, maintenance and automation. The RSM is added to the ship synthesis model to calculate required manning, and a concept exploration case study is performed for an Air Superiority Cruiser (CGX) using this model. The performance of the manning model in this case study is assessed and recommendations are made for future work. This research shows that there is a difference between minimum manning and optimal manning on US Navy Ships. / Master of Science

The purpose of this study is to extend previous findings of Mouloua, Parasuraman, and Molloy (1993), Parasuraman, Mouloua, and Molloy (1996), Hilburn, Parasuraman, and Mouloua (1996), and Oakley, Mouloua, and Hancock (2003) by: 1) examining the effect of repeated adaptive function allocation to manual control of minimal length (5 minutes) to reduce of human error and minimize workload; 2) explore the placement or timing of adaptive function allocation intervals (approximately 20 minutes of automation control to reduce the human operators' monitoring decrement between intervals, maintain adaptive recovery performance levels, and improve response times); 3) examine different levels of automation reliability (30%, 60%, and 90% reliable); 4) explore factors that may be manipulated to reduce automation-induced monitoring inefficiency, increase detection of automation malfunctions, improve situation awareness, reduce response/reaction times, and reduce workload in a simulated complex aviation system. The study was a 2 (non-adaptive control vs. adaptive group) x 3 (30%, 60%, and 90% automation reliability condition) x 4 (repeated 25 minute session) mixed factorial design. Fifty-four undergraduate participants' (i.e., 27 participants per group; 9 participants per condition; at least 18 yrs. of age) percentage of detected malfunctions, response times, and subjective workload were gathered from the Multi-Attribute Task Battery and the NASA TLX. Results indicated a significant improvement in detection of malfunctions and response times during adaptive-function allocation to manual control but without adaptive recovery. There was a significant effect for workload found between baseline measures and experimental sessions by group in the first session but not across experimental sessions. Theoretical and practical implications, limitations and future research are discussed.

Automated decision making is becoming the norm across large parts of society, which raises interesting liability challenges when human control over technical systems becomes increasingly limited. This article defines "quasi-automation" as inclusion of humans as a basic rubber-stamping mechanism in an otherwise completely automated decision-making system. Three cases of quasi- automation are examined, where human agency in decision making is currently debatable: self- driving cars, border searches based on passenger name records, and content moderation on social media. While there are specific regulatory mechanisms for purely automated decision making, these regulatory mechanisms do not apply if human beings are (rubber-stamping) automated decisions. More broadly, most regulatory mechanisms follow a pattern of binary liability in attempting to regulate human or machine agency, rather than looking to regulate both. This results in regulatory gray areas where the regulatory mechanisms do not apply, harming human rights by preventing meaningful liability for socio-technical decision making. The article concludes by proposing criteria to ensure meaningful agency when humans are included in automated decision-making systems, and relates this to the ongoing debate on enabling human rights in Internet infrastructure.

Cette thèse s’inscrit dans le cadre théorique très général de la conception des systèmessociotechniques complexes. Elle est appliquée au système de l'Appareil à Gouverner des sous-­‐marins. Elle traite, tout particulièrement, de l'étape d’allocation des fonctions entre l’homme et la machine.L’objectif des premiers travaux a été de tester un système automatisant la phase d’analyse de l’information mise à disposition du barreur. Deux expérimentations ont évalué les effets d’une interface "écologique" sur la performance, la charge de travail et la conscience de la situation de barreurs réalisant une tâche de pilotage sur simulateur.Une deuxième phase a consisté à définir différentes solutions d’allocation dynamique des fonctions entre l’homme et la machine dans le système de pilotage de sous-­‐marin. Une expérimentation a évalué l’apport d’une réallocation à la machine de la gestion du cap et/ ou de l'immersion, en fonction de la performance et de la charge de travail de l'opérateur.Les résultats de ces différentes études conduisent à proposer une démarche de conception d'un système sociotechnique qui reprend et enrichit le cadre du CWA (Cognitive Work Analysis) et à discuter du positionnement et de la formalisation del’allocation des fonctions dans le processus de conception des systèmes homme-­‐machine. / This thesis comes within the scope of sociotechnical systems design. It is applied to thesubmarine rudder control system. It focuses, in particular, on the stage of functions allocation between helmsman and machine.The objective of the first experiments was to test an "ecological" interface automating the function of information analysis. Two experiments evaluated its effects on performance, workload and situation awareness of helmsmen performing steering tasks.A second phase consisted in defining different proposals of dynamic allocations of functions between helmsman and machine. An experiment was carried out, in order to evaluate a solution of adaptive automation.The results of these studies lead to propose an approach for designing a socio-­‐technical system, which relies on and enhances the CWA (Cognitive Work Analysis) framework. They lead also to discuss place and formalisation of the functions allocation's stage into the process of sociotechnical systems' design.

Business process modeling is utilized by organizations for defining and reengineering their business processes. On the other hand, software requirements analysis activities are performed for determining the system boundaries, specifying software requirements using system requirements and resolving conflicts between requirements. From this point of view, these two activities are considered in different disciplines. An organization requiring its business processes to be defined and supported with information systems would benefit from performing business process modeling and requirements analysis concurrently. In this study, an approach enabling concurrent execution of business process modeling and requirements analysis is developed. The approach includes two business process modeling notations adapted to the research needs, a process defining the steps for implementing the approach and the requirements generation tool that generates natural language specification documents by using business process models. Within this study, two case studies are introduced / one describing the development of the approach and the other exploring if the total efficiency of performing business process modeling and requirements analysis activites would be increased by using the approach.

L’évolution de la technologie se traduit par l’émergence de systèmes de plus en plus complexes et automatisés. Dans les situations dynamiques, l’activité de l’opérateur est désormais plus centrée sur la supervision que sur l’exécution. Ces situations sont caractérisées par la complexité, l’incertitude, un contrôle partiel et un risque important de perdre la maîtrise de la situation. L’objectif principal des travaux présentés dans ce manuscrit est de créer un indicateur synthétique de charge mentale, qui servira à piloter une réallocation des fonctions en temps réel. Il s'agit, in fine, de permettre à l’opérateur de mieux réguler sa charge mentale, et de l'aider à garder ou à récupérer la maîtrise de la situation. Ces travaux sont appliqués à l’activité de supervision d’un essaim de drones étudiée sur une plateforme de simulation (la plateforme SUSIE). Nous avons pour cela revisité et combiné le modèle de Hart et Staveland (1988), qui considère la charge mentale comme une construction multidimensionnelle et celui de Sperandio (1971), qui se focalise sur les régulations de l’activité. Cela nous a amené à proposer un modèle dynamique et tridimensionnel (basé sur l’analyse des contraintes, des effets de la charge sur la tâche et l’opérateur, et des régulations), sur lequel nous nous sommes appuyés pour construire une démarche de création d’un indicateur synthétique de charge mentale. Cette démarche se décompose en quatre étapes. La première consiste à réaliser une analyse de la tâche en suivant la méthodologie de Sperandio (1988), afin de définir et d’identifier les descripteurs de la charge mentale suivant les trois catégories de notre modèle dynamique et tridimensionnel. La seconde étape s’appuie sur la réalisation d’une expérimentation, qui vise à valider ou exclure les indicateurs identifiés, en fonction de leur sensibilité aux variations de charge mentale (telle que définie par Cegarra et Chevalier (2008)). Les résultats montrent notamment que le diamètre pupillaire et l’observation des régulations permettent une mesure sensible. La troisième étape a pour objectif de créer un indicateur synthétique de charge mentale en fusionnant par réseaux bayésiens et logique floue les indicateurs retenus. Enfin, la dernière phase supporte la conception d’une allocation dynamique des fonctions, déclenchée à partir de l’indicateur synthétique de charge mentale. Une deuxième expérimentation a été menée afin d’évaluer l’apport de la réallocation. Elle montre que l’aide permet à l’opérateur de mieux réguler son activité ; elle modifie la charge de travail (contrainte), mais ne réduit pas la charge mentale. / The evolution of technology has resulted in the emergence of increasingly complex and automated systems. In dynamic situations, the activity of the operator is now more centred on supervision than on execution. These situations are characterized by complexity, uncertainty, partial control and a significant risk of losing control of the situation. The main objective of the work presented in this manuscript is to create a synthetic mental workload indicator, which will be used to trigger a reallocation of functions in real time. The ultimate goal is to enable the operator to better regulate his/her mental workload and to help him/her maintain or regain control of the situation. This work is applied to the supervision activity of a drone swarm studied on a simulation platform (the SUSIE platform). We have revisited and combined the model of Hart and Staveland (1988), which considers the mental workload as a multidimensional construct and that of Sperandio (1971), which focuses on the regulation of activity. This led us to propose a dynamic and three-dimensional model (based on the analysis of the constraints, the effects of the workload on the task and the operator, and the regulations), on which we relied to build a creative approach of a synthetic mental workload indicator. This process is divided into four stages. The first consists in carrying out an analysis of the task following the methodology of Sperandio (1988), to define and identify the descriptors of the mental workload according to the three categories of our dynamic and three-dimensional model. The second step is based on the realization of an experiment, which aims to validate or exclude the identified indicators, per their sensitivity to variations in mental workload (as defined by Cegarra et Chevalier (2008)). The results show in particular that the pupillary diameter and the observation of the regulations allow an appreciable measurement. The third step aims to create a synthetic indicator of mental load by merging the chosen indicators by using Bayesian networks and fuzzy logic. Finally, the last phase supports the design of a dynamic allocation of functions, triggered from the synthetic indicator of mental workload. A second experiment was carried out to evaluate the contribution of reallocation. It shows that the aid allows the operator to better regulate his activity; it modifies the task workload (constraint), but does not reduce the mental workload.

The desire to send humans to Mars will require a change in the way that extravehicular activity (EVA) is performed; in-space crews (including those within a vehicle or habitat monitoring others conducting EVA) will need to be more autonomous and that will require them to monitor large amounts of information in order to ensure crew safety and mission success. The amount of information to perceive and process will overwhelm unassisted intra-vehicular (IV) crewmembers, meaning that automation will need to be developed to support these crews on Mars while EVA is performed (Mishkin, Lee, Korth, & LeBlanc, 2007). This dissertation seeks to identify the information requirements for the performance of scientific EVA and determine which information streams will need to be allocated to in-space crew and which are the most effective streams to automate. The first study uses Mars rover operations as a homology—as defined by von Bertalanffy (1968)—to human scientific exploration. Mars rover operations personnel were interviewed using a novel method to identify the information requirements to perform successful science on Mars, how that information is used, and the timescales on which those information streams operate. The identified information streams were then related to potential information streams relevant to human exploration in order to identify potential function allocation or automated system development areas. The second study focused on one identified mission-critical information stream for human space exploration: monitoring astronaut status physiologically. Heart rate, respiration rate, and heart rate variability measurements were recorded from participants as they performed field science tasks (potentially tasks that are similar to those that will be performed by astronauts on Mars). A statistical method was developed to analyze this data in order to determine whether or not physiological responses to different tasks were statistically different, and whether any of those differences followed consistent patterns. A potential method to automate the monitoring of physiological data was also described. The results of this work provide a more detailed outline of the information requirements for EVA on Mars and can be used as a starting point for others in the exploration community to further develop automation or function allocation to support astronauts as they explore Mars.

Mit der Automatisierung geht stets erneut die Frage nach der Funktionsallokation einher. Diese lässt sich vereinfacht wie folgt formulieren: Welche Aufgaben bzw. Funktionen soll die Automation und welche der Mensch ausführen? Abwendend von traditionellen Ansätzen maximaler Automatisierung und statischer Aufgabenteilungen haben sich heute dynamische Ansätze der Funktionsallokation bzw. kooperativen Automation durchgesetzt, bei denen sich Operateur und Automation die Kontrolle für bestimmte Aufgaben während des Prozesses teilen oder übergeben können. Die Ausgestaltung dieser kooperativen Automation wirft insbesondere zwei Fragen auf: Wer hat (wann wieviel) Kontrolle? Und wer entscheidet darüber? Diese Fragen nach Kontrolle und Autorität bilden die Grundlage für das theoretische und empirische Anliegen dieser Arbeit. Auf Basis einer umfangreichen literaturbasierten Diskussion der Fragen wird ein eigenes Rahmenmodell vorgeschlagen. Dieses erlaubt, anhand der Aspekte Authorität und Kontrolle, konkrete und trennscharfe Kooperationsformen zwischen Operateur und Automation zu beschreiben. Der einfachste Fall einer Zusammenarbeit zwischen Operateur und Automation wird sodann in einer Studie mit 81 Operateuren untersucht. Es interessiert, ob sich die Leistung von Operateuren, welche bei der Prozessüberwachung und -führung optional mit einer automatischen Unterstützung zusammen arbeiten können, im Vergleich zu Operateuren ohne eine solche automatische Unterstützung verbessert. Für bestimmte Leistungsindikatoren können in der Tat höchst positive Effekte der automatischen Unterstützung ermittelt und konkrete Interaktionsmuster der Operateure in der Kooperation mit der automatischen Unterstützung beschrieben werden. Die theoretischen und empirischen Ergebnisse der diesere Arbeit liefern eine Grundlage für differenzierte Bewertungen und somit auch Empfehlungen spezifischer Kooperationsformen, die gerade im Hinblick auf die Ausgestaltung kooperativer Automation noch rar sind. / Automation has always been accompanied by the question of function allocation, that is: which tasks and/or functions should be executed by the machine vs. by the human? Today, dynamic function allocation or cooperative automation has become the dominant approach, which allows for a more flexible sharing and trading of control between operator and automation. However, how cooperative automation is put into practice remains mainly an open question. Two central aspects have to be dealt with: Who should execute a certain function, i.e. have control (and when)? And who should decide about it? These questions of control and authority are central to this thesis. Based on an extensive literature review, a theoretical model is proposed. This model allows for a classification of specific forms of control between operator and automation based on authoriy and control. In an empirical study the effect of the most simple form of cooperation is investigated regarding its effect on performance of 81 operators. It is assumed that supervisory control performance of operators benefits from this form of cooperation with automation. Thereby supervisory control performance of operators cooperating with automation is compared to performance of operators who were not provided with such automation, i.e. had to supervise and control fully manually. Results show, that cooperation with automation has a positive effect on certain performance indicators. Also interaction patterns of operators cooperating with automation are investigated in detail. Both, the theoretical and empirical approach in this thesis provide a profound basis which allows for a more sophisticated evaluation and recommendation of forms of cooperation between operator and automation. So far, such recommendations are rare in Human Factors Research but they are desperately needed, especially when it comes to the implementation of forms of cooperative automation.

Function allocation is the design decision which assigns work functions to all agents in a team, both human and automated. Efforts to guide function allocation systematically has been studied in many fields such as engineering, human factors, team and organization design, management science, and cognitive systems engineering. Each field focuses on certain aspects of function allocation, but not all; thus, an independent discussion of each does not address all necessary issues with function allocation. Four distinctive perspectives emerged from a review of these fields: technology-centered, human-centered, team-oriented, and work-oriented. Each perspective focuses on different aspects of function allocation: capabilities and characteristics of agents (automation or human), team structure and processes, and work structure and the work environment. Together, these perspectives identify the following eight issues with function allocation: 1)Workload, 2)Incoherency in function allocations, 3)Mismatches between responsibility and authority, 4)Interruptive automation, 5)Automation boundary conditions, 6)Function allocation preventing human adaptation to context, 7)Function allocation destabilizing the humans' work environment, and 8)Mission Performance. Addressing these issues systematically requires formal models and simulations that include all necessary aspects of human-automation function allocation: the work environment, the dynamics inherent to the work, agents, and relationships among them. Also, addressing these issues requires not only a (static) model, but also a (dynamic) simulation that captures temporal aspects of work such as the timing of actions and their impact on the agent's work. Therefore, with properly modeled work as described by the work environment, the dynamics inherent to the work, agents, and relationships among them, a modeling framework developed by this thesis, which includes static work models and dynamic simulation, can capture the issues with function allocation. Then, based on the eight issues, eight types of metrics are established. The purpose of these metrics is to assess the extent to which each issue exists with a given function allocation. Specifically, the eight types of metrics assess workload, coherency of a function allocation, mismatches between responsibility and authority, interruptive automation, automation boundary conditions, human adaptation to context, stability of the human's work environment, and mission performance. Finally, to validate the modeling framework and the metrics, a case study was conducted modeling four different function allocations between a pilot and flight deck automation during the arrival and approach phases of flight. A range of pilot cognitive control modes and maximum human taskload limits were also included in the model. The metrics were assessed for these four function allocations and analyzed to validate capability of the metrics to identify important issues in given function allocations. In addition, the design insights provided by the metrics are highlighted This thesis concludes with a discussion of mechanisms for further validating the modeling framework and function allocation metrics developed here, and highlights where these developments can be applied in research and in the design of function allocations in complex work environments such as aviation operations.