This thesis is concerned with modelling the properties of human ventilation during steady-state conditions and during acute and sustained isocapnic hypoxia. <strong>Chapter 1</strong> reviews some of the relevant studies in animals and humans. <strong>Chapter 2</strong> describes the origins of the data studied in this thesis. In particular, it describes the experimental apparatus and the technique of dynamic end-tidal forcing used to gather the data, as well as the particular protocols employed. <strong>Chapter 3</strong> studies the breath-to-breath variations in ventilation during steady breathing in both rest and during light exercise with the end-tidal gases controlled. The results suggest that: 1) both simple ARMA models and a simple state-space model can describe the autocorrelation present in the data; 2) variations in spectral power were present in the data which cannot be described by these models; and 3) these variations were often due to a uniform modulation and did not significantly affect the coefficients of the models. For these kinds of data, a heteroscedastic form of state-space model provides an attractive theoretical structure for the noise processes. <strong>Chapter 4</strong> studies human ventilation during sustained isocapnic hypoxia. Two models are used. The first, developed by Painter et al. (J. Appl. Physiol. 74:2007-2015, 1993) describes hypoxic ventilatory decline (HVD) as a decline in peripheral chemoreflex sensitivity. The second is an extended model which incorporates a component of HVD that is independent of peripheral chemoreflex sensitivity. The models incorporate a parallel noise structure. It is concluded that, in some subjects but not others, there is a component of HVD which is independent of peripheral chemoreflex sensitivity. <strong>Chapter 5</strong> studies the human ventilatory response to cyclic isocapnic hypoxia. Both a simple proportional dynamic model suggested by Clement and Robbins (Respir. Physiol. 92:253-175, 1993), and an extended model with an additional non-linear rate-sensitive component are studied. The models incorporate a parallel noise structure. The results show that, although the extended model improves the fit to the data for some subjects, both models failed to explain the data fully, especially the occasional large breaths, which were shown to occur more frequently in some parts of the hypoxic cycle than other parts.

No description available.

Humans learn to interact with many complex physical systems. For example, humans learn to fly aircraft, operate drones, and drive automobiles. We present results from human-in-the-loop (HITL) experiments, where human subjects interact with dynamic systems while performing command-following tasks multiple times over a one-week period. We use a new subsystem identification (SSID) algorithm to estimate the control strategies (feedforward, feedforward delay, feedback, and feedback delay) that human subjects use during their trials. We use experimental and SSID results to examine the effects of system characteristics (e.g., system zeros, relative degree, system order, phase lag, time delay), reference command, and command-following objectives on humans command-following performance and on the control strategies that the humans learn. Results suggest that nonminimum-phase zeros, relative degree, phase lag, and time delay tend to make dynamic systems difficult for human to control. Subjects can generalize their control strategies from one task to another and use prediction of the reference command to improve their command-following performance. However, this dissertation also provides evidence that humans can learn to improve performance without prediction. This dissertation also presents a new SSID algorithm to model the control strategies that human subjects use in HITL experiments where they interact with dynamic systems. This SSID algorithm uses a two-candidate-pool multi-convex-optimization approach to identify feedback-and-feedforward subsystems with time delay that are interconnected in closed loop with a known subsystem. This SSID method is used to analyze the human control behavior in the HITL experiments discussed above.

The purpose of this thesis is to explore the theoretical and practical issues associated with the design, execution and analysis of an empirical evaluation of novel process control displays in a collaborative control room environment. An important feature of the thesis is how practical constraints associated with limited access to industry controllers were handled and how convergent lines of evidence were used to achieve the evaluation. As a novel research domain, hydropower systems (HPS) present many design challenges, because controllers must coordinate multiple domains across different time frames. If controllers are to maintain effective situation awareness and if they are to exercise effective control, the displays must integrate controllers’ problem solving across water, generation, market, and transmission concerns and across time frames. However, these needs are ignored in the current displays in a hydropower control room. The two new ecologically-inspired displays, called “Functional Displays”, evaluated in this thesis, are intended to overcome the above shortcomings. The evaluation was done with industry hydropower controllers and coordinators on a medium fidelity simulator which represented the working environment of a large hydropower control room. The core issue of this thesis is how the evaluation was conceptualized, operationalized and analysed given practical constraints arising from limited access to industry controllers, limited capacity of the simulator, and the complexity of the hydropower domain. Of key concern was the design of test scenarios, selection of measures of performance, and data analysis in the face of these constraints. The rationale of the scenario design was to combine representative contingencies within and outside the hydropower scheme, emergent storage problems and different market contexts so that controllers or coordinators would be required to act in different domains. Only by placing controllers in such challenging and time pressured situations could we maximize our chances of seeing the expected benefits of the new display, given the limited length of the experiment. Multiple measures were proposed to capture the quality of in-the-loop coupling between human controllers and the system under various disturbances. The measures of performance include: (1) control quality, which investigates how the new displays tame “temporal complexity” by helping the control team construct a more effective pattern of activities to handle contingencies; and (2) control strategy, which exposes how the new display support the coordination between storage, generation and market to meet the strategic or tangible objectives. The measures of cognition and affect include: (1) situation awareness (SA), which represents controllers’ or coordinators’ degree of cognitive coupling with different time frames and different levels of the hydropower scheme and (2) controllers’ or coordinators’ trust in the supporting displays, and their self-confidence in their own ability to control the hydropower scheme. Through the effort of triangulating the different measures, balancing the elements in designing scenarios, and extracting and integrating raw data from the study, convergent lines of evidence were achieved for evaluation. It was found that the new Functional Displays led the participant teams to a better pattern of work which was reflected in their situation awareness, their discretionary control activity, and the outcome of their control with respect to system productivity and stability. However, limitations of these findings because of the constraints of the experiment are discussed. This research contributes to many theoretical and practical issues in Human-System Interface (HSI). For example, it outlines how some of the principles of ecological interface design (EID) can be used and it highlights the value of using time as a basis for distinguishing interfaces. Moreover, this research provides fruitful insights into the selection and development of measures to assess human control in real world complex work environments. Because a key issue of this thesis is how practical constraints were handled; pragmatic approaches for measuring SA and control were developed. Because industry often performs evaluations under similar intensive constraints, the approaches and solutions developed in this thesis for evaluating novel interfaces could be easily adapted to various industrial settings.

Objective: The challenge of keeping humans in meaningful control of highly automated systems is growing as these systems become more common in high-risk domains like aviation, the military, and emergency services. One suggested method to ensure human control and responsibility is to apply the principles of meaningful human control. This study aimed to explore the applicability of meaningful human control through a case study involving operators interacting with a prototype interface designed for controlling multiple unmanned aerial vehicles fighting forest fires.  Method: A simulated scenario was created and implemented through a prototype interface for human-swarm interaction. Empirical data included screen- and audio recordings of participants engaging with the simulated scenario through the prototype. The Joint Control Framework was used to transcribe and analyze the interaction.  Results: The results indicate that the level of meaningful human control in the interaction between the operator and the unmanned aerial vehicle system is dynamic and context-dependent. It varies based on the type of task, the operator’s level of cognitive control, the level of interaction with the swarm, and the system’s level of autonomy. It is realized through the joint actions performed by the operator and the system.  Conclusion: For meaningful human control to be applicable, it needs to be operationalized as a situated and contextual measure, rather than a binary concept. Future measures of meaningful human control, whether subjective or objective, should reflect this approach.

Train traffic control is a complex task in a dynamic environment. Different actors have to cooperate to meet strong requirements regarding safety, punctuality, capacity utilization, energy consumption, and more. The GMOC model has been developed and utilized in a number of studies in several different areas. This thesis describes GMOC and uses train traffic control as the application area for evaluating its utility. The GMOC model has its origin in control theory and relates to concepts of dynamic decision making. Human operators in complex, dynamic control environments must have clear goals, reflecting states to reach or to keep a system in. Mental models contain the operator’s knowledge about the task, the process, and the control environment. Systems have to provide observability, means for the operator to observe the system’s states and dynamics, and controllability, allowing the operators to influence the system’s states. GMOC allows us to constructively describe complex environments, focusing on all relevant parts. It can be utilized in user-centred system design to analyse existing systems, and design and evaluate future control systems. Our application of GMOC shows that automation providing clear observability and sufficient controllability is seen as transparent and most helpful. GMOC also helps us to argue for visualization that rather displays the whole complexity of a process than tries to hide it. Our studies in train traffic control show that GMOC is useful to analyse complex work situations. We identified the need to introduce a new control strategy improving the traffic plan by supporting planning ahead. Using GMOC, we designed STEG, an interface implementing this strategy. Improvements that have been done to observability helped the operators to develop more adequate mental models, reducing use of cognitive capacity but increasing precision of the operative traffic plans. In order to improve the traffic controllers’ controllability, one needs to introduce and share a real-time traffic plan, and provide the train drivers with up-to-date information on the surrounding traffic. Our studies indicate that driver advisory systems, including such information, reduce the need for traffic re-planning, improve energy consumption, and increase quality and capacity of train traffic. / KAJT / FTTS

One of the common research goals in disciplines such as computer graphics and robotics is to understand the subtleties of human motion and develop tools for recreating natural and meaningful motion. Physical simulation of virtual human characters is a promising approach since it provides a testbed for developing and testing control strategies required to execute various human behaviors. Designing generic control algorithms for simulating a wide range of human activities, which can robustly adapt to varying physical environments, has remained a primary challenge. This dissertation introduces methods for generic and robust control of virtual characters in an interactive physical environment. Our approach is to use the information of the physical contacts between the character and her environment in the control design. We leverage high-level knowledge of the kinematics goals and the interaction with the surroundings to develop active control strategies that robustly adapt to variations in the physical scene. For synthesizing intentional motion requiring long-term planning, we exploit properties of the physical model for creating efficient and robust controllers in an interactive framework. The control design leverages the reference motion capture data and the contact information with the environment for interactive long-term planning. Finally, we propose a compact soft contact model for handling contacts for rigid body virtual characters. This model aims at improving the robustness of existing control methods without adding any complexity to the control design and opens up possibilities for new control algorithms to synthesize agile human motion.

Air traffic controllers (ATCos) monitor flight operations and resolve predicted aircraft conflicts to ensure safe flights, making them one of the essential human operators in air traffic control systems. Researchers have been studying ATCos with human subjective approaches to understand their tasks and air traffic managing processes. As a result, models were developed to predict ATCo actions. However, there is a gap between our knowledge and the real-world. The developed models have never been validated against the real-world, which creates uncertainties in our understanding of how ATCos detect and resolve predicted aircraft conflicts. Moreover, we do not know how information from air traffic control systems affects their actions. This Ph.D. dissertation work introduces methods to evaluate existing ATCo action prediction models. It develops a prediction model based on flight contextual information (information describing flight operations) to explain the relationship between ATCo actions and information. Unlike conventional approaches, this work takes data-driven approaches that collect large-scale flight tracking data. From the collected real-world data, ATCo actions and corresponding predicted aircraft conflicts were identified by developed algorithms. Comparison methods were developed to measure both qualitative and quantitative differences between solutions from the existing prediction models and ATCo actions on the same aircraft conflicts. The collected data is further utilized to develop an ATCo action prediction model. A hierarchical structure found from analyzing the collected ATCo actions was applied to build a structure for the model. The flight contextual information generated from the collected data was used to predict the actions. Results from this work found that the collected ATCo actions do not show any preferences on the methods to resolve aircraft conflicts. Results found that the evaluated existing prediction model does not reflect the real-world. Also, a large portion of the real conflicts was to be solved by the model both physically and operationally. Lastly, the developed prediction model showed a clear relationship between ATCo actions and applied flight contextual information. These results suggest the following takeaways. First, human actions can be identified from closed-loop data. It could be an alternative approach to collect human subjective data. Second, the importance of evaluating models before implications. Third, potentials to utilize the flight contextual information to conduct high-end prediction models.

The thesis has aimed to study the emergence of a new weapons technology, also known as ‘killer robots’ or lethal autonomous weapon system. It seeks to answer what factors drive the development and deployment of this weapon system without ‘meaningful human control’, a component that allows the decision to kill to be delegated to machines. The research question focuses on seeking the motivations to develop and deploy LAWS, as well as the consequences this would have on military conduct and conflict characteristics.The incentives they bring up and the way of adopting them has been studied by synthesizing antinomic democratic peace theory and adoption capacity theory respectively. The findings of this qualitative content analysis lead to two major conclusions. (1) That LAWS present severe risk avoidance and costs reduction potential for the user. These factors have a more prevalent pull on democracies than autocracies, since they stand to benefit from LAWS’ specific capabilities more in comparison. (2) That their adoption is aided by low financial intensity needed to adopt it, due to the high commercial profitability and applicability of AI technology, and the ease of a spillover to military sphere. Their adoption is hindered by high organizational capital needed to implement the drastic changes LAWS bring. All of this leads to the prediction that LAWS are likely to proliferate further, at a medium speed, and potentially upset the balance of power.