Automation errors may result in human performance issues that are often difficult to grasp. Skraaning and Jamieson (2023) proposed a taxonomy for classifying automation errors into categories based on the visible symptoms of design problems, so as to benefit the design of training scenarios. In this paper, we propose a complementary classification that is based on the mechanisms of human-automation interaction guided by Rasmussen’s Skill, Rule and Knowledge (SRK) taxonomy. We identified four main failure classes and expect that this classification can support automation designers.@en

Previous research indicated a need to improve pilot training with regard to understanding of autopilot logic and behavior, especially in non-routine situations. Therefore, we tested the effect of problem-based exploratory training on pilots’ understanding of autopilot functions. Using a moving-base flight simulator, general aviation pilots (n = 45) were trained to diagnose failures either without foreknowledge and guidance (exploratory group), without foreknowledge but with some guidance (exploratory-guidance group) or with foreknowledge and full guidance (control group). They subsequently performed six test scenarios in which their understanding of the effects of failures was tested by requiring them to deduce the failures and select autopilot modes that were still functioning. Those who received exploratory training with guidance were significantly more likely than the other groups to diagnose failures correctly. The exploratory training group also selected the most appropriate functioning autopilot modes significantly faster than the control group. The results suggest that exploratory training with an appropriate level of guidance is useful for gaining a practical understanding of autopilot logic and behavior. Exploratory training may help to improve transfer of training to operational practice, and prevent automation surprises and accidents.

@en

Understanding human perception of haptic feedback is critical when designing and regulating these interfaces. In recent years, experiments have been conducted to determine the just-noticeable difference (JND) in mass-spring-damper dynamics, using a hydraulic admittance display in the form of a side-stick. These experiments have resulted in a model of JNDs when interacting with linear second-order dynamics. In real-world applications, however, control force dynamics also commonly include nonlinearities, such as friction. This research extends the current understanding of JNDs in linear systems by including the nonlinear case, where friction is also present. Experiments were conducted to determine JNDs in friction when combined with second-order system dynamics. Results indicate that friction JND can be independent of linear system dynamics as long as its value compared to the linear system's impedance is sufficiently large. As a consequence, friction JND follows Weber's law, also when it is combined with mass-spring-damper dynamics, unless the level of friction approaches the detection threshold, which in turn can be influenced by the linear system dynamics. Based on the findings presented, it is possible to conduct targeted experiments to confirm and add to these initial results.

@en

Accurate process control through automation is the key to achieving efficient and stable operation of a blast furnace. In this study, we developed an automatic control system of hot metal temperature (HMT). To cope with the slow and complex process dynamics of the blast furnace, we constructed a control algorithm that predicts eight-hour-ahead HMT using a two-dimensional (2D) transient model and calculates optimal target pulverized coal ratio (PCR) and pulverized coal flow rate by non-linear model predictive control (NMPC). An evaluation in a real plant showed that the developed control system suppressed the effects of disturbances, such as changes in the coke ratio and blast volume, on the HMT. The root mean square (RMS) of the control deviation of HMT was successfully reduced by 1.6 °C compared to the conventional manual operation.

@en

The Continuous Descent Approach (CDA) offers reduced aircraft noise emissions and fuel consumption, but the obstacle limiting it to reduced traffic density conditions, is the low predictability of the trajectory and Estimated Time of Arrival (ETA). The solution proposed by this research is to develop a pilot support interface to facilitate the execution of a fixed flight path angle CDA, where thrust is not limited to idle and a velocity profile can be tracked, which will lead to a selected ETA. Initially, the CDA trajectory was investigated through an aerodynamic model, while the solution space of the ETA and the calculation of the stepwise velocity profile were defined. Following the analysis of the pilot’s role, a two-fold support interface was designed based on Ecological Interface Design (EID), with a Vertical Situation Display (VSD) playing a central role for planning and execution. The interface was tested in a MATLAB®/Simulink® setup and five pilots were recruited to execute simulations over different wind conditions. Their on-time performance was satisfactory, while they worked with the provided cues and suggested some interface changes to offer more flexibility. A more robust application of the proposed approach can lead to a wider adoption of the CDA, as a validated procedure.@en

In academic air traffic control research, traffic scenarios are often repeated to increase the sample size and enable paired-sample comparisons, e.g., between different display variants. This comes with the risk that participants recognize scenarios and consequently recall the desired response. In this paper we provide an overview of mitigation techniques found in literature and conclude that rotating scenario geometries is most frequently used. The potential impact of these transformations on participant behavior, as described in this paper, is however not sufficiently addressed in most studies. As an example we, therefore, analyze previously collected eye tracking data from ten professional air traffic controllers, each presented with three repetitions in various rotations of several distinct scenarios. Results imply that researchers wishing to repeat scenarios should more carefully consider whether mitigation techniques might have an impact on their results.@en

Expert participants may not always be available for evaluation of new displays or support systems, and in some cases, it might be better to use novice participants, particularly when the display or support significantly changes existing work practices. To provide tools and arguments for selecting the expertise level of participants, we propose the use of Rasmussen’s decision ladder to analyze where and how a new visualization or a support tool changes the task, and identify steps where a novice participant may learn to perform the task to an acceptable level. A comparison to the support with the current operational interfaces then shows where an expert might have difficulty in stepping away from learned practice. This analysis is applied to the domain of air traffic control, and a selected set of relevant past research with both expert and novice participants is reviewed, revisiting the decision for a participant level in the study. @en

An improved understanding of pilot’s control behavior adaptations in response to sudden changes in the vehicle dynamics is essential for realizing adaptive support systems that remain effective when task characteristics suddenly change. In this paper, we replicate, extend, and validate the ‘adaptive pilot model’ proposed by Hess to verify its effectiveness for predicting human adaptive behavior in pursuit tracking tasks. The model relies on a Triggering function, that compares the current tracking performance to a stored nominal (pre-transition) state, and an Adaptation mechanism which determines new adapted human operator gain settings proportional to the magnitude of the off-nominal error occurrences. For model validation data from a previous experiment were used, where ten participants performed a pursuit tracking task with transitions in controlled element dynamics from a single to a double integrator, and vice versa. Overall, with an added human operator delay and participant-specific inner- and outer-loop gain adjustments, the model was found to accurately describe the measured steady-state tracking behavior for the participants in our data set. The results for the time-varying single integrator to double integrator transitions showed that the model can capture the transient control behavior of participants. However, the adaptive logic could only be tuned to activate for participants that had a pre-transition crossover frequency above 0.9 rad/s. Furthermore, the model was not able to capture the change in control behavior for transitions from a double to a single integrator. Here, as no distinct degradation in tracking performance occurs for such a transition to a more easily controlled system, the model's proposed Triggering logic will not activate. Further investigation and more experiment data are required for improving the applicability of the model's adaptive logic and to enable more accurate prediction of adaptive human control behavior.@en

To alleviate the workload of air traffic controllers, part of the air traffic may be handled by a future automated system. When deciding which flights to delegate, a distinction can be made between basic and non-basic flights, with the former being prime candidates for delegation. The human controller can then focus on the non-basic flights, where human competencies are most valuable and more difficult to automate. The classification of flights is preferably based on objective measures relating to the traffic situation. Existing complexity models are, however, often used for capacity predictions or airspace restructuring and primarily to assess the complexity of a sector as a whole. In this paper we use empirically collected flight complexity ratings from 15 professional en-route air traffic controllers. They indicated which other flights contributed to their complexity assessment of a single flight of interest. This exploratory study was able to build a machine-learning model which adequately classifies these flights, based on a qualified majority of controllers. By analyzing the interactions between the included flights, we discuss whether a classification model can differentiate between basic and non-basic flights, and which traffic features play the largest role. Once this can be done reliably and an appropriate complexity threshold has been chosen, a model can be developed as a starting point for an automatic allocation algorithm that distributes flights between a human controller and the computer.@en

Background. Mnemonic procedures are currently being taught to airline pilots to manage startle and surprise. We previously tested the effectiveness of a four-item mnemonic. Pilots generally rated it as useful but some remarked that it induced too much additional workload. Therefore, we tested whether a simpler mnemonic, Aviate-Breathe-Check, would be more useful. Method. The experiment took place in a hexapod simulator with a Piper Seneca aerodynamic model and a generic cockpit. Airline pilots (n = 25) were divided into an experimental (“ABC”) and control group. All received ground training on startle and surprise, which included instructions on the ABC mnemonic for the ABC group. The mnemonic aims to support prioritization of flight-path management (Aviate), followed by physiological and mental stress management (Breathe), followed by troubleshooting (Check). All pilots performed four familiarization scenarios, during which the ABC group practiced the ABC mnemonic. Two test scenarios were then performed to evaluate performance, mental effort, stress, and pilot evaluations of the ABC mnemonic. Results. The pilots’ evaluations of the ABC mnemonic were significantly higher than those were for the previously-tested mnemonic in the same scenarios. There were no significant differences between the ABC and control group in mental effort and stress, whereas there were trends towards higher mental effort and stress with the previous mnemonic. No significant effects on performance were found. Conclusions. The results suggest that the ABC mnemonic was more useful and easier to apply than a previously tested mnemonic. This is promising for the development of effective pilot training interventions for startle and surprise. @en

Haptic cues on the side stick are a promising method to reduce loss of control in-flight incidents. They can be intuitively interpreted and provide immediate support, leading to a shared control system. However, haptic interfaces are limited in providing information, and the reason for cues may not always be clear to pilots. This study presents the results of the conceptual development of visual display symbology that supports haptic feedback on the side stick in communicating flight envelope boundaries to pilots. Novel indications for the limits of airspeed, load factor, angle of attack, and angle of bank, which for the first time simultaneously indicate magnitude and direction of the haptic cues, were integrated in an Airbus primary flight display. The symbology was tested in a pilot-in-the-loop experiment with professional Airbus pilots (N=16) flying several approaches in alternate law with haptic feedback. Objective results do not show clear improvements, although the time spent outside the flight envelope is slightly reduced. Subjective results indicate a preference, however, for the new display and an increased understanding of the haptic feedback. Further research is recommended to improve the interface design, remove unused indications, and test a bank scenario using current operational bank limits.@en

In the design of human-like steering support systems, driver models are essential for matching the supporting automation's behavior to that of the human driver. However, current driver models are very limited in capturing the driver's adaptation to key task variables such as road width and visibility (i.e., 'preview' of the road ahead). This paper uses a recently proposed, novel control-theoretical model for centerline tracking to investigate driver steering in lane-keeping tasks with restricted and unrestricted preview, in an attempt to substantially extend this model's validity. Using data from a tailored driving simulator experiment, three driver control loops (feedforward, heading and position feedback) are separately quantified using system identification techniques. The results show that when preview is restricted, drivers use all of the remaining preview to anticipate the curves of the road ahead, and are no longer able to 'smooth' tight curves in the road trajectory (i.e., corner cutting). When sufficient preview and lane width are available, the time to line crossing increases, and steering behavior is less aggressive and more intermittent, or more 'satisficing'. The novel driver steering model captures these adaptations very well (over 95% of the steering actions) and can thereby be instrumental in realizing human-like steering automation and support systems.

@en

This paper considers the combined effect of two trends in commercial aviation. On the one hand, there is a continuing demand for pilots, implying that a new generation of pilots, will soon be flying our aircraft. On the other hand, legal aspects have had an adverse effect on innovation in the safety level of established procedures, leading to a trend for aircraft operating companies to adopt manufacturer’s recommended flight deck procedures rather than reviewing these and adapting these where appropriate to local needs. However, with the influx of new pilots to the workforce, the lack of innovation and adaptation of flight deck procedures poses a safety treat. The safety level of the aviation industry relies on the experience of the individual operator (pilot), the demands of the tasks and the available support tools. If the experience is not in the operator (pilot) we have to put the experience in the procedures to maintain the existing safety level in aviation. This conference paper is a plea to develop type, and operation, specific flightdeck procedures, adapted to present day operation and usable by the future pilot population. @en

Background. Previous studies and accident analyses have shown that pilots can make roll reversal errors when responding to bank angles shown by the artificial horizon in the Primary Flight Display (PFD). In the current study, we tested whether adding stereoscopic depth cues to the artificial horizon may lead to better bank angle representation due to an improved figureground separation between the symbols. Method. Stereoscopic depth cues were created by using a half-silvered mirror multi-layer PFD, which presented the horizon symbol on a lower layer and the aircraft symbol on a higher layer. A group of 23 non-pilots and 18 general aviation pilots were shown left or right bank angles on this multi-layer PFD as well as on a normal single-layer PFD, with the task to roll the wings level using a joystick. Results. In the pilot group, a similar amount of roll-reversal errors was made with both displays (median = 3.3%) with no significant difference, p = 0.635. In the non-pilot group, fewer roll-reversal errors were observed with the multi-layer display, but this difference did not reach significance either (median = 3.3% vs. 5.0%, p = 0.182). In both pilots and non-pilots, the reaction time was longer in the multi-layer display, which reached significance in the non-pilots (p = 0.016) but not in pilots (p = 0.215). Participants noticed that the depth was only visible during the start of the session. Conclusions. The results suggest that using stereoscopic depth cues are not a viable manner to enhance the figure-ground relation in the artificial horizon. @en

Cyberneticists develop mathematical human control models which are used to tune manual control systems and understand human performance limits. Neuroscientists explore the physiology and circuitry of the central nervous system to understand how the brain works. Both research human visuomotor control tasks, such as the pursuit tracking task. In this paper we discuss some commonalities and differences in both approaches to better understand the adapting human controller. Special attention is given to Adaptive Model Theory, which studied adaptive human control using several linear and nonlinear control engineering techniques. The insights gained yield schemes and concepts which pave the way for key future work on this topic.

@en

Air traffic controller workload is a limiting factor in the current air traffic management system. Adaptive support systems have the potential to balance controller workload and gain acceptance as they provide support during times of need. Challenges in the design of adaptive support systems are to decide when and how to trigger support. The goal of this study is to gain empirical insights into these challenges through a human-in-the-loop experiment, featuring a simplified air traffic control environment in which a novel triggering mechanism uses the quality of the controller's decisions to determine when support is needed. The designed system seeks to prevent high workload conditions by providing resolution advisories when the controller exceeds a threshold of 'self-complicating' decisions. Results indicate that the new system is indeed capable of increasing the efficiency and safety compared to full manual control without intervention. More adaptive support, however, increased the frustration of participants, decreased acceptance, and did not result in improved workload ratings. These findings suggest that, unless we can better infer human intent in complex work environments, adaptive support at the level of decision-making is problematic. A potentially more fruitful direction is to provide support at the level of information integration, with full decision-making authority with the human.

@en

Mathematical human control models are widely used in tuning manual control systems and understanding human performance. Human behavior is commonly described using linear time-invariant models, averaging-out all non-linear and time-varying effects, which are gathered into the remnant. These models are limited in their capability to capture particular tracking strategies that an experienced subject may learn to use. In this paper, we consider manual control from a different perspective, namely through investigating the probability densities of the tracking error for different regions of the target signal amplitude. Results show that distinct strategies become apparent for compensatory, pursuit and preview tracking tasks. Effects of these strategies are often averaged-out by current models and can only be captured in situation-dependent models. Modeling this systematic human adaptation not captured in linear models could potentially lead to better model fits and explain/reduce part of the remnant.

@en

To achieve the automation of blast furnace operation, an automatic control system for hot metal temperature (HMT) was developed. Nonlinear model predictive control (NMPC) which predicts up to ten-hour-ahead HMT and calculates appropriate control actions of pulverized coal rate (PCR) was constructed. Simulation validation showed that the NMPC algorithm generates control actions similar to those by the operators and that HMT can be maintained within ±10 °C of the set point. The automatic control system using NMPC was then implemented in an actual plant. As a result, the developed control system suppressed the effects of disturbances, such as the changes in the coke ratio and blast volume, and successfully reduced the average control error of HMT by 4.6 °C compared to the conventional manual operation. The developed control system has contributed to the reduction of reducing agent rate (RAR) and CO2 emissions. @en

In the quest for more efficient air traffic management, a common approach is to allocate an increasing amount of functionality to higher levels of automation, with a supervisory role for humans. This potentially leads to forthcoming issues such as skill degradation and out-of-the-loop phenomenon. If the traffic in an airspace is instead shared between a human operator and an automated system, with specific flights fully delegated to automation, operators can maintain their skills and stay actively involved in controlling the rest of the traffic. This does, however, lead to new forms of mixed conflicts, where two flights are controlled by different agents. A smart flight allocation strategy, starting with the delegation of basic flights requiring little monitoring or cognitive effort, is expected to improve combined human-automation performance. In this paper, we present flowcharts to model en-route air traffic controller cognitive think and action processes in two core tasks: conflict detection and resolution. We qualitatively describe the impact of delegating flights to automation and the associated introduction of mixed conflicts. Once empirically validated and quantified in follow-up research, these models can be used to design flight allocation strategies for future human-automation teams.@en

Without intervention the widespread adoption of autonomous vehicles could be compromised by an increased incidence of motion sickness compared to conventional cars. To investigate whether passengers' motion sickness can be reduced by manipulating an autonomous vehicle's accelerations on a fixed route without altering the travel time, a human-out-of-the-loop experiment was performed in the SIMONA Research Simulator at Delft University of Technology. The experiment consisted of two different driving conditions, in which an identical 22-km road including 52 curves was travelled in 30 minutes. Condition 1 comprised larger longitudinal, but smaller lateral, acceleration values compared to Condition 2. Experimental results suggested that Condition 1 resulted in more severe motion sickness than Condition 2, with fitted learning curves providing final MIsery SCale scores of 1.19 vs. 0.80. A similar relative difference between the two conditions had been predicted by the 6-DOF Subjective Vertical Conflict model. Hence, this model has the potential to, once further developed, support the design of autonomous vehicles by reducing the need to perform costly, time-consuming experiments.

@en