This paper presents a three-step validation approach for subjective rating predictions of driving simulator motion incongruences based on objective mismatches between reference vehicle and simulator motion. This approach relies on using high-resolution rating predictions of open-loop driving (participants being driven) for ratings of motion in closed-loop driving (participants driving themselves). A driving simulator experiment in an urban scenario is described, of which the rating data of 36 participants was recorded and analyzed. In the experiment's first phase, participants actively drove themselves (i.e., closed-loop). By recording the drives of the participants and playing these back to themselves (open-loop) in the second phase, participants experienced the same motion in both phases. Participants rated the motion after each maneuver and at the end of each drive. In the third phase they again drove open-loop, but rated the motion continuously, only possible in open-loop driving. Results show that a rating model, acquired through a different experiment, can well predict the measured continuous ratings. Second, the maximum of the measured continuous ratings correlates to both the maneuver-based (ρ =0.94) and overall (ρ =0.69) ratings, allowing for predictions of both rating types based on the continuous rating model. Third, using Bayesian statistics it is then shown that both the maneuver-based and overall ratings between the closed-loop and open-loop drives are equivalent. This allows for predictions of maneuver-based and overall ratings using the high-resolution continuous rating models. These predictions can be used as an accurate trade-off method of motion cueing settings of future closed-loop driving simulator experiments.

This manual provides guidance for human factors researchers and applied psychologists, on the standardized and scientifically rigorous use of the instruments. It is structured to include an overview of the instruments, administration guidelines, and a summary of their psychometric properties to support accurate application and interpretation in research and operational contexts.

Providing adequate simulator motion cues for simulated upset and stall scenarios remains challenging. This paper evaluates the potential of novel optimization-based motion cueing algorithms for upset and stall simulation. An offline analysis is performed to compare three Model Predictive Control (MPC) algorithms with varying prediction horizon lengths and prediction strategies (i.e., "Oracle", "Perfect", and "Constant") against two baseline classical washout algorithm implementations from literature. The analysis is performed for a symmetric stall scenario flown with TU Delft's Cessna Citation II laboratory aircraft. Overall, the analysis shows that the objective motion cueing quality expressed in terms of the Root Mean Square Error (RMSE) improves by 29.8% (for specific forces) and 18.7% (for rotational velocities) with the "Oracle" and "Perfect" MPC implementations compared to the reference classical washout results. For the "Constant" MPC algorithm, which in fact does not include any explicit prediction across the MPC algorithm's future prediction window, only a marginal improvement in motion quality was found. Overall, these results imply that, assuming a sufficient future reference motion prediction can be achieved, optimization-based motion cueing algorithms have the potential to provide significantly better motion cueing quality compared to classical motion cueing algorithms.

Energy management is essential in low-energy flight conditions. Changes in fixed-wing aircraft configuration affect performance and energy boundaries, and improved insight therein should allow pilots to better predict potentially dangerous situations, maintain suitable safety margins, and more effectively react to unforeseen events. This paper presents the design and experimental evaluation of a vertical situation display with enhancements portraying changes in the flight performance envelope. Sixteen pilots were tasked to fly approach and go-around scenarios with both a baseline and an enhanced display, with some of the scenarios including unexpected failures in configuration changes. Results show that the new display makes pilots maintain larger margins in velocity, thus spending less time below the advised minimum speed limit in final approach. However, these larger velocity margins also led to larger errors with respect to target velocities. Failures in configuration changes were more quickly discovered with the new display, although these results could not be substantiated due to a lack of statistical significance. However, pilots did report feeling better able to predict dangerous situations. Overall, pilots preferred the novel display; no significant differences in workload were found.

This paper outlines the three-phase construction of the Startle and Surprise Inventories (Startle-I; Surprise-I) and Visual Analogue Scales for Startle and Surprise (Startle-VAS; Surprise-VAS). In Phase 1, seven experts in the field assessed the content validity of 14 items for surprise, 7 items for startle derived from fundamental and applied literature. Elimination of items was based on a 50% agreement of relevance. In Phase 2, 81 participants completed the retained 19 items nine times, each time immediately after watching a video clip. A multilevel exploratory factor analysis was applied to assess the construct validity of items. In Phase 3, concurrent validity of the Startle-VAS and Surprise-VAS was tested by comparing with the Startle-I and Surprise-I scores, respectively. The first two phases yielded a 11-item two-factor solution, corresponding to the constructs of startle and surprise. These results supported Startle-I and Surprise-I as measures of self-report startle and surprise, with Startle-VAS and Surprise-VAS as efficient alternatives.

We tested whether pilots would detect low-salient controllability problems more quickly during manual compared to automated flight. Using a moving-base simulator and a Piper Seneca aerodynamic model, airline pilots (n = 20) performed scenarios in which either a gradually ensuing single-engine failure or an icing accumulation occurred. Both scenarios were performed once during manual flight and once during automated flight, and were alternated with distraction scenarios. The icing accumulation was detected marginally significantly more quickly during manual flight, while there was no significant difference for the engine failure. Problems in manual flight were, as expected, most likely discovered from aircraft motions or control forces. Interestingly, there were several late detections during manual flight which appeared to be caused by subconscious manual corrections. In automated flight, the engine failure was discovered most often from the engine manifold pressure indication, while the icing accumulation was most often discovered from control column movement. The results therefore underline the importance of using back-driven controls, and further indicate that manual flight does not necessarily improve detection of problems that occur without display indications.

We investigated the effect of personality traits and flight experience on pilot cognitive and affective responses across seven startling and surprising scenarios performed in motion-based simulators. A dataset of 89 airline pilots from four studies was used. The personality traits measured were trait anxiety, decision-related action orientation (AOD), and failure-related action orientation (AOF). Pilot self-reported responses in scenarios were standardized by obtaining z scores of startle, surprise, stress, and mental workload. Only trait anxiety was found to be significantly positively correlated with stress. No significant effects of AOD, AOF, or flight hours were found on pilots’ responses. The results indicate trait anxiety may affect pilots’ responses to stressful scenarios, even though pilots are selected based on low trait anxiety.

Startle and surprise can impair pilot performance and affect flight safety. This study investigates the prevalence of different startle and surprise events among helicopter pilots, its impact on pilot stress and mental effort and the influence of training background. It also looks at currently used startle mitigation strategies and evaluates the usability of a previously proposed “Aviate, Breathe, Check (ABC)” startle management method (Piras et al. 2023). A survey among 234 helicopter pilots revealed that 96% had experienced impactful startle or surprise events during operations. Scenarios such as disorientation, tail rotor incidents, and flight into instrument meteorological conditions (IMC) were considered particularly stressful. Reported levels of stress and mental effort during startle and surprise events did not differ between pilots with higher and lower experience levels or between pilots with a different training background (military or civilian). Only 38% of pilots indicated they were specifically trained to deal with startle and surprise and only 1% were trained to use a breathing technique. Most pilots (90%) expressed openness to implementing the ABC method and expected benefits from using it. Concerns regarding time constraints in critical situations emerged as the primary objection to adopting this technique. Overall, the findings indicate that the introduction of a startle management method tailored for helicopter operations could significantly enhance safety, especially given the higher accident rates compared to fixed-wing operations. Future research should focus on developing effective training protocols that account for the unique challenges of helicopter flying.

The introduction of more advanced automation in air traffic control seems inevitable. Air traffic controllers will then take the role of automation supervisors, a role which is generally unsuitable for humans. Gamification, the use of game elements in non-gaming contexts, shows promising results in mitigating the effects of boredom in highly automated domains requiring human supervision. An example is luggage screening, where dangerous items are rarely found, through projecting fictional threats on top of x-ray scans. This paper presents and experimentally tests a proposed implementation of gamification within highly automated en-route air traffic control. Fictional flights were superimposed among automatically controlled real traffic, thus creating fictional conflicts that needed resolving. System supervisors were tasked to supervise the behaviour of a fully automated conflict detection and resolution system, while manually routing fictional flights safely and efficiently through the sector, avoiding conflicts with both real and fictional flights. Automation anomalies were simulated, as well as an automation failure event, after which the system supervisor needed to assume manual control over all traffic. The presence of fictional flights increased self-reported concentration levels and reduced boredom. However, some participants reported that fictional flights were distracting. Thus, while the use of fictional flights increases engagement, it might negatively affect other cognitive functions, and with that, compromise safety. Thus, while the implementation of such a tool might provide benefits in terms of skill retention and engagement, further research is recommended involving professional air traffic controllers, improved measurement tools and a longitudinal study that better excites boredom, complacency, and skill erosion in order to understand and mitigate its negative effects.

While human control behavior is well-understood in continuous control tasks, little is still known about how human operators detect sudden changes in the controlled element dynamics. This paper focuses on modeling this detection phase for pursuit tracking tasks. Potential triggers for the human operator to detect changes in the controlled element dynamics were investigated via a time-varying computer simulation. Based on the results, hypotheses were generated and later tested in a single-axis pursuit tracking experiment with fifteen participants. Transitions from approximate single to approximate double integrator dynamics and vice versa were investigated, for which participants indicated if they detected the transition by pressing a button. Using the button push data, a model for each transition was developed and validated. The models work under the assumption that human operators use a threshold, a multiple of the steady-state standard deviation, on certain signals to detect transitions. The models developed for the transition from single to double integrator dynamics and vice versa are proposed to trigger on the tracking error and system output acceleration, respectively. They have an accuracy of 88.9% and 99.4%, respectively. However, a consistent underestimation of the detection lag remains a limitation of both models. Nonetheless, this research helped confirm the tracking error can be used in a model for the transition from single to double integrator dynamics, proposed a model for the opposite transition, and identified that the relationship between control inputs and the system's response as a crucial factor for the detection.

This erratum applies to the following published paper [1]. In Fig. 10(e) and (f) of the published version of the paper, the measured values for the t _f and T _l,f (both having values around 1 s for the considered dataset) were interchanged. This erratum includes both the published and corrected versions of Fig. 10 and the related paragraph in the paper's Results section. PUBLISHED VERSION In preview tasks, bandwidth changes yield only minor adaptations in the model parameters, see Fig. 10. Only the average look-ahead time t _f decreases slightly with bandwidth from around 1.05 to 0.9 s, Fig. 10(e) but with substantial between-participant variability, as indicated by the overlapping confidence intervals. The lower t _f may not reflect a systematic adaptation to the bandwidth, but a more subtle adaptation to minimize the errors due to the additional high-amplitude sinusoids at 2.5 and 4 rad/s, see also Fig. 8c. The general way in which participants use the available preview for control is, however, not affected by the target signal bandwidth: the target response gain (K _f ≈ 0.95, Fig. 10(d)) and lag time-constant (T _{l, f} ≈ 1.15 s, Fig. 10(f)) are approximately invariant. The estimated control dynamics in Fig. 12 show that the target trajectory is tracked almost perfectly at all frequencies below 4 rad/s, mostly because the phase lead due to τ _f allows for synchronizing the CE output with the target signal (as opposed to pursuit tasks, see Fig. 11, bottom right). Therefore, different-bandwidth target signals provide no incentive for HCs to strongly adapt their control behavior in preview tasks. (Figure Presneted) CORRECTED VERSION In preview tasks, bandwidth changes yield only minor adaptations in the model parameters, see Fig. 10. Only the average target smoothing time-constant T _{l, f} decreases slightly with bandwidth [from around 1.05 to 0.9 s, Fig. 10(f)], but with substantial between-participant variability, as indicated by the overlapping confidence intervals. The lower T _{l, f} indicates that slightly more smoothing is applied to reduce tracking of the more high-amplitude high-frequency sinusoids in the 2.5 and 4 rad/s bandwidth signals through the feedforward response, see also Fig. 8(c). The general way in which participants use the available preview for control is, however, not affected by the target signal bandwidth: the target response gain [K _f ≈ 0.95, Fig. 10(d)] and look-ahead time [τ _f ≈ 1.15 s, Fig. 10(e)] are approximately invariant. The estimated control dynamics in Fig. 12 show that the target trajectory is tracked almost perfectly at all frequencies below 4 rad/s, mostly because the phase lead due to τ _f allows for synchronizing the CE output with the target signal (as opposed to pursuit tasks, see Fig. 11, bottom right). Therefore, different-bandwidth target signals provide no incentive for HCs to strongly adapt their control behavior in preview tasks. (Figure presented).

An online pilot manual control behavior identification method, based on recursive low-order time-series model estimation, is presented and validated using experimental data. Eight participants performed compensatory tracking tasks with time-varying vehicle dynamics, where, at an unpredictable moment during a run, a sudden degradation in dynamics could occur. They were instructed to push a button when they detected a change in dynamics. Two methods to automatically detect the moment when pilot adaptation occurs from online estimated parameter traces are discussed. Results show that pilots are more accurate in detecting changes than either algorithm. But when the algorithms are correct, they are often quicker to detect pilot adaptation than pilots themselves. The presented techniques have potential but need improvements.

The identification of time-varying, adaptive behavior of a human operator in basic manual control tasks is currently still a focus area, since most methodologies only account for time-invariant system dynamics. Previous authors have proven that estimation techniques based on ARX model structures can be used to identify time-varying HO model parameters. However, ARX methods do present several problems, such as a persistent bias in the obtained estimates of the HO model poles (neuromuscular parameters) that increases due to coupled noise and system models. Therefore, in this paper a novel identification technique based on Box-Jenkins (BJ) models is proposed, to achieve a better match between the BJ estimator's inherently uncoupled system and noise models and measured HO control dynamics. The identification process was tested offline (batch-fitting) using Ordinary Least Squares and the Prediction Error Method for both ARX and BJ models, respectively, or online when Recursive Least Squares and Recursive PEM are employed. The BJ estimator has excellent potential as an identification tool due to its bias reduction capabilities, as clearly shown in batch-fitting, although the non-linear optimization processes decrease its convergence speed by 500%. An RPEM algorithm with a forgetting factor of λ = 0.99609 and a first-order remnant model incorporated in the BJ structure was tested on Monte Carlo simulation and experimental data. While the recursive BJ estimator showed the same bias-diminishing advantages also seen in batch-fitting, the non-linear RPEM estimator's results showed much slower convergence after HO behavior adaptations and frequent instabilities of the obtained parameter estimates. Hence, further research is needed for implementing a practical bias-free HO model estimator based on the BJ model structure.

Previous studies have indicated that the attitude director indicator (ADI) used in commercial aviation is suboptimal in representing the bank angle direction, which can lead to confusion, roll reversal errors and increased workload. Confusion about the bank angle direction has been implied in several cases of loss of control in-flight (LOC-I). In the current study, we therefore tested whether bank angle representation can be improved by adding non-disruptive visual depth cues to the ADI. An enhanced ADI was created, in which three monocular cues were added: atmospheric haze (i.e. a gradient in color towards the horizon), a shadow line under the aircraft symbol, and perspective lines on the ground. Airline pilots (n = 25) were tasked with rolling back to level 96 times from unforeseen (30 or -30 degrees) bank angles after experiencing either matching or mismatching (disorienting) roll motion cues in a motion-base simulator. There was no outside visibility and pilots responded using the ADI only. Roll reversal errors and reaction times were compared within-subject between the enhanced and baseline ADI, which were both based on the B747. Pilots were tasked to respond immediately upon presentation of the display, so that their initial interpretation of bank angle direction could be measured. There was no significant difference in roll reversal errors, and a significant increase in reaction times, when using the enhanced ADI compared to the baseline ADI. This suggests that pilots had slightly more difficulty with reading the bank angle with the enhanced ADI. Of the pilots, 56% preferred the enhanced ADI over the baseline display as it is, 8% had no preference and 36% preferred the baseline ADI. The most valued addition was the perspective lines on the ground, which pilots remarked would also be helpful in recovering extreme attitudes. The most-heard concerns were about potential clutter caused by the added cues, and difficulty with accurate reading of the pitch angle due to the shadow lines. In conclusion, according to the pilots' feedback, the addition of depth cues to the ADI appears promising, but it should be tested using more challenging tasks. Further design changes also appear needed to prevent clutter and facilitate quick reading of the aircraft attitude.

This article discusses a long short-term memory (LSTM) recurrent neural network that uses raw time-domain data obtained in compensatory tracking tasks as input features for classifying (the adaptation of) human manual control with single- and double-integrator controlled element dynamics. Data from two different experiments were used to train and validate the LSTM classifier, including investigating effects of several key data preprocessing settings. The model correctly classifies human control behavior (cross-experiment validation accuracy 96%) using short 1.6-s data windows. To achieve this accuracy, it is found crucial to scale/standardize the input feature data and use a combination of input signals that includes the tracking error and human control output. A possible online application of the classifier was tested on data from a third experiment with time-varying and slightly different controlled element dynamics. The results show that the LSTM classification is still successful, which makes it a promising online technique to rapidly detect adaptations in human control behavior.

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.

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.

The collective goal of the driving simulation community should be to share ideas to improve the motion cueing across driving simulators worldwide. Due to the active research and intensive usage of driving simulators over the last decades, knowledge in the field of motion cueing has been gained from experience gathered by the community, or practical experience by performing dedicated experiments. This paper discusses several points of ‘common knowledge’ in designing and evaluating motion cueing, along with their value for driving simulation. The goal of the discussion in this paper is to compare these points of common knowledge to the experiences and ideas gathered at BMW’s driving simulation center in Munich, which was opened in 2021 and hosts a fleet of fourteen driving simulators. Furthermore, we aim to bring across points of interest and outlines for future research that should be of interest to the driving simulation community. With the common goal of improving motion cueing, this contribution thus aims to improve and extend the discussion between those working and researching in the driving simulator industry.