Completed thesis

This study investigates the relationship between the personality traits conscientiousness and openness to new experience and performance when working on the (dis)assembly of an angle grinder. Personality traits were assessed using the Big-Five-Inventory. Performance was measured by task duration. In a pilot study with 36 participants, some receiving prior training on the task, no significant correlations between personality traits and stable performance were found. However, when completing a novel variant of the angle grinder, used as a measure of adaptive performance, a significant correlation was found between higher levels of conscientiousness and longer task completion times during assembly. The confidence interval included zero, suggesting the result should be interpreted cautiously. In addition, untrained participants demonstrated significantly steeper learning curves across repeated trials. These results highlight the importance of considering individual personality differences and training effects in manual task performance and emphasize the importance of a differentiated analysis of task structures.

Despite advancements in autonomous driving technology, challenges that necessitate human intervention remain. Teleoperation serves as a viable solution by allowing remote operators (ROs) to assist highly automated vehicles (HAVs) when they encounter difficulties. These difficulties can be roadworks, diversions or unreadable signs. A critical element in developing teleoperation systems is maintaining strong situation awareness (SA), especially during nighttime when the reliability of autonomous systems is reduced. This Master’s Thesis aims at exploring how time of day influences the SA of ROs. The research questions explore how lighting conditions during day and night teleoperation affect SA and confidence levels. A quantitative within-subject study (N = 32) was conducted in which the participants engaged with simulations designed to mirror situations likely to result in vehicle disengagements. Situation Awareness Global Assessment Technique (SAGAT), Situation Awareness Rating Technique (SART), and verbal elicitation methods were employed as assessment techniques. Statistical analyses revealed significant effects of time of day on Level 1 SA (V = 274.5, p < .05), Level 3 SA (t(31) = 2.15, p < .05), and operator confidence (V = 496, p < .01). These findings underscore the necessity of enhancing operator Human-Machine Interfaces (HMIs) to better support SA during nighttime teleoperation. Recommendations for HMI design are presented based on the analysis of both statistical results and audio recordings gathered during the experiment. This work addresses the current lack of research on teleoperation under nighttime lighting conditions.

The transition toward conditionally automated driving presents new challenges in the design of effective Human–Machine Interfaces (HMIs). To support safety, trust, and technology acceptance, HMIs must communicate system states and takeover requests clearly and inclusively. Previous research has demonstrated that color-coded ambient lighting can enhance driver awareness and user experience. However, ambient lighting varies widely between vehicles and is often complex to implement. This study introduces a novel approach by transferring color-coded communication to screen-based HMIs, aiming to increase compatibility and scalability across different vehicle platforms.
A within-subject driving simulator experiment (N = 36) was conducted to compare three multimodal HMI prototypes: a conventional baseline interface with text- and symbol-based cues, a screen-based color interface, and a prototype combining a screen interface and ambient lighting. Each prototype was evaluated under varying cognitive demands with respect to takeover performance, workload, trust, acceptance, and user experience. Results revealed no statistically significant differences in takeover performance across HMI types. However, the ambient lighting-based HMI was consistently preferred by users and rated highest in perceived innovation and hedonic quality. The screen-based HMI also received significantly higher preference ratings than the baseline. It also outperformed the baseline in trust ratings. No significant gender-based differences were found, and the small number of participants with color vision deficiencies precluded subgroup analysis.
These findings suggest that while color-coded visual elements do not measurably improve takeover performance in multimodal systems, they enhance subjective user experience and trust. In particular, screen-based HMIs represent a practical and broadly applicable alternative to ambient lighting systems in the context of automated vehicle design.

The study examines the field of human-AI interaction: the influence of cognitive support strategies in explainable artificial intelligence (XAI). The aim of the study was to determine the extent to which individual differences in error rates and processing time, in particular the need for cognition (NfC), affect XAI.
Participants in a study had to complete tasks with no AI support and with various AI supports. The participants had to complete the tasks under three different conditions. The data was collected using Python programming and self-assessment by the participants. The data analysis was performed using SPSS and includes the characteristic values and correlations.
The results showed no significant correlation between NfC and the error rate. Although the data suggest this, the data were not robust enough to clearly confirm the hypothesis.

The results highlight the importance of cognitive support strategies in XAI systems. In future, participants should be provided with flexible explanations and differentiated design options.
 

One key enabler that renders end-of-use products useful again is through remanufacturing practices. By subjecting obsolete products to a series of process steps they can be reintroduced into the use cycle. This circular approach replaces the hitherto linear system of material handling. This concept benefits both firms and consumers since these goods can be offered on the market for less than the price of a newly manufactured product and still yield a greater revenue. The environment profits by a lesser devotion of natural resources. A product that cannot be disassembled renders it useless for remanufacturing. Significant lever for an unobstructed disassembly is a product’s design. This thesis evaluates an example product (power tool) with regard to its design-wise suitability for disassembly in remanufacturing. For this purpose, the sample product is manually disassembled. The graphic representation of the disassembly is carried out using a Petri net. Problems hindering the disassembly are identified and solution statements to overcome these hindrances are presented. By employing Design for Disassembly guidelines, alternative product design concepts are proposed. For the academic reference, the product-specific findings are finally raised to a scientific level.

One critical aspect within the emerging field of insect farming is ensuring the occupational safety of employees operating within controlled environments. This qualitative study, based on theoretical foundations from BAuA (Federal Institute for Occupational Safety and Health in Germany) and NIOSH (National Institute for Occupational Safety and Health in the US), aims to address this challenge. Through semi structured interviews analyzed according to Mayring’s direct content analysis method, the study highlighted significant environmental constraints faced by operators during the draining operation, considered as the most difficult operation. The main challenges identified were high temperatures, humidity, and the presence of drosophila. Temperature and humidity were measured using sensors to ensure accurate data collection for assessing environmental conditions. These conditions caused excessive sweating, discomfort, and occasionally hindered breathing, impacting operators' performance and efficiency. Physical strain was also notable, especially due to awkward postures like squatting and leaning. Addressing these issues through ergonomic improvements, proper hydration practices, and better protective equipment can help mitigate the adverse effects of heat and physical strain, ultimately improving the working conditions and efficiency of the operators.

The goal of this work is to analyze the working environment of safety drivers for autonomous shuttles. This will examine the demands and the necessary skills of the position. These demands and skills are compared with those of normal motorists. In this way it can be investigated whether the automation of the shuttles relieves the safety drivers of any of their concerns. This is done through qualitative interviews. The interviews demonstrate that the work of safety drivers in the shuttle do not decrease, but that the importance of individual competencies is shifting.

With the rapid development of the intelligent transportation field, the integration of remote autonomous driving and remote assistance technology has become a key area of innovation. These dual components work together to redefine how vehicles are operated and supported over the network, allowing human operators to not only take control but also provide critical support to autonomous vehicles from a distance. This approach not only addresses the inherent challenges that autonomous vehicles encounter in complex driving scenarios, but also enhances the reliability and decision making capabilities of these systems. By providing manual decision support and monitoring capabilities, operators can intervene in special situations, thereby enhancing the safety and effectiveness of operations.

However, practical testing of these technologies poses considerable challenges, including substantial time requirements and the potential for unforeseen risks. The possibility of test failure not only brings the risk of financial loss, but also the threat of dangerous events. Given the myriad possible scenarios, exhaustive testing is both impractical and dangerous. Therefore, seeking a safe and efficient testing method has become an important research task.

To address these challenges, our proposal introduces a comprehensive remote operations strategy that leverages the sophisticated simulation capabilities of CARLA and Robot Operating System (ROS) environments. The strategy is enhanced by a novel machine learning algorithm that proposes waypoint suggestions based on historical data to improve remote operator operational efficiency while keeping vehicles and personnel safe. Through preliminary data collection on the Town04 map in the CARLA simulation environment, followed by rigorous testing on a variety of maps, our experimental results highlight the effectiveness of our approach. As the amount of data collected increases, we observe a significant increase in the success rate of remote operations, emphasizing the practical feasibility and potential of our approach for practical applications.

This study not only makes a significant contribution to the field of autonomous driving, but also provides feasible solutions to the challenges faced by intelligent transportation systems in complex environments.

The aim of this master's thesis was to carry out an analysis of the potential for optimizing manufacturing process chains in a high-variant production system. The production of the GPH 1 plant of ZIEHL-ABEGG was considered as a case study. For this purpose, the organizational form of the production was first determined and the suitable analysis methods for finding optimization potentials were derived from this. A layout analysis determined the degree of space utilization for value-adding activities in the production to be below 40 %. For this reason, the material flow was depicted in a Sankey-diagram and a value stream analysis was carried out for product families. By combining the results from these analysis methods, the optimization potentials could be derived. The greatest potential for optimization is the elastic interlinking by introducing a FiFo system between the machines with the largest material flow.

Research into aspects of artificial intelligence has experienced a rapid increase in recent years. In particular, research into explainable artificial intelligence (XAI) has therefore been the subject of many studies.
This thesis examines the design of effective XAI systems by presenting the combination of explorative possibilities for researching explanations and the use of counterfactual explanations as a design option and evaluating their effectiveness.
For this purpose, a laboratory experiment was conducted with 70 test persons in which three different exploratory conditions were used, which were evaluated in terms of their effect on subjective and objective understanding. In addition, the effect of trust in the generated explanations was also evaluated.

The results of the study show that the exploration conditions used had no influence on objective understanding, while subjective understanding was increased by the static test condition. Furthermore, trust in the explanations has no influence on the effect that the exploration conditions have on subjective or objective understanding.

Artificial intelligence (AI) is increasingly being used in small and medium-sized enterprises (SMEs) to ensure competitiveness. The use of AI can lead to an improvement in productivity by automating tasks, analyzing data and providing insights that help in decision-making. The introduction of AI in SMEs leads to numerous changes for employees, especially in the case of interaction between humans and AI. To ensure productive collaboration, employees must be taken into account right from the start of the introduction. This paper examines the practical relevance of a theoretical model by Makarius et al. (2020). A checklist for the human-centered introduction of AI is developed from the contents of the model and the results of a qualitative data collection with consulting companies. As part of a quantitative data survey of employees in SMEs, the checklist is evaluated for its practical suitability. This is positive.

Hand gesture control is an innovative operating concept for controlling devices. operating concept for controlling devices. Gesture control can have a lasting impact on humanmachine interaction. The goal of this thesis was to investigate the influence of menu type and number of buttons on input time, error rate, and user friendliness. To investigate this research question, an experimental subject study was conducted. The task of the test subjects was to enter numbers using inertial-based gesture control. A linear menu and radial menu with five and ten buttons respectively were available for this purpose. The study showed that the linear menu was superior to the radial menu in terms of input time. When using the me-menu with five buttons, the radial menu allowed a minimally lower error rate. When using the menu with ten buttons, the error rate was significantly lower when using the linear menu. In terms of userfriendliness, the me-nus did not differ significantly. Our study concludes that linear menus should be preferred when using gesture control.

With the development of technology, autonomous vehicles will become more and more involved in road traffic in the near future. A mixed traffic involving autonomous vehicles and human drivers brings in many challenges, especially at unsigned intersections. Hence it is in particular crucial to understand and predict drivers’ behaviours. This article uses computer-aided computational methods to calculate the complexity of intersections. It then analyses whether these computational methods can predict perceived complexity through linear regression and random forests regression. Finally, these video-based computer-aided complexity calculation methods are used to perform predictions of human driver behaviours at T-intersections. The results show that the driving behaviours can be predicted by using the computer-aided complexity calculation methods.
Key words: Intersection complexity, driving behaviours prediction, video processing

Although artificial intelligence (AI) is regarded as a key technology of the 21st century and offers companies a wide range of potential, SMEs in particular are reluctant to adapt AI in their business processes despite their position in the economy. In this context, opportunities and challenges have already been explored, but little research has been done on how information offerings can support SMEs to adopt AI applications. The purpose of this paper is to explore the status quo regarding the opportunities and challenges of AI adoption in SMEs, as well as related information needs in order to derive recommendations for action to create support offerings. For this purpose, after a theory-based derivation of the opportunities and challenges associated with the introduction of AI, an online survey was conducted with N = 40 participants. The results show that SMEs have a high demand for conveying understanding and competencies regarding AI and its value added in order to overcome challenges in the introduction process of AI and to create the prerequisites for AI implementation.

Background Exoskeletons are utilized in rehabilitation and industry, impacting body mechanics and necessitating adaptations due to their weight. An examination of the gait cycle scrutinizes these procedures. As gender differences frequently result in gait abnormalities, males and females are also evaluated individually.
Methods Ten healthy individuals, comprising five females and five males, were observed in a study while walking on a treadmill at their preferred pace. The participants wore weight cuffs on their thigh and lower leg, weighing a total of nine kilograms, and their movements were recorded using marker-based analysis. The study analyzed joint angle trajectories in the sagittal plane and kinematic parameters.
Results and Discussion Weight cuffs had significant effects on the hip, knee, and ankle joints, possibly due to higher inertial forces in the hip and knee joints. Gender differences may also be attributed to the added weight, with women potentially having a higher moment of inertia in the hip joint due to their anatomy, resulting in minimal differences in joint angle trajectories.

The goal of this master thesis was to investigate the ability of the Kanban method to adapt to batch production environments with complex parameters, such as high set-up times for series changeovers. After pointing out a number of characteristics that are unfavorable to the standard method, such as high set-up times for series changeovers, large batch sizes, unstable demand, unreliable information about inventory, high variability of unit processing times, high mix of products, many suppliers, complex flow organization, and long distances between stations in the factory, a wide range of 43 variants that meet some of these requirements was collected. The Weighted Sum Model was proposed as a tool to highlight relevant variants in the context of a search for adaptations based on several criteria of various importance. A case study of Michelin's Karlsruhe plant was used to demonstrate the adaptability of the Kanban method to these specificities by implementing and testing a new variant for this production environment, characterized by high series changeover times, a wide product mix at one workstation, constraints on the minimum product aging that must be respected, and a complex flow. The approach proposed in the case study can be used as a starting point for other companies with similar characteristics to find a variant suitable for their needs. This master thesis has highlighted the ability of the Kanban method to adapt to batch size environments with complex parameters, paving the way for future research to propose other approaches to adapt the Kanban method.

The current ageing of the population (particularly in Europe) poses real challenges for companies in terms of the sustainability of know-how and the training of future generations in working processes. This Masterarbeit takes part in the KI_eeper research project, which aims to develop an AI to assist companies in the identification, storage and transfer of knowledge and know-how. In this project, the following report focuses on the modification of work environments during the implementation of this artificial intelligence and the ethical problems that this entails for employees. It will examine the contribution of an ELSI workshop to the involvement and acceptance of changes in the working environment by the employee concerned.

In their decision-making humans are assisted increasingly by artificial intelligent systems. Such systems derive references by means of machine learning experiences. However, designs of this type are often a matter of black boxes. For ethical and judicial reasons, the traceability of references is mandatory. The concept of XAI attempts to provide comprehensible explanations of how such systems act. Unfortunately, the ideal representation of explanation is underexplored.

The focus of this thesis is to investigate the design of XAI-declarations in the level of user interface. Furthermore, the purpose of this thesis is to create a guideline by means of the gained knowledge. This guideline contains significant indications to create a user-centered XAI-Design frame. Therefore, the research questions are the following: What are the gaps and challenges in the interface of human and AI? How can those gaps be filled? How do visualizations affect the human decision-making? To what extend does XAI assist the medical decision-making? How is it possible to arrange frameworks in categories?

The answers to these research questions provide essential resources to approach the concept of XAI.

The aim of this work is to develop a roadmap for the digitization of work processes in small and medium-sized enterprises (SMEs). To this end, the special features of SMEs in the context of digitization are first translated into requirements and framework conditions for a suitable roadmap. The requirements and framework conditions can be divided into the three areas of SME suitability, adaptability to the individual company and overcoming typical obstacles to digital transformation.

An overview of existing roadmaps for digitization is given and the extent to which they meet the requirements and framework conditions is analyzed. Based on this, a roadmap is developed that is suitable for the digitization of work processes in SMEs. The developed roadmap is applied to a practical example.

In the context of the project „Fahrerkabine 4.0“, an adaptive human-machine interface in agriculture, this Master's thesis examines whether a mental workload measurement can be realised by means of speech analysis. This is less invasive, easier to implement and less expensive than the alternatives. For this purpose, various speech parameters were extracted by the program „Audacity“ from recordings of an experiment already conducted at IFAB (Funk, Haase, Remmers J., & Deml B., 2021). With the help of „Matalb“ those were used to train various machine learning algorithms. These are intended to make an assertion about the current mental workload of the test person.

The tests deliver accuracies up to 76%. Thus, it is generally possible to recognise mental workload on the basis of various speech parameters. Nevertheless, there is room for improvement to increase the reliability of the algorithms.

The purpose of this thesis is to provide a better understanding of the assessment of behaviors of drivers in merging situations on highways. Both perspectives of the merging process were considered, the merging of the driver himself and the merging of another road user. Three behaviors were considered: accelerating, decelerating and maintaining speed. These were compared in two evaluation dimensions, the perceived cooperativeness and the criticality of the resulting situation.

The second scope of the study was to investigate the influence of marking vehicles as automated driving vehicles (eHMI) on the assessment of their behavior and behavior towards them. To answer these questions, a driving simulator experiment with an additional questionnaire was evaluated. Obtained results were that road users assess the behavior "deceleration" in both perspectives as more cooperative and less critical than "acceleration".

Regarding the marking of automated vehicles, this study showed that in case of the driver merging himself, a significant difference of the assessment of the criticality was reported. On average, automated vehicles were rated to be more cooperative and less critical in all driving situations. An evaluation of the observed time headway in the scenario of the merging of another road user onto the highway showed it to be significantly larger if the vehicle is marked as automated.

The aim of this thesis was to examine whether cobots or exoskeletons help to exploit open ergonomics potentials at manual workplaces. First, various assembly processes were examined in more detail and a concrete handling process was selected as a case study. A literature analysis was used to collect aspects on the suitability of the assistance systems. Subsequently, practically relevant aspects about the assistance systems were collected with the help of expert interviews. The data from the expert interviews were analyzed according to Mayring.

With regard to the use of one of the assistance systems to improve ergonomics at the selected workplace, it should be noted that the experts and the literature generally attribute a positive effect on ergonomics to the systems. However, taking into account other factors such as productivity, process complexity and costs, but also technical restrictions, it can be seen that none of the assistance systems "exoskeleton" or "cobot" would offer holistic added value at this workplace and the experts would therefore not use any of the systems without further ado.

Object handovers are commonplace for humans and occur intuitively. In robotics, attempts are being made to achieve safe, reliable and fluid object handovers between robots and humans. Robots are already capable of passing objects and adapting to the movements of humans in real time. In the course of this work, it will be investigated whether a high degree of adaptivity of the robot during the transport phase leads to a better handover than a low degree.

In order to compare both transport methods, it will be investigated how they affect the interaction time, as a criterion for the performance and fluidity of the handover. It could be shown that handovers with a low degree of adaptivity in the transport phase lead to a lower interaction time than handovers with a high degree of adaptivity in the transport phase. Furthermore, it could be shown that the interaction time in both methods improves over the course of the trials.

In the world of work the human resource is becoming increasingly short due to demographic change and the prevailing lack of skilled workers. To protect workers from damage and employ them for a long time, companies should try to design their working environment in such a way that the tasks to be performed won’t harm their employees. In the context of the present bachelor thesis a manual assembly workstation was considered and optimized. The aim was to minimize the mental and psychological workload by appropriate measures.

In a first analysis the test subjects’ stress was determined using questionnaires and motion sensors attached to their bodies. After that the workstation was optimized and the stress was measured again in a second run. On that, potential effects of the appropriate measures could be analyzed and discussed at the end of this bachelor thesis. The results show that the physical workload has been shifted to other body regions, but wasn’t eliminated completely. The test subjects felt the optimized workplace to be less stressful, both mentally and physically.

Automated vehicles are continuing to develop. This will result in mixed traffic on the roads in the future. Therefore, good communication between automated and non-automated vehicles is necessary to ensure a smooth traffic flow. In this paper, three aspects of communication and their influence on the evaluation of a driving situation at a motorway slip road are considered. The duration, the position of the lane change, and the type of vehicle (automated, non-automated) are varied. For each scenario, video recordings were made from the first-person perspective of the vehicle in the right lane. Data collection took place via an online video survey. Participants rated the situations according to how dangerous and how cooperative they found them.

The survey showed that a slow lane change is rated as less dangerous and more cooperative. There is a non-linear correlation with the position of the change. A change in the middle of the acceleration lane is seen as the most cooperative and the least dangerous. No clear statements could be made in this work on the differences between the situations in the evaluation of automated and non-automated vehicles.

A real-time evaluation of mental stress could significantly change industry and economy in the future. The monitoring of heart rate, as a cardiovascular indicator of mental workload, could represent a possible practical implementation. In this context, three methods of heart rate measurement were evaluated and compared with each other. Therefore, an attempt was made to detect differences in mental stress simultaneously by varying different tasks and task difficulty using three different measuring methods. The aim of this work is to check whether the selected methods ECG, Fitbit and Remote Photoplethysmography image recognition algorithms record the courses of the heart rate equally.

Results of the investigations suggest that there are significant differences between the recorded heart rates of the applied measurement methods. This work can serve as a basis for testing whether and to what extent these methods can be used to determine the degree of mental stress.

Individualised products and shorter product life cycles increase the demands in manual assembly. These demands can be experienced as negative stress. In the course of Industry 4.0, various visual assistance systems are being used to support assembly workers in their work through the targeted display of information and to promote their cognitive ergonomics. Human acceptance is crucial for the successful use of technologies.

In this study, the effect of four visual assistance systems on people's acceptance and subjectively perceived stress is investigated in a laboratory experiment. The results showed that the screen was significantly less accepted than the projection, the light signal and the combination of screen and light signal. A difference in their effect on the experienced stress and mental strain of the subjects, as well as on the number of errors, could not be proven. No correlation could be found between acceptance and stress. The projection and the light signal showed great potential and could be used helpfully.

This paper studies the impact of large vehicles on the driving behavior of drivers entering and in the acceleration lane. To this end, this paper uses drones to take aerial videos of the entrance of the B10 highway in the south of Karlsruhe, and then process the collected videos to obtain relevant data sets. In order to study the effect of large vehicles on driver driving behavior patterns, three hypotheses are proposed in this paper: 1) different main road vehicles flow and large vehicle mixing rates; 2) different types of main road intersecting vehicles; and 3) relative positions of main road vehicles and on-ramp vehicles on driver driving behavior in the acceleration lane. Then use significance test analysis to analyze the data set. The results show that the main road traffic flow and the mixing rate of large vehicles have significant effects on the average speed and average acceleration of drivers in the acceleration lane. The type of vehicles on the main road in front has a significant effect on the speed and acceleration of drivers of small vehicles entering the acceleration lane. The relative position of the vehicles on the main road and the vehicles on the ramp has a significant effect on the average speed of drivers in the acceleration lane.

With a view to autonomous driving and consequently towards mixed traffic, the aim of this study was to asses the effect of non-involved pedestrians on cooperation behavior of drivers on T-intersections in patt-situations.

Video clips were used to examine the driver´s behavior and behavioral expectations as well as difficulty and the clarity of the situation in an online study. The number of pedestrians, the sidewalk zone, the pedestrian barriers and the approach-direction was varied. The cooperative behavior was examined by the drivers rating on their own and the expected behaviour of the other drivers.

The study shows that only the approach-direction influences the behavior and the behavior expectation. At high numbers of pedestrian, the difficulty of the situation increased, and the clarity decreased. The interaction between pedestrian numbers and barriers shows that at high pedestrian numbers and a barrier the clarity of the situation increased, unexpectedly.

Unlike other methods of stress measurement, determining mental workload through speech is non-invasive, less intrusive, and less expensive. In this work, an experimental environment is adapted from an existing experiment to induce mental workload through two visual tasks (surveillance task and mental rotation) and one secondary auditive tasks. A study of performance indicators such as reaction time and error rate during the tasks, as well as a study of speech features, should show a correlation with the increase in the subjects' mental workload. During the study (N = 21), the subjects' responses are recorded with a microphone. Praat software is used to analyse jitter variations, and OpenSmile audio feature extraction software is used to build a database to construct a multi-layer perceptron using Weka software. The experiment shows that the reaction time and the error rate of the subjects during the auditory tasks are significant indicators of the increase of the mental workload; the jitter decreases, but not significantly; finally, the model developed by Deep Learning method is 67.7% accurate in recognizing the mental workload state.

Understanding Mental Workload in human-machine interface could bring tremendous benefits in Ergonomics. Meanwhile, the rising popularity of wearables introduces them as possible instrumentation for assessing mental effort. This study compares Fitbit Charge 3 and Garmin Vivosmart HR+, and questions their accuracy and potential as new instruments for mental workload.

To test the hypothesis that these watches can replace the standard electrocardiogram, an experiment was conducted, measuring heart rate with these three instruments, while volunteers were performing mental demanding tasks. The analysis was divided into two parts: first, to determine the most accurate watch using agreement comparison methods; second, to determine their potential to detect changes in mental load. As results, Fitbit was better than Garmin, but could not replace an electrocardiogram.

Findings indicate that both watches were not suitable for identifying mental workload. Still, wearable devices have potential for identifying high mental effort and assisting humans in increasing cognitive demands.

The introduction of automatic vehicle guidance primarily leads to mixed traffic between automatic and human guided vehicles. In road traffic, it can happen at various points that the right of way is not clearly regulated and drivers have to cooperate with each other. One such situation is a bottleneck of equal priority where two road users arrive at the same time. The automatic vehicle guidance must now cooperate with the human driver.

In addition to the two cooperating road users, other vehicles may appear in the situation at different locations. In an online survey, video sequences of same-ranking bottlenecks with different traffic were shown and analysed with regard to perceived difficulty, assessed accident risk, driving order, assessed behaviour and safety via this assessment. Not only the number of vehicles but also the location plays a role, as can be seen from the results on perceived difficulty. In general, people tend to drive defensively, but the vehicle in front tends to lead to an offensive driving style.

The aim of this bachelor thesis was to present the quality criteria of psychological procedures and various text evaluation models. Based on this knowledge, 195 tickets were to be evaluated by means of a selected text evaluation model and subsequently a calculation of the interrater reliability was to be carried out. The tickets are part of a preliminary study on the cognitive ergonomics of mobile work in public transport and are used in the context of a routine task.

By calculating the interrater reliability, it was checked whether the tickets met the quality criteria of psychological work or whether individual tickets had to be changed. Based on the calculations of the interrater reliability, it could be determined that the tickets met the quality criteria of psychological procedures and can be used in unchanged form.

In this article, we proposed an optimization method for eye-tracking. We designed a test procedure that enables the optimization of an eye-tracking system within the distance range from 20 cm to 100 cm and the angle range within 30° for both vertical and horizontal axes. This method can reduce the error between the real gaze point and the estimated point from the video of this kind of eye-tracking system. With the help of an open-source RCNN artificial neural network, the error can be detected by our program automatically, and the error data can be transferred and analyzed to build an error model with the help of space interpolation. This error model can be applied to the analyzed eye-tracking system to eliminate or reduce the error and lead the estimated point iteratively to the real gaze point. This method focuses only on the output data of the eye-tracking system instead of the internal features of the system, which makes it universally applicable to different eye-tracking systems with the same form of output data. Experiments show that this method is feasible and effective in decreasing the gaze point error by about 36.64%.

Im Rahmen der vorliegenden Arbeit wurde eine Studie durchgeführt, die Einflussfaktoren auf das menschliche Verhalten an einer T-Kreuzung im innerstädtischen Gebiet untersucht. In der untersuchten Verkehrssituation ist die Vorfahrt nicht eindeutig geregelt und es bedarf einer Form der Kooperation der Verkehrsteilnehmer, um die Blockade aufzulösen.

Im Rahmen der Studie wurden Probanden Videos der Anfahrt aus verschiedenen Richtungen und unterschiedlichen Umgebungen an die Kreuzung gezeigt. Anschließend wurden sie hinsichtlich ihres eigenen, erwarteten Verhaltens und des erwarteten Verhaltens der anderen beteiligten Verkehrsteilnehmer befragt.

Es konnte gezeigt werden, dass die Anfahrtsrichtung an die T-Kreuzung einen Einfluss auf das angegebene Verhalten hat – sowohl auf das vorhergesagte eigene Verhalten als auch auf das erwartete Verhalten der anderen beiden Verkehrsteilnehmer. Für verschieden gestaltete Komplexitäten der Umgebung der T-Kreuzung kann ein Zusammenhang mit der Bewertung der Übersichtlichkeit der Situation aufgezeigt werden. Ein Einfluss der Komplexität der Umgebung auf das erwartete eigene Verhalten und Anderer kann nicht gezeigt werden.

In the context of the presented thesis, a study was carried out to investigate factors influencing human behavior at a T-junction in an inner-city area. Within the investigated traffic situation the right of way is not clearly regulated and some form of cooperation among the road users is needed to resolve the blockade. In the course of the study, subjects were shown videos of the approach to the intersection from different directions and various environments. Thereafter they were surveyed regarding their personal expected behavior and the expected behavior of the other road users involved.

It can be shown that the direction of the approach to the T-junction has an influence on the indicated behavior – both on the predicted personal behavior and on the expected behavior of the other two road users. A dependency between differently designed complexities of the surrounding area of the T-junction and the valuation of the perceived visual clutter of the situation can be shown. An influence of the complexity of the surrounding area on the expected personal behavior and that of others cannot be shown.

The recent increasing demand of the dismantling processes left the disassembly operators suffering from loads affecting their musculoskeletal system due to the limited ergonomic interventions to the disassembly workstations. As a result, the purpose ofthis paper is to develop an ergonomic clamping system to support the disassemblersin dismantling the use case of an electric motor on a stress free workbench. The sub-handling of the disassembly task is analysed and evaluated according to the objective ergonomic DIN standards. The evaluated downsides are tackled using the physiological characteristics of ergonomics that included the suggested workplace enhancements as well as recommended handling to disassemble the motor. The broad analysis of the disassembly of the electric motor defined the features to be acquired by the clamping system that aided in minimising the loads acting on the body muscles. Then, a digital CAD-Model of the concept is created and further validation processes are suggested to ensure the concept’s assembly and functionality success.

Full automation of assembly processes is rarely economically feasible in small and medium-sized manufacturing companies (SMEs). However, partial automation, in which humans and automation work together, could increase productivity, reduce the workload on employees, and offer a high degree of flexibility at a low investment cost. Therefore, this master thesis investigates the changes in productivity and human stress caused by a reconfigurable robotic system.

To this end, an experiment with an industrial assembly task was carried out. This experiment is performed by subjects both manually and hybrid (in interaction with the robotic system). The assembly time, the stress and strain on the subject as well as his satisfaction with the task were analysed. It was found that the use of a robotic system reduces the assembly time and increases the employee's satisfaction. The use of such a hybrid system shows economic potential and rapid acceptance by the employees, which makes its use more attractive for SMEs.

Mobile working at customers' premises, in hotels or on the road has been gaining in importance among employees and employers in recent years. With modern communication technology, knowledge work in particular is no longer tied to fixed locations. However, the ergonomic properties of the working environment for mobile work are unknown in many cases. It is therefore not possible to derive any generally valid statements on the risk assessment of mental stress. The influence of non-work-related environments on performance is also not fully understood. The Institute of Human Factors and Industrial Organization (ifab) at the Karlsruhe Institute of Technology (KIT) is planning a field study to compare physical and mental stress and the resulting mental strain during knowledge work in the office and on the train. In addition, the influence of the working environments of the experimental conditions on the performance achieved will be investigated. The aim of the study is to identify possible hazards in the context of mobile work in public transportation and to derive possible measures for the health protection of employees. For the planned study, a detailed experimental concept is developed and presented in this master's thesis. This includes the underlying theoretical background, a concrete experimental setup and procedure, the development of a routine task and the programming of a ticket system for this. Furthermore, a creative task is theoretically derived in four subtasks and elaborated on the basis of tasks and evaluation guidelines. Finally, this work includes a structured interview guide for recording workload, perceived difficulty, blockages and disruptions as well as the planned evaluation of all data collected, including research questions and statistical hypotheses. The designed experimental concept is complete and can be carried out in the manner described in this master's thesis.

To drive a vehicle safely, a driver must be able to recognize emerging risks in a situation at an early stage. If the driver reacts too slowly or not at all to risks, this may be caused by an inadequate situational awareness. Based on the situation awareness model of Endsley (2013), the present work will investigate by means of an experiment how situation awareness changes qualitatively as well as quantitatively with increasing degree of experience. All three levels of situational awareness (perception, understanding, anticipation) will be considered.

For this purpose, an experiment was developed in the form of an online study, which is based on the study of Sommer (2012) and is continued with the widely used and standardized measuring method for situation awareness (SAGAT) developed by Endsley (2013). The work culminates in design recommendations for a future assistance system, with the aim to support the development of adequate situation awareness in the best possible way, taking the individual driving experience into consideration.

The aim of this work is to develop a catalogue of questions to compare production jobs in terms of their potential for human-robot-collaboration. This questionnaire serves as a stand-alone tool for the initial evaluation of an existing workplace in SMEs. Thereby, the challenges arising from the future social structure such as the ageing of individuals and with thus the longer period of employment are presented. This is followed by a chapter outlaying the burdens and diseases of workers in the workplace. Following the linking of the consequences of the characteristics and challenges within SMEs in the transformation of digitalization are considered.

Before analyzing comparing assessment tools and evaluation catalogues in Chapter 2.6, an overview of the use of robots in industry, their producers and the technical characteristics of some representatives of CoBots are given. A precise analysis of similar questionnaires resulted in the fact that there are fundamental differences in realization and key facts. As a result of the prior stated criteria a new CoBot questionnaire is outlined in Chapter 3.3. This questionnaire is then carried out with two SMEs and finally evaluated by means of an expert survey.

The outcome of this procedure is the request for an expansion of the questionnaire concerning the potential of full automation. Furthermore, it is asked for some changes in wording. Both SMEs primarily share the motivation of reducing labour costs in the field of assembly. Consequently, a re-analysis of the questionnaire regarding criterions of full automation is due. Besides the overall results, another questionnaire for Industry 4.0 implementation of existing production facilities could be developed from the basic structure of the ProBot catalogue.

This bachelor thesis deals with the optimization of a production workplace relating to ergonomic and organizational aspects. In many companies, these are insufficiently taken into account when designing workplaces. The aim of this bachelor thesis is to identify optimization potentials for the considered workplace and to implement suitable measures.

To achieve this goal, a comprehensive analysis of the workplace and the production process is carried out. In advance, the methods and evaluation tools required are described in detail. The analysis shows optimization potentials with regard to the employees’ posture as well as improvements in the organization at the workplace and of the production process. After categorization and prioritization, concepts are developed and optimization measures are implemented.

The evaluation of the implemented measures shows that especially the adjustment of the working height to the employees’ needs achieves a significant improvement in posture. By involving the employees in the optimization process, improvements in processes and organization are implemented, making work easier. Within the scope of future projects, the results of this bachelor thesis can be transferred to other workplaces and contribute to a holistic optimization.

In this thesis, the behavior of motorists at an unregulated T-junction is investigated. In particular, the influence of direction, traffic, and vehicles in front of and behind the cooperation vehicle are examined. These influencing factors are investigated by means of an online test subject study. The participants were shown videos depicting various scenarios and subsequently asked about their behavior and perceived difficulty of the situation.

The results of the study indicate that drivers on the through road behave more offensively than drivers at other positions. If there is traffic on the road before the driver reaches the junction, they show more defensive behavior. The vehicles in front and behind the cooperation vehicle do not influence the behavior of the driver. Only the vehicle driving in front of the cooperation vehicle was found to have an influence on the perceived difficulty. This situation was rated with a higher difficulty.

The results of the study are able to make an important contribution to the integration of autonomous vehicles into road traffic, thereby enabling mixed traffic.

The use of strategies for fault diagnosis was investigated using the computer-aided process simulation AWASim. Both process parameters, i.e. solution time and accuracy of the fault diagnosis, as well as recorded eye movements served as the basis for the analysis. The collected data were evaluated with regard to two questions. First, it was investigated whether a systematic approach to fault diagnosis would increase efficiency. Second, changes in the use of strategy and the use of information over the duration of the experiment were examined.

Neither the application of a topographic search routine nor the application of a half-split heuristic proved to be useful in fault diagnosis. Furthermore, it could be observed that flow path independent information throughout the duration of the experiment was increasingly considered. However, this did not result in any increase in efficiency in the fault diagnosis.

In recent years there has been a trend towards project work. Companies are faced with the challenge of integrating the knowledge gained from the projects into the organisational knowledge base. This thesis deals with how knowledge management can be promoted in the field of project management. Based on a literature search, surveys are designed in the form of expert interviews and workshops and carried out at a partner company. The aim of the surveys is to derive influencing factors on the exchange of knowledge and to identify a field of action. Based on the recommendations from the research, a methodology will be developed to promote the exchange of knowledge related to project management. This will be tested and evaluated in the partner company. The findings will be used to draw conclusions about the practical application of the methodology.

The manual execution of physical work can lead to sanitary consequences. Therefore, this paper deals with the question, which ergonomic potentials collaborating human-robot-solutions can show. The investigations of ergonomic potentials are made based on a prototypical workstation, where an implementation of a collaborating robot takes place. In addition, cooperating robots are introduced, analyzed and compared with the collaborating solution. These investigations are made with systematically selected ergonomics assessment methods and the ergonomics software Jack – Human Modelling and Simulation Tool – (Siemens PLM). Furthermore, this paper allows to have a look into the psychological impact of robots on humans.

The results of this paper demonstrate that collaborating human-robot-solutions have the potential to improve working conditions ergonomically, in the context of physical stress and fatigue. However, cooperating human-robot-solutions show a higher ergonomic potential compared to collaborating solutions.

To predict thermal perception of people, several classical models are available. However, they don’t focus on individual thermal perception. Recent research tends to use vital parameters to build models that are able to predict personal thermal perception. This thesis investigates the performance of machine learning (ML) models for prediction of personal thermal perception, whose inputs are integration of vital parameters and classical models’ inputs. Involved vital parameters are skin temperatures and heart activity metrics.

The ML models are developed and evaluated based on two available datasets, which were collected under cold and near-neutral thermal conditions, respectively. Results show that the ML models can achieve more than twice the accuracy of classical models on the dataset from cold thermal conditions, whilst little accuracy improvement is observed on the dataset from near-neutral thermal conditions.

Observational learning is the basis for human motor skills. But that is not all. Eye movements are affected by observational learning as well and give information about cognitive processes according to the eye-mind-hypothesis. Novices can learn the strategy of experts by watching their eye movements while they are solving a problem. Therefore, the design of visualisation of eye movements in Tower of London exercises is picked as a central theme of the bachelor’s thesis.

In an experiment with forty test persons, the usability of four different visualisations of eye movements is investigated. The four visualisations differ in mental load triggered by themselves, their usability, their experienced emotions and their effects on memory and performance, whereas the speed of the visualisation has a low influence on the usability criterions. In the interpretation and discussion of the results, some implications for good visualisations of eye movements are worked out.

Visual complexity is a key concept in many domains, especially the web design where it is relevant in order to create user-friendly, aesthetically pleasing and workload-optimized websites. This study aims to achieve a better understanding for visual complexity of websites by analyzing influencing factors of the construct as well as its relations to the concept of aesthetic.

By creating 60 schematic websites patterns, it was possible to investigate the subjective perception of a website’s visual complexity as well as the number of fixations of 31 subjects. Four factors were selected according to literature research, which are expected to influence visual complexity in the case of websites: quantity of elements, variety of colors, symmetry, and prototypicality.

Results show that all influencing factors had a significant effect on the rating of visual complexity, even though variety of colors has a much weaker effect than the three others. Moreover, visual complexity on websites has shown to be negatively linked to aesthetic. Finally, the number of fixations increased with a higher quantity of elements, and presence of prototypicality, but variety of colors and symmetry have no significant effect in our study.

In the future, the automated systems will take on all the tasks of humans. Human will act as a monitor. In order to detect the loss of competence after long-term monitoring, the relationship between experience and corresponding psychophysiological data is very important. Based on the EsA model (Heine, 2017), the present work attempts to find the connection between experience level and somatic marker and to validate the EsA model.

An experiment with two measurement times is performed. Twenty-one tasks were performed at each time point and electrodermal activity data were measured by probands. The 21 tasks were divided into three parts. The results showed that the SCR averages of the three parts always increased at the first measurement point. The SCR averages of the three parts at the second measurement time rose first and then lowered. There was no significant change between the SCR averages of the corresponding parts.

In times of global competition, the needs of customers are the driving force behind every product development. With the constant improvement of the consumer market, the customers have more and more demands on the comfort of the use of the car. Mercedes-Benz has been researching new car storage devices for many years. However, a perfect product has not been produced in series on the market.

Based on the guiding ideology of the human-centered / user-oriented and the several earlier student works of Daimler, AG, the Mercedes-Benz S-Class is used in this work as a carrier to develop a storage medium behind the front seats in the car. The device must be ergonomically designed to meet customer requirements for usability and safety in use, provided the storage space is large enough and the design is aesthetically pleasing.