Vol.1/2007

Empirical analysis of information input for mobile human-computerinteraction

Support possibilities of the product development: Which changes in the result are caused by question-based thinking about own solutions?

Benefits of 3D process visualization on fault detection in process plant operation

Development and validation of scales assessing different job-related self-control demands

 

Vol.2/2007

Dual-task performance while driving a car:
Age-related differences in critical situations

 
Perceived strain and well-being in different types of offices
 
Evaluation of instruments to evaluate job satisfaction
 
Learning Cultures and Different Perceptions of Learning-Conducive
Working Conditions in Five Businesses – Research Results

 

Vol.3/2007

Development of Work Physiology and Ergonomics in Germany —Tasks and Objectives in Education and Research

Room acoustics for open-plan offices

The influence of milking parlour equipment on the workload of the milker

The effect of predictability of system response times on the user’s behavior and state

Promotion of Changing Processes by Efficient Team Management in a Company Health Insurance Scheme
 

Vol.4/2007

Continuous head gestures – an alternative way of controlling menu based in-vehicle information systems?
 
Binocular coordination of eye movements in viewing conditions of head-up displays in road vehicles
 
Evaluation of different mobile map displays for geographic orientation tasks
 
Evaluation of driver drowsiness by video analysis
 
 

Vol. 1 - 2007

Empirical analysis of information input for mobile human-computerinteraction (Vol. 1 – 2007)

Author: Thomas Alexander, Dieter Leyk und Christopher Schlick

Keywords:    Mobile computing · Personal digital assistant (PDA) · input performance · Fitts´law

Summary

Mobility and mobile computing are topics of growing importance within today’s and future working processes. In general, it has to be differentiated between portable and mobile computing. Portable computing refers to stationary work at different locations while mobile computing is defined by working on the move. Both have different characteristics which have to be considered when designing the human-computerinterface. This paper refers to mobile computing while walking. Frequent applications of mobile computing are entertainment, information input and retrieval, communication, and orientation. Common devices for mobile computing are notebooks, tablet- PCs, cellular phones, and personal digital assistants (PDA). For practical weight reasons only the last two are applicable on the move. In addition to simple keypad interaction of cellular phones, touchscreens with special input pens are used as input devices when more complex inputs are required. The according graphic user interface (GUI) consists of a modified standard desktop GUI with windows, icons, menus, and pointer (WIMP). The user interacts nearly exclusively by means of point-and-click. Additionally, different text recognition modules or virtual, miniaturized QWERT-keyboards are used for text input. Characteristics of mobility or movement are usually not taken into consideration for HCI design. Instead, concepts of stationary desktop computers are simply adapted. Previous studies show that this leads to shortcomings in performance. One reason for this is that interaction while walking is obviously a dual task with walking as primary task. Walking requires attention resources, which leads to degraded HCI-performance. Previous studies have also shown that mobile use leads to a different use, performance, and workload. Thus, mobile HCI should be studied during actual movement. But especially with walking this is true for few studies only. In this study it was hypothesized that the movement state effects input performance. In order to quantify the effect, measures of performance were taken at different walking speeds on a treadmill. Input performance was determined by error rate and Fitts’ index of performance. They were measured with a target designation task on a PDA while walking on a treadmill (speed: 0, 2, 3.5, 5 km/h). Experienced participants (n=18) were asked to hit targets of different sizes (2.4 mm to 12.0 mm) with changing positions on the display. The according Fitt’s index of difficulty (ID) was varied from 2.0 to 5.6 with steps of 0.2. Results show significant effects of ID and walking speed on error rate (p<0.01). The ubsequent pairwise analysis revealed only significant differences between error rates for two groups: (ID<4) and (IDe“4). Consequently, the statistical analysis was repeated for each of them. For movements with ID<4, ID and walking speed explained similar proportions of variance (o²ID=31.4%; o²v=34.8%). However, a subsequent regression analysis showed that the values for error rate increased only slightly from 0% to 3.6%. For more difficult movements (ID e“ 4), both factors explain comparable proportions of variance (o²ID=46.6%; o²v=39.6%). The final regression analysis showed a strong increase from an average error rate of 2,8% (ID=4, standing) to 22,2% (ID=5.6, v=5 km/h). Thus, the effect is much stronger for these visually-controlled movements. With regard to Fitts’ index of performance an ANOVA proved significant effects of walking speed on this variable (o²=23.9%). The index of performance decreased from 8.5 bit/s (0 km/h) to 5.8 bit/s (5 km/h). Notice, that the effect of ID is included in the index of performance by definition. As a consequence, Fitts’ law was extended by an additional, speed-dependent term. The regressional analysis revealed a quantitative description of the relation between input time, ID, and walking speed. This study shows that walking with different speeds on a treadmill has an effect on input performance. Consequently, movement status and speed have to be taken into consideration when designing HCI for mobile computing. This is especially important for the design and size of icons and menu items of a GUI metaphor. Moreover, an error-tolerant behavior has to be implemented when critical input is required. In this study it was also found that error rate increases strongly even though precision was prioritized by the instruction. In reality, error rate might be even higher because of a missing prioritization and higher environmental disturbance. Consequently, movements with IDe“4 should be avoided. However, it is also found out that stopping is a possible way to enhance input performance. This meet reality, because most mobile users take a break when high precision input into the mobile computer is required.

Practical Relevance

The importance of mobility and the resulting technical requirements for mobile computing is rapidly growing. But present Human-Computer-Interfaces for these mobile systems are often designed with regard to stationary desktop systems only and do not consider the special context of a mobile usage. Consequently, they cannot be used when the user is walking or moving. For a sensible mobile use, reference values for estimating expected input performance and precision are required.


Support possibilities of the product development: Which changes in the result are caused by question-based thinking about own solutions? (Vol. 1 – 2007)

Author: Constance Winkelmann und Winfried Hacker

Keywords: Design Problem Solving · Question Answering Technique   Reflection

Summary

Design-problem solving is conceptualized as an iterative process of self-evaluation and correction of errors (Ullman, Dietterich & Stauffer 1988). Question-answering techniques evaluating (the) obtained own results may improve the quality of solutions of design (Wetzstein 2004; Winkelmann 2005). However, the degree of improvements is significant, but small and only about half of the participants improve their solutions at all. Therefore, this study analyses whether the types of the question-induced modifications of own results of design problem solving may hint on explanations for these limited question-induced improvements. Three groups of subjects (N = 116; laypersons, students of engineering design and experienced engineering designers) were asked to design a sophisticated version of a grill. The manual sketches were evaluated by experts before and after the intervention by a generic system of interrogative questions aiming at the semantic relationships (e.g. causal, conditional, final). The three groups of participants achieved significant mean improvements of solution quality with small to medium effect size. 82% of the participants modified their solutions at all in spite of a large window of opportunity. About 60% of all participants carried out modifications but did not really improve the solutions. 40% of the modifications in principle led to solution improvements; however, half of the participants only added some more or less relevant aspects. The other half consisted in equal parts of error corrections, continuations of the design principle begun, and of applications of a new principle („productive modifications“). Laypersons, students and engineering designers do not differ significantly as to these results. Comparing the group with improvements to the group showing none reveals that corrections, continuations of a solution begun, and applications of a new design principle occur three times more frequent in the former. This may not be taken for granted, though, since the correction of one error, for example, may produce another error. The results hint on restrictions on the application of questionanswering techniques so far: the participants neither had an opportunity to get access to additional knowledge nor to discuss their solutions with experts. Thus, the efficiency of the questions may be underestimated. Moreover, reflection by question-answering should become a component of engineering education.

Practical Relevance

The analysis of the changes of design solutions due to question-based reconsidering of these solution shows that especially „productive“ changes (error corrections, enhancements of the solution, development of a new solution principle) alone and in combination with additions lead to improvements of these solutions. In principle, it might be practical to use the question-answering technique as an aid for the discourse with other specialists when deliberating on (interim) solutions. Beyond that, we recommend to integrate this question system into the training of advanced students as they particularly can profit from the released error corrections.


Benefits of 3D process visualization on fault detection in process plant operation (Vol. 1 – 2007)

Author: Birgit Vogel-Heuser, Karin Schweizer, Antje van Burgeler, Yvonne Fuchs1 und Dorothea Pantförder

Keywords: ·  3D visualization · fault · diagnosis · HMI · process plant operation · work-load

Summary

Nowadays process plant visualization and operation take place without the operator’s physical presence at the technical device (plant, machinery). As a consequence a lot of complex systems must be visualized simultaneously on one or more monitor, maybe with several windows to be selected out of a menu. The monitoring of those processes with conventional 2D human-machine-interfaces requires high affordance in terms of operator’s attention and increases work load. In this article an approach is introduced which aims at the reduction of work load and complexity and by that faults using 3D visualization. The process to be observed is a distance control of a continuous hydraulic press, i.e. a field of analogue sensors arranged in a spatial setting. The approach is evaluated in a lab experiment.
Hypotheses: In an experimental study with 75 male and female participants (aged from 18 to 46) we investigated three hypotheses: (i) 3D visualizations as surface plots reduce the amount of faults and the time to detect faults compared to 2D visualizations when monitoring a continuous hydraulic press (ii) 3D visualizations as surface plots increase the presence compared to 2D measured through a presence questionnaire. (iii) There is a relation between the increased presence and the reduction of faults and time to detect faults.
Results: To investigate these hypotheses we simulated several different problems referring to common problems with a continuous hydraulic press as inclination of the press and torsion of the press table in different views and dimensionalities (Details are given in Fuchs 2005; van Burgeler 2005). We analyzed the number of hits, the reaction time latencies for detecting faults and the presence score measured by the questionnaire and received the following results. In contrast to hypothesis (i) participants operating with 2D visualizations have reacted as fast as participants operating with 3D visualizations. There is, however, a significant interaction with the test conditions/versions which is due to the different versions of the 3D visualization. When the problem develops and diminishes in front of the operator there is an advantage in terms of reaction time and number of faults for those participants compared to another 3D-group (the problem is far above and approaches to the operator) and compared to the 2D visualization. Our hypothesis is also confirmed by a second finding: for one special problem (the most difficult of all problems) participants of both 3D groups react faster than the participants of the 2D group.
This trend can also be stated when regarding the number of hits. Participants operating with 2D visualizations have discovered more faults than compared to the ascribed 3D-group when the problem was far above and approached to the operator.
We could also confirm our second hypothesis. There is an increase of presence in both 3D groups. We could, however, not confirm our third hypothesis. We could not find any sustainable hint that the amount of presence is related to advantages in detecting faults at all.
Discussion: We assume that monitoring in process plants can be separated in two sub-processes: one component that can be solved with conventional 2D visualizations and another component that require 3D visualizations like surface plots in order to solve more complex problems. Our findings show that in this case the visualization should start with the problem in front of the operator and should than depart subsequently.
Further research deals with the relation between attention efforts (measured through work load) and task solution.

Practical Relevance

The impact and benefit of 3D visualization in process plant operation is discussed. We assume that spatial presentation of a sensor field as surface plot is more efficient to monitor a process and to detect faults (in terms of time and number) than 2D visualizations. If faults could be detected earlier and appropriate measures could be taken, the cost reduction for operation will be tremendous, due to the avoidance of down-times.
The discussed 3D presentation has been introduced already in an industrial plant successfully.


Development and validation of scales assessing different job-related self-control demands (Vol. 1 – 2007)

Author: Barbara Neubach und Klaus-Helmut Schmidt

Keywords:  · job demands · emotion-control · burnout · self-control

Summary

Changes in the modern world of work have created new job demands. Nowadays employees are increasingly expected to be self-responsible for their work results, to react flexibly touncertainties and to behave in an achievement- and customeroriented way in order to contribute to the success of the company. All these demands require self-control insofar as the individual must behave in accordance with the companies’ interests and goals and at the same time suppress any spontaneous behaviour which deviates from the expected one. People exert self-control, for example, when they conform their behaviour to given standards, express organizationally desired emotions, or avoid distractions and overcome barriers to pursue an externally determined goal. Results of basic research suggests that the exertion of self-control is associated with psychic costs. For example, Muraven, Tice, and Baumeister (1998) demonstrated that after an act of self-control performance on a second, unrelated act of self-control was impaired. In one study, they showed that thought-control caused people to give up faster on a subsequent anagram task than those of a control group. In another study, an emotion-regulation exercise resulted in subsequent decrements on a physical endurance task. This and other findings provide strong support for the hypotheses that different acts of self-control all draw on a single limited resource, and that consecutive exertion leads to exhaustion of this resource. Research on emotion work provides initial evidence that workplace-related self-control might be a stressor. The studies currently available show that the expression of organizationally desired emotions has negative effects if emotional dissonance exists (Zapf 2002). Emotional dissonance occurs when someone’s true emotions conflict with the emotions he is required to display in a particular situation. In this case, displaying appropriate emotions requires intrapsychological processes to suppress or control spontaneously and genuinely experienced emotions. Under this point of view emotional dissonance could be interpreted as one aspect of self-control - as self-control of emotions. To investigate self-control at the workplace more directly, Neubach and Schmidt (2006) have recently developed a scale ‘required self-control’. The scale covers the suppression of spontaneous, impulsive reaction tendencies. Neubach and Schmidt (2006) showed, that required self-control and emotional dissonance reflect factorially distinct and only moderately related constructs. Significant correlations were found between ‘required self-control’ and strain. Schmidt and Neubach (2006) replicated this result with a different sample. To differentiate the measurement of self-control at the workplace, in the present study two additional facets of self-control were operationalized. According to theories of volition (Goschke 1997; Sokolowski 1996) and results of basic research (Muraven et al. 1998; Baumeister et al. 1998) we constructed the scales ‘resisting distractions’ and ‘overcoming inner obstacles’. The scale ‘required self-control’ was slightly modified to focus on ‘impulse-control’. 549 employees of an administration of a German federal state participated in the study. 15 Items were used to measure facets of self-control. The results of a principal component analysis followed by a varimax rotation showed a clear three-factor solution with simple structure. The first factor comprised six items that measured ‘impulse-control’, the second factor was composed of five items describing ‘overcoming inner obstacles’, and the third factor contained four items asking for ‘resisting distractions’. The internal consistencies of the scales were good. Cronbach’s alpha ranged from .87 to .90. Correlations with task variables and strain variables provide support for the discriminant and convergent validity of the self-control scales. Furthermore, hierarchical moderated regression analyses were conducted to assess the unique contribution of each self-control scale on strain variables. The results indicate significant main effects of ‘impulse-control’ and ‘overcoming inner obstacles’. Even when entered into the regression after social and task variables, ‘impulse-control’ and ‘overcoming inner obstacles’ showed significant beta weights. No main effect appeared for ‘resisting distractions’. This could be due to an overlap between ‘resisting distractions’ and the task variable workload. The results confirm findings from previous studies on ‘impulsecontrol’. To assess the sensitivity of the scale, scores of impulse-control was compared in three different samples (geriatric nurses, employees of a municipal administration, and employees of an administration of a German federal state). The hypothesis was that geriatric nurses display a higher level of impulse-control. This was assumed because caring for residents with dementia or depression requires extreme patience and perseverance, and because dealing with emotionally demanding situations (e. g. death and dying) requires a large amount of emotion-control. Statistical analysis revealed that, on average, impulse-control scores for geriatric nurses were significantly higher than impulse-control scores for administration employees. This result supports the ecological validity of the scale ‘impulse-control’. All in all, the results confirm that the scale impulse-control is reliable and valid. The results for ‘resisting distractions’ and ‘overcoming inner obstacles’ need to be replicated and cross-validated.

Practical Relevance

The developed scales of self-control allow specific assessments of a novel stressor from an occupational psychological perspective. Taking into account the increasing requirements for efficiency and flexibility further investigation seems warranted.

 


Vol. 2 - 2007

Dual-task performance while driving a car: Age-related differences in critical situations  (Vol. 2 – 2007)
 
Author:
Gisa Aschersleben und Jochen Müsseler
 
Keywords: dual-task situation, stimulus-response-compatibility, older drivers
 
Summary
 
Dual task situations are ubiquitous, at home as well as at work. Especially modern information technologies often require the coordination of different activities. For example, while driving a car the driver has to process additional information (e.g. from the navigation system, from the walkie-talkie or from the mobile phone) and to react in an appropriate way. This is especially the case for professional drivers like taxi drivers or lorry drivers. Under less demanding driving conditions, e.g. slow driving with low traffic density, the coordination of the primary task (driving the car) and the secondary task (manipulating the navigation system) might be relatively easy. However, under demanding driving conditions and especially in critical traffic situations, impairments in reaction times and driving errors can be expected. Dual task conditions are in particular demanding for older drivers as has been demonstrated in many experimental studies. Another important factor is the cross-task compatibility that is whether the information presented in the second task provides information, which is compatible or incompatible to the primary task. In first experiments using a driving simulation set-up we tested groups of younger (age: 18 – 30 years) and elderly (age: 56 – 65 years) drivers in their reaction times and error rates in dangerous situations (a person suddenly jumping on the street, which requires a sudden reaction with the steering wheel). First results indicate that effects of cross-task compatibility depend on the relevance of the driving situation and on the direction of the response that has to be performed in the primary task and not so much on the position at which the information is presented. Moreover, although the groups of elderly participants show increased reaction times they do not seem to be especially impaired by compatibility effects. This finding might be a result of the increased
amount of expertise in older drivers.
 
Practical Relevance
 
Information from a driver assistance system (e.g. a navigation system) often requires to react immediately thus introducing a secondary task beneath the main primary task (driving the car). The study examines with younger and elderly professional drivers whether and how driving is impaired by dual-task demands, especially in critical situations. Based on the findings we plan to propose guidelines and recommendations for the age-based design of operation devices in vehicles and to suggest training programms for older drivers.
 


Perceived strain and well-being in different types of offices  (Vol. 2 – 2007)
 
Author: Lukas Windlinger und Nina Zäch
 
Keywords: wellbeing, work environment, strain, office design
 
Summary
 
Office work places typically entail conditions in the working environment that can hardly be influenced and controlled by the users. The most important of these conditions is environmental noise. Environmental noise can be disturbing and distracting. In the present research, perceived environmental strain is analysed in large and small offices. Further, the effects of environmental strain in offices on wellbeing are analysed. Noise in offices is mostly caused by conversations of colleagues. Conversations can be disturbing and distracting on the one hand, on the other hand they can act as a resource. Unhindered communication in large offices enhances social support and contributes to stronger cohesion in workgroups. Social climate is therefore analysed as well. Relevant organisational conditions are included in the research as control for two reasons: (1) individual conditions are influenced by many factors of the work system, and (2) the control guarantees the comparability of the employees in different types of offices. The allocation of employees to different types of offices is usually based on functional and status related reasons. Comparisons between employees in different types of offices therefore are meaningful only when work related organisational conditions are controlled because these conditions define control latitudes which are reflected in work organisation and work design. In previous research on environmental strain an well-being in offices, however, organisational conditions have not been controlled. A written questionnaire was administered in two financial service providers. Subjects with a degree of employment of less than 80% and subjects that occupied their office workplace in less than 80% of their working time were excluded from the sample. The final sample consists of 51 subjects. All participants work in the same jobs in the same divisions. In each of the two companies, the participants work in different types of offices in the same building. There were no differences between the two companies, nor in sex, age and jobs between the two office conditions. No difference was found in work content and work conditions between the two office conditions either. Results show that perceived strain caused by noise is significantly higher in large offices compared to small offices. Among all environmental strains, noise is most prevalent: in both office types nearly 60% of the participants reported medium to high strain caused by nvironmental noise. Telephone conversations and conversations by colleagues were referred to as the most frequent sources of noise. Perceived distractions were significantly more frequent in large offices. Further statistically significant differences occurred in perceived strain through visual stimuli and through aircondition. In all cases the large office group reported more strain than the small office group. The differences in perceived strain are not paralleled in differences in measures of well-being. Considering middleterm effects on well-being, however, the large office group shows a more pleasurable social climate. On the other hand, the small office group experiences more mutual social support. These results suggest a possible qualitatively different development of social relations, depending on size and spatial structures in offices.
 
Practical Relevance
 
The design of office environments has effects on well-being, social climate and satisfaction of the users. The systematic analysis of these effects and the knowledge of causal influences contribute to improved design of office environments on the macro-ergonomic level.
 


Evaluation of instruments to evaluate job satisfaction (Vol. 2 – 2007)
 
Author: Yvonne Ferreira
 
Keywords: job satisfaction, job satisfaction measurement, concept of job satisfaction
 
Summary
 
In the following article the results of a study of instruments of job satisfaction are presented. Given the fact that on one hand the construct of job satisfaction is still of major interest there is on the other hand at least scepticism regarding the reliability and validity of the present evaluation instruments. The initiation of this research was the failure attempt to use an already existing questionnaire for the evaluation of job satisfaction as valid criteria for the development of a new instrument. Hereby the instability of the postulated factors became obvious. All in all over 300 instruments to evaluate job satisfaction could be determined. As a restriction it must be noticed that foreign instruments, if not published in English or German language, could only be considered in a few cases. In the following research 201 instruments could be adopted.
 
1. To permit a comparison between the investigated instruments, eligible criteria have to be determined: Four minimal requirements have been defined to integrate as much instruments as possible in the evaluation (cf. for example Lienert and Raatz 1994). With a close approach to job satisfaction should all common definitions of job satisfaction be enclosed (cf. Ferreira 2001).
 
2
. Application of the instrument within a norm sample of at least 100 persons
 
3. Investigation of the instrument regarding the quality factors of reliability and validity (the question of objectivity remains unconsidered, because nearly all given instruments were concerned questionnaires)
 
4. Availability of the instrument (publication)
 
The research and categorisation of job satisfaction instruments shows that during and especially after the conception process of the instruments scientific criteria was often not respected: Instruments have not been tested on a sufficient number of subjects, quality factors have not been calculated, the development was not based on a theory or own concept and further more. The reason for this could be that lots of instruments were either constructed for academic research or as an answer to specific organisational questions. In addition to the limitation of existing measurement instruments, the application of factor-analytically extracted instruments for job satisfaction evaluation could at least be challenged. As a result of this, the lack of prove of the correlation between job satisfaction and other factors could be due to the used measurement instruments or due to the postulated correlation. If on the one hand the correlation between job satisfaction and other factors and on the other hand the deployed evaluation instruments are at least questionable, this can be the reason to intensify the studies in both research areas. The aim of this research, to create more humane and economic workplaces, is not only worthwhile but has also socio-political and economic brisance (cf. among others Hamm 1999). It is very important to discuss and to show job satisfaction in other theoretical contexts. There is a large number of approaches to extend and to improve the concept of job satisfaction, e.g. by integrating other concepts as emotions and commitment (cf. e.g. Fischer 2006, Wirtschaftspsychologie 2005). From a scientific point of view, those developments are very promising and give hope for improved instruments.
 
Practical Relevance
 
Job satisfaction is a popular construct especially at operational leve. The correlation between economic factors as for example quality, time of absence, and performance are postulated.
 


Learning Cultures and Different Perceptions of Learning-Conducive Working Conditions in Five Businesses – Research Results  (Vol. 2 – 2007)
 
Author: Erika Spieß, Ralph Woschée und Brigitte Geldermann
 
Keywords: ·Learning cultures, development of competence, job satisfaction, learning-conducive working conditions
 
Summary
 
The project „new learning cultures in organizations“, part of the BMBF program „Learning Culture – Development of Competence: Learning in the Work Process“, accompanied five company projects to develop personal and organizational competence from 01/01/2001 to 31/12/2003. In the course of changing organizational and career models, employees change workplaces more frequently, individual learning needs increase, becoming more specialized and complex. Standardized intensive further education courses can no longer meet these needs. However, the necessary greater staff participation in organizing the work process presupposes that learning opportunities are available. A central assumption was that measures for developing employee competence in the company would result in a learning culture that is perceived to be learning conducive and thus lead to job satisfaction. The project was based on theoretical concepts of developing competence, self-organized learning and learning culture. The heart of learning competence are the individual’s selforganization abilities. Competence is considered to be a disposition to be able to self-organize. Growing self-organized learning at the workplace therefore influences the role of the superior within the company: the superior has to support these new learning processes actively. A central prerequisite for developing employee job competence is a corresponding learning culture within the company that makes this possible and promotes it. „New learning cultures“ refer primarily to learning activities and a setting in which learning occurs. Superiors become important moderators in establishing a corresponding learning culture in companies. Based on this, the research projects assumed the following:
 
1. Companies differ regarding conditions for employeecompetence development.
 
2
. Due to current developments in companies, superiors have to assume the role of a supporter and promoter of learning in employee self-organization. As a consequence of the different role functions, the superior’s and the employee’s perception of the learning conduciveness of working conditions for employee competence development, however, is not the same.
 
3. Perception of learning-conducive working conditions and job satisfaction are positively correlated. The heart of the survey were questionnaires on learning- relevant features of the job and on learning in the work process. Competence-building measures comprised developing superiors’ and employees’ competence by means of teamoriented action, discussing company targets, introducing a system of agreeing on targets and introducing a program for developing technical competence as well as introducing a yearlong program for long-term further development of organizational and staff competence. An important outcome was that the learning cultures of the five companies differed as did the superiors’ and employees’ subjective perceptions of the learning conduciveness of working conditions. The employees rated the work beneficial learning conditions from their vantage point and the superiors from theirs. The employees’ and the superiors’ perceptions of the conditions for developing employee competence varied significantly: the superior always overestimated staff conditions. Moreover, job satisfaction and learning in the work process correlate positively.
 
Practical Relevance
 
For the company, the outcome means creating a good learning culture is very important for employee job satisfaction. Changing the learning culture, however, is a slow, weary process in which superiors play a significant role and assume a model function. The aim is to develop a learning program that widens the scope of the job and promotes the versatility of requirements and transparency of processes. Offering an ability to participate and to develop contributes to staff satisfaction. spiess woschee geldermann.pmd 04.06.2007, 09:26 28
 


Vol. 3 - 2007

Development of Work Physiology and Ergonomics in Germany —Tasks and Objectives in Education and Research (Vol. 3 – 2007)

Author: Helmut Strasser

Keywords: Work Physiology, Work Science/Ergonomics, Physiological Costs

Summary

Work physiology which was established as a scientific discipline in Germany around the turn from the 19th to the 20th century, when the focus was on basic investigations into physical work and the human metabolism, became the germ cell of todays widely-spred work science and ergonomics. But it was a long way until in the sixties, several chairs and research institutes for work physiology, work science, and ergonomics had been founded mostly in faculties of mechanical engineering at Technical Universities. It is remarkable that the first generation of the full professors and most of the renowned founders of ergonomics had been working and had acquired their fundamental scientific education in the world-famous Max-Planck-Institute of Work Physiology in Dortmund. Meanwhile, there is no doubt that the ergonomic design of workplaces and the appropriate sensoric and motoric interfaces in a man-machine system according to the principle „man, the measure of all things“ would not correctly be possible if fundamental work-physiological knowledge of the characteristics of man, i.e., the capacities as well as the limitations of the human sensory organs and the hand-arm system, were not considered. Nonetheless, work physiology as a discipline in Germany has lost somewhat in popularity during recent years. This may be due to the fact that on the one hand basic knowledge has found meanwhile its way into generally applicable teaching and guiding principles of work science, and ergonomics, respectively, and on the other hand that effective research utilizing work-physiological methods needs years of experience and mostly deserves intensive and hard working as well as precisive working hypotheses and strategies. Yet, a type of working associated with something like „blood, sweat and tears“, of course, is not as popular in a new generation of researchers as experiencing „fun at work“, and getting fruitfull success, well in time via less strenuous and more softly methods. But in ergonomics research, never all claims may be given up to develop and apply methods which allow to measure „physiological costs“ of work. Assessing the ergonomic quality of, e.g., hand-held tools and computer-input devices solely by expert rating, „paper and pencil“ tests or subjective ratings could be insufficient and misleading. In order to clearly quantify what has to be invested or „paid“ by the musculature involved in work when utilizing more or less ergonomically designed tools, also physiological measures such as multi-channel electromyography have to be applied. Similar is true for evaluating noise exposures who’s physiological costs exactly can be quantified in „figures and numbers“ by measuring threshold shifts, and their restitution if there are still reversible and not already irreversible physiological responses associated with noise. So, comparable to work-related heart rate increases and their restitution after work as specific physiological responses to dynamic muscle work during classical work-physiological investigations, also in the future it will be crucial in ergonomics to apply objective, reproducible, specified and valid experimental methods and to address concrete questions which always require a clear commitment to working on details as it was and is the case in work physiology.

Practical Relevance

Comprehensive work-physiological knowledge of the characteristics of the human organism, always was, and will still in the future remain, a prerequisite for the truly ergonomic design of workplaces and user-friendly tools and products. Besides physiological core competencies for a competent ergonomics education, work-physiological principles and paradigms are absolutely necessary for ergonomics laboratory research approaches as well for integrative applied field research aiming at the analysis, assessment and humanoriented design of the working environment.


Room acoustics for open-plan offices  (Vol. 3 – 2007)

Author: Helmut V. Fuchs

Keywords: Open-plan office, noise control, sound absorber, sound barrier, communication

Summary

Noise in open-plan offices, schools, lobbies, restaurants and other communication areas is predominantly self-generated by their respective users. The lack of appropriate sound absorption in traditionally furbished classrooms and scarcely furnished modern office environments forces all occupants to inevitably raise their voices, as soon as the necessary communication among them or/and on the phone sets in. Conventional damping at medium and high frequencies by means of absorptive furniture and screens may somewhat reduce the detrimental effect of the room on speech intelligibility and the resulting noise development. Current research, consultant and restoration activities have, however, revealed that novel low-frequency, yet broadband sound absorber modules can tackle this problem more efficiently. The new room acoustic concept was earlier described more fundamentally in „Relevance and treatment of the low frequency domain for noise control and acoustic comfort in rooms“ (Acta Acustica 91 (2005), p. 920-928). Numerous refurbishment projects for meanwhile more than 100 000 valuable workplaces showed amazing improvements of their acoustic comfort in enclosures where more conventional noise control measures had failed or were impracticable for several reasons. A more rigorous acoustic treatment of rooms with high communicative demands has gained considerable attention for three reasons:

♦ Acoustically „soft“ outfits with carpets, curtains, upholstery and acoustic ceilings, plasterwork, baffles or sails have completely gone out of fashion, not only but mainly for utility reasons.

♦ Acoustically „rigid“ concrete, stone and glass surfaces are not only considered as an indispensable part of modern architectural design but also form the necessary presupposition for construction elements to be thermally activated.

♦ For up-to-date administration, business, education and communication centres tolerable noise levels and high speech intelligibility have become important ergonomic qualities, the absence of which increasingly provokes severe criticism and complaints from users as well as frominvestors.

An always prevailing lack of space and shortage of resources hardly ever led to a well-balanced or even low-pitched acoustic treatment of, in particular, the smaller or flat rooms. It may therefore come as a surprise for many a skilled acoustician that one may now - with relatively simple measures and at moderate costs - cope with the long-standing low-frequency problems by employing especially developed building elements of a novel, visually attractive design. The concentration of millions of work places especially in openplan offices definitely exceeds ergonomic limits, as the noise generated by its users unavoidably reaches intolerably high levels of occasionally more than 80 dB(A). Following traditional concepts, one tries to damp the sound generated through multiple communication processes within the room by means of absorbent ceilings and conventional office dividers. This, however, is now in conflict with up-to-date trends towards thermal activation and visual transparency of all bounding surfaces. Above all, these out-of-date acoustic measures do not tackle the problem at its roots, i.e. at the vivid sound sources, as may be concluded from numerous negative experiences. Acoustic comfort and noise control as indispensable presuppositions for concentrated intellectual work can only be achieved, if equally low reverberation of the room down to very low frequencies supports best speech intelligibility. Novel high-performance broadband compact absorber modules and fully transparent sound screens completely incorporating such modules now enable an optimal visual and acoustical layout for truly open-plan offices. The article shows the visual differences of the innovative as compared to the conventional acoustic treatments in openplan offices and describes the respective efficiency of two alternatives in terms of the

♦ pressure level decay from an artificial „pink noise“ source along measuring paths spanning the bureau landscape as a whole as compared to that in a „free field“ (with no reflections) and

♦ reverberation time measured as a function of frequency. After this acoustic dilemma could be solved very convincingly in prominent cases, many fallow lying office immobilia may be revived for open-plan (group- or team-) office landscapes with only 6 to 12 m2 floor area occupancy per user. With their high flexibility and relatively low price the proposed innovative measures can help to raise the value of numerous objects – attractive aspects for clear-sighted project developers and investors in a fast growing market.

Practical Relevance

The often deafening noise in service centres generated by the users themselves can be lowered by more than 30 dB(A) by employing novel sound absorbing and shielding glass dividers with integrated broadband sound absorber modules – a practicable guideline for a persistent revalorization of many fallow lying office immobilia.


The influence of milking parlour equipment on the workload of the milker (Vol. 3 – 2007)

Author: Martina Jakob, Sandra Rose und Reiner Brunsch

Keywords: Work load · strain · milkers · musculoskeletal disorders · motion analysis · body posture analysis

Summary

The technical development in milk production has lead to changes in milking parlours concerning the workload of milkers. Physically demanding work tasks have become less, but at the same time the variety of tasks has decreased. Modern milking parlours are functioning highly efficient. One of the remaining physically demanding tasks is attaching the milking unit; strain depends on its weight, the task duration as well as the body posture. The body posture is influenced by the size of the person and the distance between the worker and the udder. The results of an evaluation of temporary disability data showed a significant number of female milkers being absent due to musculoskeletal disorders (Liebers & Caffier 2005). A swedish long term survey also reported a raise in discomfort among milkers (Pinzke 2003). This is contradictory to the actual decrease of physically demanding work during milking. The objective of this study was to analyse the actual work routine and the influence of several factors like the equipment and the work style. For this the body postures were measured with a motion analysis system (Simi Reality Motion Systems, Germany) in two different milking systems (Herringbone, Side-by- Side) while the milker was attaching the cluster. The applied motion analysis system was based on video tracking of the work process and the consecutive generation of 3-Droom- coordinates calculated from marker positions on the worker’s body. The marker positions were determined from two videos for each frame by tracking algorithms of the software. The postural analysis brought up that it is necessary to bend forward, lift the arm and stretch it to be able to reach the teats. The arm holding the cluster showed constant values for the duration of the work task, but also variations of 10° within the same setting. The strain on the arm depends on the cluster weight and the upper arm elevation. The moment of torque was calculated for a light and a heavy milking cluster based on the measured length of the lever arm. The attaching arm showed a dynamic progression during work; the variation was larger in practice compared to the experimental setting. The exact body posture depended on the size of the person, the work style, the working environment and other parameters. For the herringbone milking system the data variation tended to be larger. The herringbone is characterised by a larger distance between the milker and the furthest teat. The duration for attaching the cluster was also longer compared to Side-by-side. It took round about 10 seconds from grabbing the cluster until it was attached to the udder. The exposure analysis according to DIN EN 1005-4 brought up extreme joint postures for the upper arm elevation. The weight of the milking cluster is not regarded and inevitably causes higher strain. There are some technical devices that help to ergonomically improve the work task. It is possible to adjust the floor height, service arms are available to hold the cluster and take it off and indexing is used to move the cow in the right position. Further investigations are planned to measure the impact of cluster weight in combination with different body postures. The proposal of ergonomic guidelines for the implementation of technical equipment in milking parlours shall be achieved.

Practical Relevance

Appropriate work place design to prevent disorders amongst milkers is based on the knowledge of the work routine as well as factors that may cause disorders. The presented results shall contribute to develop guidelines to implement ergonomic devices to reduce strain and improve work place design in modern milking parlours.


The effect of predictability of system response times on the user’s behavior and state (Vol. 3 – 2007)

Author: Werner Kuhmann, Florian Schaefer

Keywords: system response times, predictability of work flow, user behaviour, user state, human computer interaction

Summary

How well do unexperienced users estimate the length of enforced waiting periods in human-computer interaction, and in which way does the predictability of such waiting periods influence the working behavior and the psychophysical state of the user? These were the central questions of a study in which simple tasks were separated by waiting periods of 0.5 to 5.5 seconds and mean of 3 seconds. Subjects performed so called ‘Sterzinger lines’ of letters in random sequence in which targets had to be identified and marked by the cursor. Each line contained a target which occurred in one of ten possible positions within the line. Waiting periods were either coupled with the target position, and thus predictable, or were randomly distributed, and thus unpredictable. For predictable waiting periods an additional experimental condition was introduced in which the length of the waiting period was visually indicated on the screen. This display was analoguos to a ‘bar display’: when the target was marked correctly, the left-hand part of the line remained on the screen. After the ending of the waiting period the next task was presented. In 20% the presentation of the next task was, however, postponed until the user pressed a ‘restart-button’ (Stop-Response-Time-Task, Stop-RT), with an additional waiting period of 8 seconds. Pressing the restartbutton too early, i.e. befor the waiting period was expired, therefore resulted in an unproportional long additional waiting period, and thus in prolonged working time. Response times in the Stop-RT-Tasks thus give an estimate of the internal representation of the length of waiting periods. The results show, that Stop-RTs of the subjects under the predictable conditions corresponded with the target positions of the preceeding task. For subjects working under the unpredictable condition Stop-RTs were uncorrelated with the target position and exceeded the longest waiting period. Obviously, subjects under the predictable conditions tried to optimize their Stop- RT-response-times and to respond differentially to the length of the waiting periods. Subjects under the unpredictable condition followed the longest of the waiting periods in the Stop-RT-tasks. Both strategies, differential responding and responding to the longest waiting period, indicate that subjects tried to optimize the total amount of working time. Differential responding under predictable conditions, however, was associated with an additional cognitive work load as compared to concentration upon the longest waiting period only under the unpredictable conditions.

Practical Relevance

In modern software design it is well established to support the users with information about actual status of the task perfomed and time to completion of the task. For example, the status is displayed as a moving bar, and the completion time is calculated and displayed in minutes and seconds. Completion times represent waiting periods for the user, which usually are called ‘system response times’. When deciding which display of system response times during human-computer interaction should to be used one has to consider weather the predicted time intervals are basically computable due to their regularity or if they cannot be calculated due to that they depend on external factors. The results of the present study show that in the first case the user is able to detect these regularities by implicit learning and thus to optimize his or her work flow according to the system’s timing behaviour. In this case the additional visual display is of no use and may even irritate the user. In the second case – when the system response times cannot be calculated and are unpredictable – a display of predicted waiting periods lacks accuracy and thus is less helpful. Therefore, a display of actual progress should generally be preferred.


Promotion of Changing Processes by Efficient Team Management in a Company Health Insurance Scheme  (Vol. 3 – 2007)

Author: Stefan Beyer, Hans-Jörg Gittler, Barbara Neubach und Klaus-Helmut Schmidt

Keywords: Health insurances, organizational development, teamwork, productivity of work groups, ProMES

Summary

The general development of the health insurance market, decreasing financial resources and increasing competition pressure, necessitate a reorientation of compulsory health insurances in Germany. In the future, a health insurance company’s success will increasingly be determined by the ability to realize continuous improvements of productivity and to guarantee good service and product quality. Health insurances must also be able to react flexibly to changing legal duties or competitive conditions. In this regard, team based structures have more innovative potential than hierarchically and functionally divided structures. The answer to the question whether the introduction and use of teamwork in fact lead to the desired effects, depends on numerous variables and limiting conditions which have to be considered during the introduction and further development of teamwork. The present article describes how BAHN-BKK, an important company health insurance scheme, introduced team structures in its regional branches through a multi-staged organization development process. During the first stage the company improved the surrounding conditions to encourage performance. The second stage entailed purposeful training measures to support the formation of teams, while the third stage introduced a Productivity Management and Enhancement System (ProMES), giving the teams a useful tool to develop strategies and measures in order to improve their productivity and to help the people do their work. The results of a concurrent research project confirm that the use of ProMES allowed the teams to improve customer orientation, economic efficiency and work efficiency in a pragmatic manner. The verified results showed improved productivity and quality in most cases. By determining major areas, deploying delegates during the development process and introducing employees gradually to all relevant sectors, the management of BAHN-BKK was able to treat the resources of decision makers and the project participants with care. Furthermore, opposition and problems of acceptance could be identified and removed in order for the changing process to be persistent.

Practical Relevance

The field report provides an example of how enterprises can be enabled to arrange important preconditions for efficient teamwork applying a multi-staged organization development process. The report also confirms that the management system ProMES introduced to support team processes provides an appropriate basis for measuring performance, collecting systematic feedback and fostering target agreement. The use of ProMES thus guarantees comprehensive enterprise objectives to be converted into team actions. Additionally, the stability and efficiency of team processes can be improved, creating an atmosphere of learning, development and innovation.

 


Vol. 4 - 2007

Continuous head gestures – an alternative way of controlling menu based in-vehicle information systems? (Vol. 4 - 2007)
 
Author: Claus Marberger und Michael Kondzior
 
Keywords: User interface, gesture control, head orientation, head movement, multifunctional controller, menu
systems
 
Summary
 
Infotainment systems inside modern vehicles include a great number of features such as multimedia players, car diagnosis systems or communication functions. In order to keep these systems usable and safe while driving special care needs to be taken for the user interface design. Multi-functional Rotary-Push Systems (RPS) seem to become a standard input device for controlling these kinds of systems in Europe and the USA (Rößger & Hofmeister 2003). However, from an ergonomic point of view this user interface concept has also weaknesses which cause usability difficulties for many drivers. This report describes an experimental study that compares an alternative interaction concept based on continuous head gestures against the standard approach using a rotary-push-device. The new concept utilises a sensor that measures the driver’s head orientation in real time. This parameter can be used to control a pointer or to highlight an area on the central screen. The confirmation of a highlighted area can be achieved by pressing a simple button at the steering wheel. A theoretical analysis of the head gesture based interaction concept reveals potential advantages: First, the homing operation (moving finger/hand towards the input device) should be simpler and take less time (mainly looking at display vs. moving hand towards rotary-push-device. Second, the hands don’t need to be taken off the steering wheel for controlling the graphical user interface on the central screen. Another advantage over the rotary-push-device is that the movements of the head are compatible with the graphical feedback of the system. This is not the case with rotary-pushdevices since turning the device clockwise often causes system reactions either up or down. However, there are also potential disadvantages of the head based concept: Since there is no haptic feedback the interaction may capture more visual attention than the (haptic) rotary controller. The experiments covered various comparisons between the head based interaction concept and the RPS concept in terms of ♦ driving performance (while interacting with the system), ♦ task performance (while driving) and ♦ subjective criteria such as mental effort and acceptance. Both interaction concepts were implemented in the immersive driving simulator run by Fraunhofer IAO. 10 subjects (8 males, 2 females) took part in the experiment. The primary task was to drive a 40 km stretch of highway at a speed of 100 km/h while performing a specially designed secondary task. The graphical user interface of the secondary task was based on an abstract menu with six buttons at each level. Following a within-subject design the task was to select a series of buttons with a particular given number either by head movements (experimental condition) or by turning the rotary-push-device (control condition). The order of the conditions was counterbalanced to avoid learning effects. All participants were able to practice each interface concept while standing still and also while driving. The results related to driving performance can be summarised as following: The Standard Deviation of the Driving Speed (SDDS) while performing the secondary task did not significantly differ between the interaction concepts (Wilcoxon, z = -1.274, p(2-sided) = .232), nor did the Standard Deviation of Lateral Position (SDLP) (Wilcoxon, z = -0.968, p(2-sided) = .375). However, the mean values were slightly smaller for the RPS condition. The results related to secondary task performance showed that task completion time increased by 10% when using head movements as input method. This difference was statistically significant (Wilcoxon, z = -2.701, p(2-sided) = .004). Also the number of input errors increased significantly by an average of 0.3 errors per task when using head movements as input device (Wilcoxon, z = -2.620, p(2-sided) = .008). Looking at subjective criteria the results revealed that the perceived mental effort was significantly higher in the head control condition (Wilcoxon, z = -2.703, p(2-sided) = .004). Mean RSME (Rating Scale Mental Effort: Min=0, Max=150) values differed by approx. 30 units (RSME head control: 75.5 vs. RSME rotary-push-device: 46.7). Finally, acceptance ratings showed a tendency that interaction with the rotary-push-device was perceived as being more useful and satisfactory. In summary, the obtained results are not in favour of using continuous head gestures. This effect can hardly be attributed to technical shortcoming of the head tracking technology since real time performance and optimal calibration was guaranteed for each subject. Only the difference in training (rotary-push-devices were more familiar than continuous head gestures) could explain the obtained results to a certain degree. However, high mental workload (RSME) values suggest that the head control interaction mode was visually more loading for this type of task. Further research will reveal whether simpler tasks (e.g. vertical scrolling in only one list or simple horizontal selection tasks) can bring out the potential advantages of the head based interaction mode.
 
Practical Relevance
 
Multi-functional central controllers (e.g. rotary-push-devices) are becoming more and more a standard for controlling menu based information systems inside the vehicle. However, the rotary-push-device has clear weaknesses from an ergonomic point of view. It is therefore essential to develop and evaluate new interaction concepts in order to ensure a more comfortable and safer use of in-vehicle information systems, especially while driving.
 


Binocular coordination of eye movements in viewing conditions of head-up displays in road vehicles (Vol .4 - 2007)
 
Author: Wilhelm Bernhard Kloke, Wolfgang Jaschinski, Gerhard Rinkenauer
 
Keywords: Head-up display, eye movement, vergence.
 
Summary
 
Since several years, the automotive industry has promoted the use of head-up displays for the presentation of information in road vehicles. For head-up displays the windscreen is used as a mirror to present information at a virtual location above the vehicle’s front end, which is close to the main viewing direction. When head-up displays are used to display the velocity, this information appears at a larger viewing distance than when the conventional location of a speed-indicator is used. These designs can be compared for ergonomic differences. Among the advertised advantages of the head-up display setup it has been supposed that a smaller change in viewing distance relative to the front traffic could reduce the strain of the driver's ocular system by reducing the need for accommodation and vergence.
Fast ballistic eye movements (saccades) are usually performed when changing the focus of attention from one object to another. Simultaneously, the vergence angle between the visual axes of the two eyes may have to be changed to adjust for the difference in depth. It has been observed in previous studies that a changes in vergence may be facilitated when they are executed in conjunction with saccades, especially when both changes are combined in an ecologically valid way. However, in these previous studies considerably large saccades and near viewing distances have been used, thus the results of the studies cannot be generalized to head-up displays. Therefore, the experiment in our study has been designed to measure the vergence dynamics during fixation changes in head-up display situations. For this reason two conditions were simulated in our experiment: First, moving the eyes from the road traffic to the conventional location of the speed-indicator and back again. Second, moving the eyes from the road traffic to several locations of the head-up display and back again. These locations differed with respect to direction and required saccade size.
In our laboratory simulation, the dashboard was located at viewing distance of 60 cm and a downward gaze inclination of 24°. For the head-up display a viewing distance of 2 m was used at various positions in the vertical and horizontal direction relative to the main direction. The main direction consisted of the viewing direction to a screen in a distance of 10m. This far distance corresponded to the traffic situation. At the far distance a sequence of numbers was presented randomly. After a certain number of digits the letter “S” indicated to the participant to change gaze from the far distance to either the speed-indicator position or to the head-up display. For the speed-indicator and the head-up display position a simple nonius task was used to assure correct fixation. For each of 8 experimental conditions in 2 replicated sessions, the participants made series of 20 cycles of gaze change between the traffic and instrument positions. The movements of both eyes were recorded using an Eyelink II system with a sampling rate of 500 Hz. The vergence angles were computed as the difference between the horizontal angles of the two eyes; the maximum velocity of vergence change was computed in the different conditions of convergence (far to near, i. e. from the traffic to instrument position) and divergence (near to far, i. e. from the instrument position to the traffic).
The observed maximum vergence velocity was higher for the change between traffic and conventional position of the speed-indicator compared to the change between traffic and head-up display positions; vergence velocity was almost proportional to the difference in vergence demand. For the divergent experimental conditions these velocities were about 30% lower than for the convergent ones. Differences of vergence velocity between the head-up display positions were generally small. No significant difference between the horizontal or vertical directions were observed for equal saccade sizes. For the downward positions of the head-up display the vergence velocity was facilitated with increasing saccade size. This effect was statistically significant, though small.
In sum, the time needed for saccade execution was sufficient to converge to the correct vergence angle for both the head-up display and the dashboard positions. Different possible locations of the head-up display appear to have little effect on vergence performance. Our results show that the vergence system is sufficiently prepared to cope with the larger vergence demands of the gaze change to the conventional position of the speed-indicator. If there are no limitations in accommodation the results on vergence facilitation could support the assumption that the accommodation is also facilitated to an appropriate degree, as vergence is a strong stimulus for accommodation.
 
Practical Relevance
 
By use of head-up displays in road vehicles, information such as the velocity indication is presented at a different viewing distance compared to the conventional setup. This induces different demands for the binocular coordination of the eyes under conditions which are not covered yet by the experimental literature. The results of our experiment simulating the conventional speed-indicator position in the car dashboard and various positions of a head-up display suggest that the advantage of the head-up displays for the assistance of binocular eye movement coordination is rather limited. Furthermore, the variation of the vertical position of the head-up display showed a statistical significant effect on the observed maximum vergence velocity. However, the small size of this effect, does not support the notion that specific positions are preferable to others for the matter of binocular coordination. Though the vergence demands were higher for the conventional instrument position, participants were able to compensate these demands by higher vergence velocities. Thus, the reduction of vergence demands at head-up displays in road vehicles seems not to be an ergonomic design issue of primary importance.
 


Evaluation of different mobile map displays for geographic orientation tasks (Vol. 4 - 2007)
 
Author: Claudius Pfendler
 
Keywords: Head-Mounted Display, Workload, Simulator Sickness, Visual Fatigue, Secondary Task
 
Summary
 
Military as well as civilian forces (e.g., police, ambulance) often face the problem of geographic orientation in unknown areas, which increases the time for reaching the target location. Especially in emergency cases this may result in serious consequences. Orientation problems can occur e.g. in abandoned areas, cities as well as within huge buildings, such as factories or airports.
In the experiment described in this paper three different types of electronic displays were compared in respect to performance in an extended geographic orientation task in which the subjects had to cross a simulated virtual urban environment.
The visual simulation consisted of an interactive 3D scenario projected on a screen in front of the subjects. The 2D maps for orientation were displayed on one of the three electronic displays: a monocular head-mounted LCD-display (HMD), a monocular head-mounted virtual retinal-display with laser technology (RD) and a hand-held LCD-display (HHD). The HMD as well as the HHD were polychrome, whereas the RD was monochrome red. The task was to follow a given path shown on the map, and to navigate through the urban environment. On the map the actual position in the city was indicated by a moving dot. A joystick was used as a control device to move across the urban area. Time to complete the task – which took about 1.2 hours on average - was registered as performance measure.
Peripheral attention was assumed to be higher with the HMD and RD as all relevant information is permanently present in the visual field and extensive eye and head movements are not necessary in contrast to the HHD. To measure this effect, a secondary task had to be performed in parallel to the primary task in which subjects had to detect digits (1, 2, 3 or 4) in the periphery of the visual simulation. Selection of the digits, of the length of the interstimulus intervals as well as of the location of the presentation were determined randomly. The subjects responded with one of four corresponding buttons on the joystick. Reaction times as well as percentage of errors, hits, and omissions were measured as dependent variables.
As also effects from the displays on workload, simulator sickness and visual fatigue were expected, these variables were assessed too. Workload was measured with the NASA Task Load Index (Hart & Staveland 1988), which is based on bipolar ratings of mental, physical, and time demands, as well as on evaluation of own performance, effort, and frustration. As simulator sickness might be higher with HMDs, which results, e.g., from dissociations between visual and vestibular stimuli, the Simulator Sickness Questionnaire (SSQ) of Kennedy et al. (1993) was also utilized in the experiment. Based on 17 ratings the questionnaire provides scores for Nausea, occulomotor symptoms and Disorientation, and a total simulator sickness score. Another problem with visual displays especially in extended tasks is visual fatigue. In order to measure this variable the Visual Fatigue Questionnaire (VFQ) of Bangor (2000) was completed by the subjects, which comprises 17 ratings, addressing symptoms like headache, dry eyes, difficulty focusing, nausea and mental fatigue. The SSQ and VFQ were administered before and after the experiment, so that values before the experiment, which might be already elevated, could be taken into account by calculating difference scores. At the end of the experiment subjects also completed a questionnaire regarding experience with PCs, video games and joysticks. These data were correlated with the experimental results.
In total 48 male subjects (mostly students) performed the task. The subjects were randomly assigned to three groups (HMD,HHD, RD) from which each one used only one of the displays (between subjects design). Statistical analysis of subjective and objective data was based on single factor ANOVAs (between factor=display with three levels: HMD, HHD, RD).
The results show no significant differences between the displays. That means that all displays lead to comparable performance, workload, simulator sickness and visual fatigue levels in geographic orientation tasks of the type investigated. Nevertheless, a series of consistent trends were observed which conform to the hypotheses. In this respect the HMD had the shortest completion times in the geographic orientation task as well as the fastest reaction times, the lowest percentage of omissions and the highest percentage of correct responses in the secondary task. The average completion time with the HMD was about 10% shorter than with the HHD. Obviously peripheral attention profits from the advantage that all relevant information is presented in a single visual area, so that extensive eye and head movements become redundant. In contrast, the HHD shows the lowest levels of workload, simulator sickness and visual fatigue. The results can be tentatively interpreted as supporting the hypotheses of performance advantages of the HMD whereas these advantages are accompanied with higher costs in terms of workload, simulator sickness and visual fatigue. The RD scores are usually worst in subjective and objective variables reflecting the disadvantages of monochrome displays. With this display it is more difficult to locate the own position on the map, but this interpretation needs more empirical support.
With all displays workload was high, demonstrating that the effort invested in the task was high. Also simulator sickness was high regardless of the displays. Obviously there are at least two factors in the experiment which might contribute to simulator sickness: The simulation and the visual displays. The first contribution might result from inconsistencies in the simulation like time delays between the locomotion of the own position on the map and in the visual scenario. Such factors seem to contribute more to simulator sickness than the display, as also strong symptoms were associated with the HHD, which should not produce simulator sickness. In contrast to the other subjective data, visual fatigue was low in most aspects. Furthermore, significant correlations between questionnaire data and experimental results show that subjects experienced with video games have a better performance in the primary and secondary task and rate the task lower in respect to mental demands and effort.
Finally it must be pointed out that the conclusions must be supported by further experiments with more sophisticated methods. E.g., instead of the present secondary task with fixed presentation times for the peripheral stimuli it is intended to use variable times. Apart from the fact that it is a secondary task, the approach has much in common with the determination of a recognition threshold with the method of constant stimuli or an adaptive method and is assumed to be a better measure of spare capacity left by different displays.
 
Practical Relevance
 
For geographic orientation of civilian and military forces more and more mobile electronic map displays are used while it is still not clear, which are suited best for this purpose. That is why three displays (HMD, HHD, RD) have been compared in a geographic orientation task with respect to performance, workload, simulator sickness, visual fatigue, and peripheral attention. The results show that the displays don’t differ significantly regarding these variables and therefore no one can be preferred.
 


Evaluation of driver drowsiness by video analysis (Vol 4 - 2007)
 
A
uthor: Axel Muttray, Lorenz Hagenmeyer, Bastian Unold, Jean-Baptist du Prel, Britta Geißler
 
Keywords: Video analysis, sleepiness, driver
 
Summary
 
Introduction Video analysis of driver behaviour is a means to evaluate the degree of sleepiness, using a rating scale (Wierwille & Ellsworth 1994). In a field study, we videotaped bus drivers in order to find out if Wierwille’s descriptors were suitable to evaluate the sleepiness of these bus drivers. Because the rating of video films is very time consuming, we compared Wierwille method with two different methods which might be time sparing. The aim of our pilot study was to evaluate if the methods are equivalent.
Methods Wierwille’s scale, a description of the wake-sleep continuum, was translated into German. Then, we evaluated the descriptors by means of video films of bus drivers. As high levels of sleepiness were missing in these films, we completed the films with recordings from a driving simulator experiment with sleep deprived subjects in order to cover the whole sleepiness continuum. Wierwille’s scale descriptors were slightly augmented and the visual analogue scale proposed was substituted by an ordinal scale as such one could be easier managed by different raters. The scale extends from 0 (awake) to 4 (very sleepy, including microsleep), intervals are 0.25. Videos of six bus drivers and six simulator drivers were evaluated by a trained rater (B.U.), the material included more than 30 hours of video film. Three different analysis algorithms for 5-minutes segments were compared. For algorithm 1, the films were divided into 2.5-minutes segments, which were scored in a randomised order. Then, the highest scores of pairs of chronologically subsequent segments were used for calculation (algorithm 1, gold standard). Algorithm 2 is similar to the procedure of Wylie (Wylie 1998). If the sleepiness score of the first 5-minutes segment of a 30-minutes epoch was < 1, the rater “jumped” to the first segment of the next 30-minutes epoch. If the score was < 1, the rater “jumped” again to the first segment of the next 30-minutes epoch. Segments that were not rated were set to the score of the first segment of the according epoch. However, if the score of a first segment was ≥1, all adjacent segments were rated both forward and backward in time, until segments were found with a score < 1. Algorithm 3 prescribed to view all segments with double film speed. If signs of sleepiness were detected, only the respective 5-minutes segments were scored with normal speed. In order to calculate kappa, all values measured were divided into four subgroups. Agreement between methods was also investigated by means of Bland-Altman plots (Bland & Altman 1986).
Results Most segments were scored ≥ 1. Therefore, the alternative methods did not save much time. For the comparison of algorithms 1 and 2, kappa was 0.67 (95 %-confidence interval (CI) 0.60 – 0.75). The comparison between algorithms 1 and 3 yielded a kappa of 0.69 (95 %-CI 0.62 – 0.77). A kappa from 0.61 to 0.80 indicates a good agreement (Altman 1999). This finding was supported by the Bland-Altman plots which also showed a good agreement between the different methods.
Discussion The comparison between the methods showed a good agreement. However, there are some limitations of this pilot study. Experiments with a greater number of subjects are necessary in order to be able to generalise the results. Furthermore, measurements of intra- and inter-rater reliability are necessary. Our preliminary experiments with two trained raters (L.H. and A.M.) point to a possibly good inter-rater reliability, as the maximum difference of scores of video segments of different drivers with a different level of sleepiness was 0.25. As the agreement between the different methods was good, we conclude that intra-rater reliability of the one rater was good, too.
 
Practical relevance
 
Driver fatigue is a major cause of road accidents. Video analysis of driver behaviour is a means to assess the level of sleepiness. This study is the first step to validate the method. Video analysis may contribute to accident prevention.