Towards Modern Inclusive Factories: A Methodology for the Development of Smart Adaptive Human-Machine Interfaces
Valeria Villani, Lorenzo Sabattini, Julia N. Czerniak, Alexander Mertens, Birgit Vogel-Heuser, Cesare Fantuzzi
TTowards Modern Inclusive Factories:A Methodology for the Developmentof Smart Adaptive Human-Machine Interfaces
Valeria Villani ∗ , Lorenzo Sabattini ∗ , Julia N. Czerniak † ,Alexander Mertens † , Birgit Vogel-Heuser ‡ and Cesare Fantuzzi ∗∗ Department of Sciences and Methods for Engineering (DISMI)University of Modena and Reggio Emilia, Reggio Emilia, ItalyEmail: { valeria.villani, lorenzo.sabattini, cesare.fantuzzi } @unimore.it † Institute of Industrial Engineering and ErgonomicsRWTH Aachen University, Aachen, GermanyEmail: { j.czerniak, a.mertens } @iaw.rwth-aachen.de ‡ Institute of Automation and Information Systems,Technical University of Munich, Munich, GermanyEmail: [email protected]
Abstract —Modern manufacturing systems typically requirehigh degrees of flexibility, in terms of ability to customize theproduction lines to the constantly changing market requests. Forthis purpose, manufacturing systems are required to be able tocope with changes in the types of products, and in the size ofthe production batches. As a consequence, the human-machineinterfaces (HMIs) are typically very complex, and include awide range of possible operational modes and commands. Thisgenerally implies an unsustainable cognitive workload for thehuman operators, in addition to a non-negligible training effort.To overcome this issue, in this paper we present a methodologyfor the design of adaptive human-centred HMIs for industrialmachines and robots. The proposed approach relies on threepillars: measurement of user’s capabilities, adaptation of theinformation presented in the HMI, and training of the user. Theresults expected from the application of the proposed method-ology are investigated in terms of increased customization andproductivity of manufacturing processes, and wider acceptanceof automation technologies. The proposed approach has beendevised in the framework of the European project INCLUSIVE.
I. I
NTRODUCTION
Modern automatic machines and robotic cells in productionplants are becoming more and more complex because of higherdemands for fast production rate with high quality. Over thesebasic functions, today’s factories need to allow for higherlevels of product customisation and variable requirements. Tothis end, advanced functions are implemented, such as faultdiagnosis and fast recovery, fine-tuning of process parametersto optimize environmental resources, fast reconfiguration ofthe machine and robot parameters to adapt to productionchange.Despite high levels of automation of machines and robots,humans remain central to manufacturing operations sincethey take charge of control and supervision of manufacturing activities. Human operators interact with machines and robotsby means of user interfaces that are the modern cockpitof any production plant. For example, they set up machineproduction parameters, identify and solve faults, coordinatemachine and robot re-configuration to enable adaptation toproduct changes. These activities are all performed by meansof computerized human-machine interfaces (HMIs) that areinevitably becoming more and more complex, as new functionsare implemented by the production system [1], [2].In this new scenario, human operators experience manydifficulties to interact efficiently with the machine; this isparticularly true for middle age workers who feel uncomfort-able in the interaction with a complex computerized system,even if they have a great experience with the traditionalmanufacturing processes. On the other hand, complex HMIslinked to complex machine and robot functions create a barrierto young inexperienced or disabled people, who are thenunable to effectively manage the production lines.Such an increasing gap between machine complexity anduser capabilities calls for smart and innovative human-centredautomation approaches that lead to the determination of ad-equate levels of automation for optimal flexibility, agilityand competitiveness of highly customised production on theone side, and, on the other side, a sustainable effort for all workers. Accordingly, novel automation systems shouldembed HMIs that accommodate to the workers’ skills andflexibility needs, by compensating their limitations (e.g. dueto age or inexperience) and by taking full advantage of theirexperience.Moving along these lines, in this paper we present amethodology for the design of adaptive human-centred HMIsfor industrial machines and robots. It consists in enabling userinterfaces to measure the user capabilities, experience andcognitive burden and adapt the complexity and information (cid:13) a r X i v : . [ c s . H C ] J un oad accordingly. In particular, according to the proposedmethodology, an adaptive user interface for industrial ma-chines can be developed that fully adapts to user’s (1) phys-ical status and impairments, (2) cognitive status and mentalworkload, and (3) experience in the working scenario and inthe use of computers. Adaptation concerns visual presentationof information, selection of displayed content, selection ofmachine functionalities enabled to the user and guidance inthe interaction with the process through default recipes andworking strategies. Additionally, the interface provides off-line and, more importantly, on-line training to the user inorder to increase her/his performance and prevent errors. Thesesolutions aim at improving worker’s situation awareness fora more effective, reliable and prompt interaction with thesystem, thus allowing workers to have a full comprehensionof the system behavior and facilitate intervention in dynamicand unforeseen situations.The final goal is to create an inclusive [3], [4] and flexibleworking environment for any kind of operator, taking intoaccount multiple cultural background, skills, age and differentabilities. To achieve this, it is needed to reverse the paradigmfrom the current belief that ”the human learns how the machineworks” to the future scenario in which ”the machine adaptsto the human capability” accommodating to her/his own timeand features. This is realized by adaptively simplifying theHMI based on the user’s features and complementing her/hiscognitive capabilities by advanced sensing and the higherprecision of machines. However, this simplification might leadto the increase of the time needed to perform a processfunction and the reduction of productivity due to limitedfunctionalities enabled to low skilled users. To overcome thisissue, a training facility needs to be integrated in the adaptiveHMI that embeds a virtual (or augmented) environment toguide and teach the user to evolve her/his capability aimingat a more efficient process, both in terms of time and quality.The approach presented in this paper has been devised withinthe framework of the European project INCLUSIVE, whichseeks to develop smart and adaptive interfaces for inclusivework environment.The paper is organized as follows. In Section II we presenta review of the state of the art on adaptive automationsystems. In Section III the proposed methodology is described,with a special focus on the proposed rules for adaptation inSubsection III-B. The expected impact of the application of theproposed methodologies is investigated in Section IV. Finally,Section V follows with some concluding remarks.II. S TATE OF THE ART
In human-computer interaction, the interface is what userssee and work with to use a device [5]. In industrial scenarios,the HMI takes care of all visualizations and user’s interactionswith the data coming from technological processes, and thusallows the user to operate the machine, to observe the systemstatus and, if necessary, to intervene in the process. Custom-arily, HMIs used in industrial process control applications provide no means to control the amount and form of infor-mation displayed during operation. While the user is flexibleand adaptable, the system is not. Control systems commonlyrespond in the same way without regard as to whether theflow of information is low or extremely high, or the level ofexpertise of the user is good or bad [6]. As a consequence,the responsibility for the interaction is placed on the user,who has to adapt to processes determined by the technicalsystem. Moreover, the flexibility required to deal with difficultsituations must be provided by the operators alone acting underthe pressure of unexpected and rapidly changing hazardoussituations. This issue is even more severe if we consider thatthe amount of monitored data that come from modern plantprocesses keeps increasing and control systems are becomingmore and more complex [7], [6], [8]. Therefore, automationresults in working methods that demand increase with regardto stamina, time pressure and the pace of work [7]. This leadsto detrimental effects on workers’ health and safety givingrise to occupational diseases, such as stress or musculoskeletaldisorders, as well as to occupational accidents [7].To tackle this issue, context-dependent automation, alsoknown as adaptive automation, has been considered [9], [10].Context awareness is the ability of programs, applications orcomputer devices to sense, interpret, respond and act basedon the context. Context refers to any information that can beused to characterize the state of an entity, that can be a person,place, or object considered relevant to the interaction betweena user and an application, including the user and applicationsthemselves [11]. According to this design paradigm, levels ofautomation need not be fixed at the system design stage, butshould be designed to vary depending on situational demandsduring operational use. In this regard, the distinctive featureof adaptive user interfaces is the possibility to change how theinformation is presented so that only relevant information isprovided to users by including the environment and the useras part of the monitored system through adaptive HMIs.Adaptive user interfaces have been developed and imple-mented in different domains, such as automotive [12], [13],[14], aeronautics [15] and smartphone and hand-held devices[16]. However, very few partial attempts and preliminary re-sults on the development of adaptive HMIs for complex indus-trial systems have been reported [6], [10]. In [6] a preliminaryconceptual architecture is introduced that allows to definingan HMI that adapts the presentation of information based onthe operator responsiveness. In [10] different user profiles,such as manager, supervisor and maintenance personnel, areidentified and adaptation is limited to present informationselectively according to the logged account. Going beyondthese preliminary efforts, the methodology we propose in thispaper allows for the development of a complete ecosystemof technological innovations that includes the measurement ofhuman capabilities, the adaptation of the user interface and thetraining of unskilled users. ig. 1. Overview of the proposed approach.
III. P
ROPOSED METHODOLOGY
The methodology presented in this paper aims at developinga smart user-machine interface that adapts the information loadof the HMI and the automation capability of the machine tothe physical, sensorial and cognitive capabilities of workers.The smart interface is based on three main modules, as shownin Figure 1:1) human capabilities measurement (
Measure ): the smartinterface measures the human capability of understand-ing the logical organization of information and thecognitive burden she/he can sustain (automatic humanprofiling). The interface identifies also the real skill levelof the user analysing how she/he operates in the commonworking processes (e.g. measuring the time needed tomove among different screens of the HMI, measuringthe eyes activity in seeking information, etc.);2) adapt interfaces to human capabilities (
Adapt ): the smartinterface adapts the organization of information (e.g.the complexity of the information presented), the meansof interaction (e.g. textual information, only graphics,speech, etc.), and the automation task (normal operation,adaptation to new processes, predictive maintenance,etc.) that are accessible to the user depending on her/hismeasured capabilities;3) teaching and training for unskilled users (
Teach ): thesmart interface is used to teach the unskilled users howto interact with the machine. Depending on the skilllevel of the user and the operation performed by themachine, the interface can train the user by using astep by step procedure, also supported by simulation ona virtual environment. This teaching mode can be on-line or off-line, depending on the level of automationand the criticality of the job operated by the machineor robot. Moreover, in this module, an industrial social network app (Android and iOS) is developed to facilitatethe sharing of knowledge among the users about theindustrial processes and the machine operational modes.Since the behaviour of the interface depends on the ac-tual process organization and operational modes, which arespecifically related to the particular industrial process underconsideration, it is important to establish a general method-ological approach that, then, can be specified by buildingcustomized HMIs according to applications. Thus, the designof such universal adaptation patterns leads to a core meta-HMI ,which is general and dialogue-independent from hardware.This meta-HMI needs, then, to be customized to the specificapplication scenarios, functionalities and hardware targets ofthe use cases.In the following, the three modules will be presentedseparately, with a special focus on the adaptation module,which is the core of the system.
A. Measurement of human capabilities
The first step towards adaptation is the measurement of theindividual capabilities and strain level while fulfilling operativetasks. Firstly, the effect of age (changed perception, cognitionand motor skills), dyslexia, second-language speaking, disabil-ities (one handed operation, colour blindness, etc.), missingexperience in the context and impaired abilities in acquiringknowledge are measured a priori, at the first involvementof the user with the automatic system. Then, the strain ofthe operator is continuously assessed in real time. To thisend, contactless and body-worn sensors are used to measureseveral physiological indicators, such as heart rate variabilityand electrocardiographic activity, galvanic skin response, eyetracking, blink reflex, skin temperature, cerebral electricalactivity, and adrenaline/noradrenaline levels [17], [18].
B. Adapt interfaces to human capabilities
Results of the measurement module are directly mappedinto a suited degree of adaptation of the interface. Adapta-tion occurs at sensorial, cognitive and interaction levels. Theproposed approach is summarized in Figure 2.Sensorial adaptation is meant to tackle physical, visual,auditory and dexterity impairments of users. In this regard,the first step towards adaptation consists in meeting userphysical impairment mainly by varying the presentation ofinformation, e.g., adapting font size, accompanying iconsto short text description, enabling audio input and output.Although such features can be manually enabled/disabled bythe user, the optimal configuration is automatically selected bythe interface on the basis of user’s claims and measurements.Also environmental conditions, such as lighting, noise and useof protective gloves, are considered. A deep analysis based onergonomics factors [19], [20] drives the selection of such anoptimal configuration.Cognitive adaptation provides the adequate level of instruc-tions and details in order to not exceed the cognitive capa-bilities of less experienced workers and increase performanceof more experienced workers. It is implemented in terms of:) amount of information presented, 2) guided interactionwith the productive system, and 3) amount of functionalitiesenabled to users. Displayed information is adapted accordingto two factors: user’s experience in the task to accomplish andexperience in the use of the HMI (e.g. novel, occasional orhabitual user).In the presence of inexperienced workers, it is considereduseful also providing an extended tutorial concerning thedescription of machine functionalities and/or the use of theHMI (e.g. interaction mode, details on icons, menus, settingof preferences, etc.). This tutorial can be easily accessed whilethe machine is running and represents a brief and easy toaccess summary of the teaching module. Novel or occasionalworkers are provided with a brief description of each buttonin the interface when the user moves over it with a mouse orfinger, depending on the input device. This feature is usefulalso for elderly users having a long-time experience in the taskbut limited computer alphabetization, thus being unfamiliarwith computer jargon. Conversely, the feature is not enabledin the case of experienced workers, since continuous andubiquitous explanations are superfluous and would slow downthe interaction with the system. Furthermore, depending onworker’s familiarity with the interface and the machine, asfrom findings of the measurement module, information on themachine and the whole process chain are selectively presentedto users. Expert workers have access to global view of theprocess, enriched with information on production rates andtrends, levels of input raw material, and due maintenance ac-tivities. Additionally, an overview of the plant (or a subpart) isprovided by means of an interactive map reporting informationabout alarms, failures, ongoing tasks and production rates.As regards alarms, experienced users are provided with thedetailed list of all active alarms, including those warningsthat do not stop production. Conversely, adaptation for lessexperienced users prescribes a more restricted view of the pro-cess, focused on the activity the user is currently performing.Alarms are filtered so that only the most severe ones are shownto the worker, together with a detailed description of causesand actions to take, supported by pictures, videos and technicaldrawings of the machine. If necessary, secondary alarms areshown at the end of the working day so that the worker canask for assistance to solve open issues and restore the correctmachine status.Additionally, for users with limited experience in the pro-cess or task to be accomplished, the HMI needs to guide theworker in the task. To this end, predefined working recipes(e.g. default values of parameters, working strategies or com-bination of working parameters) are presented by the interface,covering the range of strategies that can be implementedon the machine. In this way, the cognitive gap between theworker and the system is covered by the interface, and thecomprehensibility of the interface is improved. To support theconcept of adaptive automation systems, further adaptation inthe HMI is implemented conditioning the accessible function-alities of the machine to user’s experience. Indeed, productionis adapted to workers’ capabilities by disabling most advanced functionalities for workers with limitations due to inexperienceor disability, with the goal to decrease mental workload.For example, considering the scenario of an old worker notfamiliar with the process, exploiting the measurement module,the HMI will automatically disable complex or unusual tasksthat require many inputs and are not supported by establishedrecipes.
C. Teaching and training for unskilled users
The last pillar of the proposed approach is an adaptive teach-ing system that trains the user according to her/his capabilities,identified in the measurement module, and understanding ofthe working system. Teaching is provided to unskilled usersboth off-line, before starting a working session, and on-line.The off-line training helps the user to get familiar with theautomatic system and learn the task to perform. This is donein a virtual environment replicating the real scenario andworking situations. At this stage, the received training istailored to meet the measured user’s capabilities and mentalmodel. Additionally, while the process is ongoing, the userreceives additional on-line training that provides guidance inthe use of the machine or the robot by means of augmentedreality [21]. This module adapts the training level also to thecurrent understanding of the process, assessed, for example,by tacking user’s errors and eye.Additionally, the teaching module hosts an industrial socialsystem, providing a contextual help menu that broadcasts arequest for help using a social network media (e.g. app foriOS or Android). In the case of a problem, the operator cancontact other qualified experts within the company or, in casethat no sufficient help can be provided by them, further experts,e.g. from the plant manufacturer, can be contacted easily. Therequest is sent to the local workers community or the servicelevel, who has the app installed on her/his mobile device. TheHMI augments the message with the detailed context of useof the machine, in order to facilitate the expert to correctlyaddress the problem experienced by the local inexperiencedoperator. The system routes automatically the request to theusers, who qualify at app login as experts in that particulartask/machine function. The experts can then respond using amessage or voice call to support the unskilled user.IV. E
XPECTED RESULTS
The results of the approach proposed in this paper willbe measured at the end of the INCLUSIVE project on threedifferent industrial use cases which come from three differentmarket sectors and address different user groups: young,elder artisans, seasonal workmen, people with low level ofeducation, people with certified limited cognitive abilitiesand physical impairments ranging from mild to severe. Inparticular, a use case refers to a large bottling company, withautomatic filling and packaging machines, whose employeesinclude people with certified limited cognitive abilities andphysical impairments. The second use case is a companyproducing woodworking machines for artisan shops and small
ENSORIALCOGNITIVE
TYPE OF ADAPTATION ‣ Age ‣ Physical impairment ‣ Environmental conditions ‣ User’s experience in the task ‣ User’s experience in the HMI (and computer alphabetization)
CRITERIA FOR ADAPTATION HMI ADAPTATION ‣ According to ergonomics factors ‣ Amount of information presented ‣ Interaction with the productive system (working recipes) ‣ Functionalities enabled to the user
META-HMI universal adaptation pattern
APPLICATION-MATCHED META-HMIHARDWARE-TARGETED HMI
INTERACTION ‣ Characteristics of the operator (e.g., dyslexia, physical handicaps, stress level) ‣ Received training ‣ Interaction methods, e.g., touch, speech gesture
Fig. 2. Methodological rules for the adaptation module. companies. Thus, the final users of the HMI are elder sub-jects with low education level and computer alphabetization.Finally, the last use case aims at matching a system integratorfor robotic applications to a manufacturing company producingmachines for bending metal parts and components, which arecurrently manually fed mainly because of the variability of theprocess itself and the lack of skilled personnel, able to manageautomatic machines or robots.For the time being, the expected impacts for the increasedcustomization, flexibility optimization of the production andthe widest acceptance of automation technologies have beeninvestigated. They are summarized in Fig. 3.
Effect on customization of manufacturing processes
The results of the approach proposed in this paper areexpected to have a significant impact on the customizationof manufacturing processes, guaranteeing the possibility ofintroducing significant levels of customization in the productsand in the production processes. This will be achieved thanksto the developed smart HMI that will adapt its behaviour atrun time, accommodating time-variable needs together withthe users’ capabilities.In particular, the HMI developed according to the proposedmethodology will make it possible to introduce high levelsof customization in manufacturing process machines, whilereducing the complexity of the interaction to a sufficiently lowlevel, to enable also non-specialized personnel and operatorswith disabilities or with low education levels to effectivelyinteract with the machines. Moreover, as mentioned above for the third use case ofthe INCLUSIVE project, several manufacturing processesare mainly performed in a manual manner nowadays, dueto the high variability of the production batches. Despitethe availability of automatic machines able to perform suchoperations, their potential is often limited by the inability ofhuman operators to interact with such complex systems. Inthis scenario, the availability of such adaptive user interfaces,which support also on-line and off-line training of operators,allows them to effectively utilize automatic machines.
Effect on productivity of manufacturing processes
Additionally, we expect that the application of the proposedmethodology to the design of adaptive HMIs will have asignificant impact on the productivity of the overall manu-facturing processes. Indeed, the performance of the operators,in particular of elderly, inexperienced or disabled ones willbe significantly improved since they will be able to dealwith complex machines and production systems in a profitablemanner.In particular, the proposed inclusive HMI is expected toallow a significant reduction of the time needed to completeeach production task and the down-time for adaptation ofrobotic cells or automatic machines to a variation of theproduction, and increase of the overall line productivity, interms of overall equipment effectiveness. daptive HMIs for inclusive work environment
CUSTOMIZATIONThe adaptive HMI allows to introduce high levels of customization in the products and in the production processes, without necessarily increasing the complexity of the interaction. PRODUCTIVITYThe operators, in particular of elderly, inexperienced or disabled ones will be able to deal with complex machines and production systems in a profitable manner.SOCIAL IMPACT ON EMPLOYMENTA smart HMI that will adapt to the current user’s skill will help in making vulnerable users comfortable with innovative technologies.SOCIAL IMPACT ONWORKING CONDITIONSThe proposed methodology will have an impact related to employment and working conditions in terms of increased usability and diminished cognitive load. ECONOMIC IMPACTThe adaptive HMI will increase the acceptability of advanced automatic machines and robotic cells, thus opening new market opportunities, for producers and integrators of automatic machines and robotic cells.
Fig. 3. Results expected from the proposed methodology.
Social impact on employment and working conditions
Some categories of workers are widely recognized as partic-ularly vulnerable, specifically in the presence of a worldwideeconomic crisis. Old employees, low educated people, anddisabled people fall among those categories. The vulnerabilityof those people is related to the fact that they are the mostlikely to lose their job, and the less likely to be re-trainedand re-employed. This is due either to the difficulty in effec-tively utilizing complex modern computer aided manufactur-ing equipment, or to physical impairment that prevents somekinds of activities.A significant impact on the employment of elderly, loweducated and disabled people is expected to result fromthe application of the proposed methodology. The resultingadaptive HMI, in fact, will automatically adapt to the skillsof the current user, supporting the initial (off-line and on-line) training phase, and letting each user reach high levels ofproductivity in a short time. This will significantly reduce therisk for those people to lose their job due to lack of specificskills. At the same time, in case of loss of the job, it willincrease the possibility of re-employment, since the re-trainingphase is significantly reduced.Furthermore, as is well known, very often people refuseinnovation and automation. Main reasons are related to fearfor technological unemployment (i.e. loss of a job due toa technological change) and, more in general, to difficultiesin adapting to new technologies and procedures. However, technological innovation is mandatory for achieving the con-stantly increasing productivity and quality requirements. Theproposed methodology will have a significant impact on theacceptability of automatic machines and robotic cells in tradi-tional production lines. In fact, providing a smart HMI that willadapt to the current user’s skill will help in making the userscomfortable with innovative technologies and procedures.We expect that the proposed methodology will have animpact related to employment and working conditions in termsof usability and cognitive load. Usability will be evaluatedbased on surveys that will monitor the degree of satisfactionof users comparing traditional HMIs with the smart HMI de-veloped according the proposed approach. The cognitive loadwill be computed by non-invasive measurement of differentphysiological quantities, such as heart rate, blood pressure orpupillary response.
Impact on the market for automatic machines and robotic cells
As detailed above, we expect a significant impact in thecapability, for elderly, low educated or disabled operators, toprofitably utilize advanced (and complex) automatic machinesand robotic cells. This increased level of acceptability willopen new market opportunities, for producers and integra-tors of automatic machines and robotic cells. In particular,the smart HMI system designed according to the proposedapproach will open new market opportunities for automaticachines and robotic cells in traditionally hostile manufactur-ing environments, such as SMEs and artisan workshops.
Impact on the market for HMI systems
According to a report published at the end of 2015 [22],the value of the worldwide HMI market is estimated to reachUS$5,579.3 by 2019, expanding at a CAGR (CompoundedAverage Growth Rate) of 10.4% during the period from 2013to 2019. According to this report, one of the key factors inthe growth of HMI market is to be found in the high rateof development in industrial automation: in fact, complexautomatic machines and robotic cells require modern HMIsystems to be effectively utilized by non-specialized workersin an useful manner. The market of HMIs is composedof different items: touchscreens or displays, industrial PCs,interface software, and various other controllers. Among these,the market for interface software leads the global HMI marketat present: analysts project this market to report the fastestgrowth during the forecast period. The application of theproposed methodology will further push this positive trend,as a consequence of the increasing market opportunities forautomatic machines and robotic cells.V. C
ONCLUSIONS
In this paper, we presented a methodology for the designof adaptive human-centred HMIs for industrial machines androbots. The interfaces developed according to the proposedapproach adapt the information presented to the user and itsvisualization to the user’s capabilities and strain level. Thus,they allow for inclusive and flexible working environmentsaccessible to any kind of operator, regardless of age, educationlevel, cognitive and physical impairments and experience inthe tasks to be performed. Additionally, the proposed approachconsiders a teaching module that adaptively provides trainingto unskilled users on the basis of their capabilities and actualunderstanding of the working scenario.The approach presented in this paper has been devisedwithin the framework of the European project INCLUSIVE,which is ongoing. Thus, the results of the proposed method-ology will be measured at the end of the INCLUSIVE projecton three different industrial use cases. For the time being,the expected impact for the increased customization, flexibilityoptimization of the production and the widest acceptance ofautomation technologies is investigated in this paper.A
CKNOWLEDGEMENT
This work has been supported by the INCLUSIVE collabo-rative project, which has received funding from the EuropeanUnion’s Horizon 2020 Research and Innovation Programmeunder grant agreement No 723373.R
EFERENCES[1] T. B. Sheridan,
Humans and Automation: System Design and ResearchIssues . New York, NY, USA: John Wiley & Sons, Inc., 2002.[2] F. Nachreiner, P. Nickel, and I. Meyer, “Human factors in process controlsystems: The design of human–machine interfaces,”
Safety Science ,vol. 44, pp. 5–26, 2006. [3] J. Abascal and C. Nicolle, “Moving towards inclusive design guidelinesfor socially and ethically aware HCI,”
Intercating with Computers ,vol. 17, pp. 484–505, 2005.[4] C. Stephanidis and A. Savidis, “Universal access in the informationsociety: Methods, tools, and interaction technologies,”
Universal Accessin the Information Society , vol. 1, no. 1, pp. 40–55, 2001.[5] A. Sutcliffe,
Human-Computer Interface Design . Springer, 1988.[6] G. Viano, A. Parodi, J. Alty, C. Khalil, I. Angulo, D. Biglino, M. Cram-pes, C. Vaudry, V. Daurensan, and P. Lachaud, “Adaptive user interfacefor process control based on multi-agent approach,” in
Proc. WorkingConf. Advanced Visual Interfaces , ser. AVI ’00. New York, NY, USA:ACM, 2000, pp. 201–204.[7] E. Flasp¨oler, A. Hauke, D. Reinert, T. Bleyer, N. Henke, S. Kaluza,A. Schieder, A. Windel, W. Karwowski, S. Salminen, J.-C. Blaise,L. Claudon, J. Ciccotelli, L. Eeckelaert, M. Verjans, K. Muylaert, andR. Op De Beeck, “The human machine interface as an emerging risk,”European Agency for Safety and Health at Work, Tech. Rep., 2010.[8] T. Skripcak, P. Tanuska, U. Konrad, and N. Schmeisser, “Towardnonconventional human-machine interfaces for supervisory plant processmonitoring,”
IEEE Trans. Human-Machine Systems , vol. 43, no. 5, pp.437–450, Sep. 2013.[9] R. Parasuraman, T. Sheridan, and C. Wickens, “A model for types andlevels of human interaction with automation,,”
IEEE Trans. Systems,Man, and Cybernetics - Part A: Systems and Humans , vol. 30, no. 3,pp. 286–297, May 2000.[10] A. Lee and J. Martinez Lastra, “Enhancement of industrial monitoringsystems by utilizing context awareness,” in
IEEE Int. Multi-DisciplinaryConf. Cognitive Methods in Situation Awareness and Decision Support(CogSIMA 2013) , 2013.[11] A. K. Dey, “Understanding and using context,”
J. Personal and Ubiq-uitous Computing , vol. 5, no. 1, pp. 4–7, Feb. 2001.[12] H. Sharma, R. Kuvedu-Libla, and A. Ramani, “ConFra: A contextaware human machine interface framework for in-vehicle infotainmentapplications,” in
Int. MultiConf. Engineers and Computer Scientists(IMECS 2008) , 2008, pp. 19–21.[13] A. Amditis, H. Kussmann, A. Polychronopoulos, J. Engstr¨om, andL. Andreone, “System architecture for integrated adaptive hmi solu-tions,” in
Intelligent Vehicle Symp. , 2006, pp. 388–393.[14] S. Garzon and M. Poguntke, “The personal adaptive in-car HMI:Integration of external applications for personalized use,” in
Advancesin User Modeling , S. B. Heidelberg, Ed., 2012, pp. 35–46.[15] T. Inagaki, “Situation-adaptive autonomy: Dynamic trading of authoritybetween human and automation,” in
Proc. Human Factors and Er-gonomics Society Annual Meeting , vol. 44, Jul. 2000, pp. 13–16.[16] T. Gu, H. Pung, and D. Zhang, “A middleware for building context-awaremobile services,” in
IEEE Vehicular Technology Conf. (VTC 2004) , 2004,pp. 2656–2660.[17] D. Kulic and E. A. Croft, “Affective state estimation for human–robotinteraction,”
IEEE Trans. Robotics , vol. 23, no. 5, pp. 991–1000, 2007.[18] P. Rani, J. Sims, R. Brackin, and N. Sarkar, “Online stress detection us-ing psychophysiological signals for implicit human-robot cooperation,”
Robotica , vol. 20, no. 06, pp. 673–685, 2002.[19] ISO:9241-110, “Ergonomics of human-system interaction – part 110:Dialogue principles 2006..” Tech. Rep., 2006.[20] J. N. Czerniak, C. Brand, and A. Mertens, “Designing human-machineinteraction concepts for machine tool controls regarding ergonomicrequirements,” in
Proc. IFAC 20th World Congress Int. FederationAutom. Control IFAC , ser. IFAC-PapersOnLine, 2017.[21] G. Kipper and J. Rampolla,