izpis_h1_title_alt

VEČMODALNA INTERAKCIJA ČLOVEK-STROJ V VOZILIH
ČEGOVNIK, TOMAŽ (Author), Sodnik, Jaka (Mentor) More about this mentor... This link opens in a new window

.pdfPDF - Presentation file, Download (3,29 MB)

Abstract
V doktorski disertaciji obravnavamo področje interakcije med človekom in strojem, in sicer se osredotočamo na poseben primer interakcije med voznikom in različnimi napravami v vozilu. Tu nas zanima predvsem večmodalna interakcija, kar pomeni uporabo več različnih načinov interakcije in združevanje človekovih čutil za razvoj učinkovitih uporabniških vmesnikov. V uvodnih poglavjih obravnavamo teoretične osnove interakcije človek-stroj, saj je ta omejena na eni strani s človekovimi zmožnostmi in na drugi strani z omejitvami uporabniškega vmesnika, ki mora biti prilagojen človeškim omejitvam. Ugotavljamo, da je človekova kapaciteta za izmenjavo informacij omejena s čutili, ki se med seboj precej razlikujejo in zahtevajo specifične prilagoditve tudi na strani uporabniškega vmesnika oz. zahtevajo prilagojen uporabniški vmesnik. Poleg tega je pomembna tudi omejitev človeških sposobnosti obdelovanja informacij in omejitve t. i. delovnega spomina. Posebno poglavje je zato namenjeno prav človekovemu delovnemu spominu, saj je razumevanje koncepta njegovega delovanja bistveno za razvoj večmodalne interakcije. V tem delu ugotavljamo, da človekov delovni spomin informacije, pridobljene iz različnih čutil, lahko do neke mere obdeluje vzporedno. To pomeni, da lahko človek učinkovito sprejema in obdeluje informacije iz različnih kanalov, kar predstavlja največjo prednost večmodalne interakcije. Kljub temu pa ima delovni spomin svojo omejitev, saj lahko v določenem trenutku hrani in obdeluje le omejeno količino informacij. Obremenjenost delovnega spomina imenujemo tudi kognitivna obremenitev in igra pomembno vlogo v interakciji človek-stroj. Ker v doktorski disertaciji obravnavamo takšno interakcijo na primeru voznika oz. voznega okolja, moramo upoštevati, da je vožnja primarno opravilo z visoko prioriteto, ki je tudi varnostno kritično. Voznik mora v vsakem trenutku skrbeti za položaj vozila, spremljati okolico, cestna pravila in morebitne nevarnosti. To opravilo že samo po sebi povzroča precejšnjo kognitivno obremenitev, ki jo je treba upoštevati pri načrtovanju sekundarnih opravil, kamor spada tudi interakcija z različnimi napravami v vozilu. Ustrezno ocenjevanje kognitivne obremenitve voznikov je zato zelo pomembno pri obravnavi interakcije človek-stroj, zato smo v okviru raziskav najprej razvili in analizirali sistem, ki z merjenjem očesnih aktivnosti podaja informacijo o obremenjenosti človekovega delovnega spomina. Uporabili smo enostavno napravo za sledenje pogleda, ki smo jo primerjali s standardizirano metodo merjenja kognitivne obremenitve voznikov. V disertaciji podajamo rezultate uporabniške študije v simulatorju vožnje, s katero smo ovrednotili ta način merjenja in potrdili njegovo pravilnost. Ugotavljamo, da je možno z enostavnimi in cenovno zelo dostopnimi napravami izmeriti kognitivno obremenitev voznikov z merjenjem velikosti zenice ter frekvence mežikanja. Rezultati so še posebej obetavni zato, ker kaže, da lahko razlikujemo celo med različnimi stopnjami kognitivne obremenitve. V nadaljevanju se osredotočamo na interakcijo človek-stroj v vozilu. Po pregledu znanstvenega področja ugotavljamo, da se področje tehnološko hitro razvija in da se v vozilih vsako leto pojavijo nove oblike informacijskih sistemov ter načinov interakcije. V doktorski raziskavi smo analizirali interakcijo človek-stroj v dveh delih. Najprej smo ločeno opazovali in analizirali različne načine interakcije kot samostojne vhodne enote. Zanimal nas je predvsem njihov vpliv na kognitivno obremenitev voznika, varnost vožnje, uporabniško izkušnjo in samo učinkovitost. V drugem delu pa smo glede na izkušnje in ugotovitve prve uporabniške študije izboljšali sistem ter ga analizirali kot večmodalen uporabniški vmesnik in ovrednotili njegovo uporabnost za opravljanje različno težavnih nalog v različnih prometnih situacijah. V sklopu doktorske raziskave smo razvili prototip lastnega uporabniškega vmesnika, ki omogoča komunikacijo med voznikom in napravo preko več različnih načinov interakcije. To so tradicionalni taktilni vmesnik oz. gumbi na volanu, sledilna ploščica, sistem za prepoznavanje prostoročnih kretenj in zaslon na dotik. Poleg tega pa je sistem vseboval različne izhodne enote – zaslon, ki simulira osrednji zaslon na vozilu, projekcijski zaslon pred voznikom in zvočni vmesnik. Ta uporabniški vmesnik je bil integriran v simulator vožnje in je tako predstavljal orodje za ekološko veljavne raziskave. V sklepnem delu primerjamo različne tehnologije in povzemamo nekaj glavnih prednosti novih ter še neuveljavljenih tehnologij, kot je npr. uporaba prostoročne interakcije. Izpostavljamo pa tudi nekaj najpomembnejših pomanjkljivosti obstoječih načinov interakcije in uporabniških vmesnikov, kot je npr. zelo popularen in pogosto prisoten zaslon na dotik.

Language:Slovenian
Keywords:interakcija človek-stroj, voznik, kognitivna obremenitev, varnost, uporabnost, uporabniška izkušnja
Work type:Doctoral dissertation (mb31)
Organization:FE - Faculty of Electrical Engineering
Year:2019
Views:251
Downloads:103
Metadata:XML RDF-CHPDL DC-XML DC-RDF
 
Average score:(0 votes)
Your score:Voting is allowed only to logged in users.
:
Share:AddThis
AddThis uses cookies that require your consent. Edit consent...

Secondary language

Language:English
Title:MULTIMODAL HUMAN-MACHINE INTERACTION IN VEHICLES
Abstract:
In the thesis we study the area of the interaction between a human and a machine, specifically we are focus on the special case of the interaction between a driver and the vehicle. We are interested in the multimodal interaction between those two, which means using more interface designs and combing human senses for developing effective interfaces. In the introductory topics we study theoretical basics of human-machine interaction, because it is limited on one side with human capabilities and on the other with the interface, which must be adapted to those human capabilities. We find that the human capacity to exchange information is limited by the senses, which are quite different from each other and require specific adjustments on the side of the user interface. In addition, the limitation of human information processing capabilities and the limitation of working memory. A special chapter is therefore intended specifically for studying human working memory, as understanding the concept of its operation is essential for the development of multimodal interaction. In this topic, we find that a person's working memory can process information obtained from different senses in parallel, to some extent. This means that one can effectively receive and process information from different channels, which is the biggest advantage of multimodal interaction. However, working memory has its limitations, because it can only store and process a limited amount of information at some point. The human working memories workload is also called cognitive workload and plays an important role in human-machine interaction. In the doctoral thesis we consider such interaction in the case of the driver or of the driving environment, therefore it must be borne in mind that driving is a primary high priority task, which is also a safety critical. The driver must at all times take care of the position of the vehicle, monitor the surroundings, road rules and any other hazards. This task in itself causes considerable cognitive load, which must be taken into account when planning secondary tasks, which also include interaction with various devices in the vehicle. Adequate assessment of the cognitive load of drivers is therefore very important in researching human-machine interaction, so we first developed and analyzed a system that, by measuring eye activity, provides information about the drivers cognitive workload. We used a simple eye-tracking device that was compared to a standardized method of measuring cognitive load on drivers. In the thesis we present the results of a user study in a driving simulator, which we used to evaluate this measurement method and confirm its correctness. We find that it is possible to measure the cognitive load of drivers by measuring the pupil size and blinking frequency with simple and affordable devices. The results are particularly promising because it shows that we can even distinguish between different levels of cognitive load. In the following part of the thesis, we focus on the human-machine interaction in the vehicle. After reviewing the scientific field, we find that ii is developing rapidly and that new information systems and modes of interaction are emerging in vehicles every year. In the doctoral research, we analyzed the human-machine interaction in two parts. First, we separately observed and analyzed the different modes of interaction as independent input devices. We were particularly interested in their impact on drivers cognitive load, driving safety, user experience, and performance alone of those devices. In the second part, based on the experience and findings of the first user study, we improved the system, analyzed it as a multimodal user interface, and evaluated its usefulness for performing differently difficult tasks in different traffic situations. As part of the doctoral research, we have developed a prototype of our own user interface that enables communication between the driver and the device through several different modes of interaction. These are a traditional tactile interface in form of buttons on the steering wheel, touchpad, free hand gesture recognition system and touch screen. In addition, the system contained various output devices - a display simulating a central cockpit display, a head-up screen in front of the driver, and an audio interface. This user interface was integrated into a driving simulator to provide a tool for ecologically valid research. In the final part, we compare different technologies and summarize some of the main advantages of new and not yet implemented technologies, such as use of free-hang gesture interaction. We also highlight some of the major drawbacks of existing interaction methods and user interfaces, such as very popular touchscreens.

Keywords:Human-machine interaction, driver, cognitive load, safety, usbility, user experience

Similar documents

Similar works from RUL:
Similar works from other Slovenian collections:

Comments

Leave comment

You have to log in to leave a comment.

Comments (0)
0 - 0 / 0
 
There are no comments!

Back