what are the research topics of mastar project? · the beijing olympics under the mastar project,...

1 NICT NEWS, NOV, 2008

What are the research topics of MASTAR Project?Nakamura: MASTER project is a research project,which started in April 2008, and the topic of ourresearch is speech and languages. The research scope ofspeech and languages is quite extensive, but one of themost well-known applications of the research is thetranslation technology used to translate writtensentences into that of another language. You can find avariety of translation softwares on the Internet thesedays; however, they all fall short for practical use. In ourproject, we are doing research on speech translation thatcan translate spoken languages. It is a technology moreakin to interpretation than translation. We are also doingresearch on the spoken dialogue system where amachine understands human speech and gives pertinentreplies. Our goal is to accomplish this with multiplelanguages. Not only is the MASTAR Project a large-scale project by world standards but also a provocativeand challenging project because it encompasses multiplelanguages in its scope. We want NICT to become theglobal center for speech and language relatedtechnologies and resources.

Please tell us more in detail about your research.Nakamura: There are four main categories in theresearch. The first category is the development oftechnologies related to the network-based speechtranslation. It is the technology required for machines torecognize and translate human speech and output repliesusing the same language the listener uses. Our projecthas been pursuing the research for the last 20 years andit is starting to bear fruit in recent years. During theBeijing Olympics, we have asked some of the Japanesetourists going to Beijing and the Japanese living inBeijing to use our prototype equipment, and havereceived valuable feedback from them.

So, you have already succeeded in developingthe prototype equipment?Nakamura: That is correct. We have developed two

types of equipments. One of them is a small PC that isused as a stand-alone device. The other system is a typethat uses a mobile phone—the data is first sent to theNICT server, where speech recognition, translation,voice synthesis are carried out, and the output is thensent back to the mobile phone. Some users havecommented that the processing took too long, but theupside of this type of equipment is, of course, that youcan enjoy the benefit of an interpretation service by justhaving your mobile phone. For example, there are manyBrazilian factory workers in Nagoya; the speechtranslation service is very useful there. A localgovernment office equipped with 24-hour Portuguesespeech translation service can answer questions byBrazilian workers visiting the office. This speechtranslation technology was approved to be included inthe list of technologies of “Innovation 25,” which is along-term strategy initiative for the technologyinnovation contributing to the growth with an eye on theyear 2025, put forth by the Cabinet Office. Further, thespeech translation technology will be a part of “ThePioneering Projects” for Accelerating Social “Return”,which had started in April 2008. It is not going to be anendeavor that takes 10 to 20 years to produce results; thetechnology is expected to be available in five or sixyears’ time.

You think you will have created something thatpeople can use in five or six years?Nakamura: I think so. First, we want to developsomething that can be used during travel. The languagesused by the system will be limited to Japanese, English,and Chinese at first, but the number of supportedlanguages will increase in time.

What is the second category in the research?Nakamura: The second category is the machinetranslation services, which I mentioned earlier. We arethinking of industrial usages as the main application ofthis technology—translation of instruction manuals andresearch papers are good examples of it. We will beworking together with the companies using the serviceto collect data from the networks so that we can build

Speech and language research project

Machine translations using the large databaseof examples

Network-based speech translation—interpretation by machines

MASTAR Project

Aiming to Be the Global Centerfor Speech and LanguageResearch

Leadoff Interview

Satoshi NakamuraExecutive Researcher

After having completed his university study. he hadreceived his Ph.D. in Information Science from KyotoUniversity. From 1981 to 1993, he worked for Sharp’sCentral Research Laboratory in Nara. From 1994 to2000, he was the associate professor of the graduateschool of Information Science at the Nara Institute ofScience and Technology. From 2000 to 2005 was the

department head and the director of ATR SpokenLanguage Communication Research Laboratories. He iscurrently the executive researcher and the director of theMASTAR project at Knowledge Creating CommunicationResearch Center, NICT. He also serves as an honoraryprofessor of University Karlsruhe, Germany and aprofessor at Keihanna Joint Graduate School.

Dr. Nakamura holding a mobile device

NICT NEWS, NOV, 2008 2

the necessary database of vocabulariesand original and translated sentences;this will help in the continuousdevelopment of the machine translationtechnology. The data that is specific tothe company is handed over to thecompany, while the other data that ismore general in nature can be utilized invarious other fields of new businesses.

What are the methods used for the machinetranslation?Nakamura: Dr. Makoto Nagao, the former president ofNICT, is a pioneer in machine translation in Japan, andNICT has a history of pursuing the development of theexample-based translation that he advocated. Ourdevelopment effort has now expanded into the statisticaltranslation method, which takes probabilities intoaccount. To be able to machine translate, you must firstdevelop “bilingual corpus.” It is a collection of sentencepairs; for example, a Japanese sentence and acorresponding English sentence. For a Japanesesentence “Watashi wa gakko ni iku,” an Englishequivalent, “I go to school,” is paired. You prepare acollection of such pairs—a lot of them. We use bilingualcorpuses that comprise more than million of such pairs.Further, when you put the bilingual corpus into machinelearning, it would generate the probability data. Wewould know, for example, the usage percentages of“Watashi ga” corresponding to “I.” Or, we would knowthat “to school” corresponds to “Gakko e” morefrequently than “to university.” Therefore, theprobability component is used in machine translation.Further, the word order differences that exist in differentlanguages are made consistent using bilingual corpusescomprising a large number of examples. Once thebilingual corpus for a specific language such as Chinese,Indonesian, and German is developed, machinetranslation across multiple languages becomes possible.The accuracy of the translation is another matter, but themerit lies in the fact that just pairs of sentences arerequired for this type of translation to become possible.

What is the third category in your research?Nakamura: The third category is the spoken dialoguesystem. I have briefly touched on this earlier; anexample is a robot that understands what I speak andgives pertinent replies. There is a system we areworking on right now, which is being used in Kyoto’stourist information center. If you told the machine, “Tellme about the Golden Pavilion,” it would, for example,give you some information gathered from Wikipedia. Ifappropriate information was not found in Wikipedia,web search would be used to obtain the information. Ifyou asked, “How do I get there,” it would answer “Youcan go there by bus” and display the bus timetable. It ismore helpful for such information system to use

animation characters on the display in addition to usingsynthetic voice to help convey messages; the one inKyoto uses the animation characters as well. We are stillat an early stage, but we aspire to develop in the future,a type of machine thatcan carry on naturalconversations withpeople, like the ones wesee in the animation orscience fiction movies.

Like Doraemon?Nakamura: Yes.

It seems to me, there are many other usefulapplications of this system.Nakamura: Yes, there are plenty of other possibleapplications besides the use in tourism. Especially, whenthe multi-language support is implemented, the localgovernments of communities with many foreignresidents will be greatly benefited by this system. Thecall centers are another example.

And, the last category of your research?Nakamura: The fourth category is the distribution oflinguistic resources worldwide. Dictionaries anddatabases of languages are used to provide the type ofservices we are developing, but the languages arealways changing. We would like to incorporate newwords and word usages that are gathered from theInternet and media, and compile databases and distributethem periodically.

It seems to me that the coordination with thebusinesses is important for the advancement ofyour research.Nakamura: Yes, exactly. Till now, this kind of researchtended to be outsourced, but in the MASTAR Project,we will be asking the companies to send personnel tojoin the research of the MASTAR Project. We are alsopreparing to launch R&D forum, where the companiescan participate in discussions.

It looks like the development of the MASTARProject will be headed in diversified directions.Nakamura: The spoken language and written languagethat we use in our everyday life is called naturallanguage. Until recently, machine translation was theonly application of the technology that utilizes naturallanguage, but now the technology is starting to be usedin various other fields as well. A very exciting futureawaits us.

Thank you very much.

Natural communication between man and themachine

Collection and distribution of linguisticresources

Multiplicity of development for naturallanguage processing

Demonstration of Speech Translation


Experiments of the Cellular Phone Speech Translation Servicein a Real Environment

Toru ShimizuKnowledge CreatingCommunication Research CenterResearch Manager, SpokenLanguage CommunicationGroup

After completing his master’scourse and before taking up hiscurrent position, he worked inKokusai Denshin Denwa Co., Ltd.,

(current KDDI Corporation) R&D Laboratories and AdvancedTelecommunications Research Institute International (ATR). He isengaged in research and system development of speech synthesis,natural language processing, and speech recognition.

●P r o f i l e●

Toru Shimizu / Satoshi Nakamura

Field Experiments of theCellular Phone SpeechTranslation Service atthe Beijing Olympics

Under the MASTAR Project, Knowledge CreatingCommunication Research Center strives for therealization of technologies that enable recognition,translation, and synthesis of spoken language in variousenvironments, and translation of various information onthe Internet. This project is expected to both eliminatelanguage barriers and to realize borderlesscommunication.

As a part of this project, we conducted fieldexperiments with the cellular phone speech translationservice in Beijing from August to September of 2008.This experiment served to advance the multi-lingualspeech translation technologies, and to promote ourlatest speech translation system.

We have developed the world’s first “Japanese-Chinese cellular phone speech translation service” whichtranslates daily conversation in the travel domain. In thisexperiment, Japanese travelers who were attending theBeijing Olympics Tours and Japanese residents inBeijing tested our service for communication with localChinese people while sightseeing, shopping, and gettingdirections.. Recorded conversations and questionnairesare used to analyze the system performance.

The photo shows an image of how this cellular phonespeech translation service is used. The service introducesthe following new features.

ï The voice characteristics of the user can be pre-registeredby speaking a few sentences into the mobile phone.

ñ The user can choose an appropriate dictionary accordingto the location and the situation.

ó The translation result can be voice-synthesized.

ò The translation result can be checked by re-translating itinto the original language.

For this experiment, we have developed a propernoun lexicon with Tsinghua University which includesseveral thousand Beijing specific proper namesnecessary for sightseeing, shopping, transportation, ordining. In addition to the “cellular phone speechtranslation service”, we have also used the “handheldspeech translation system” in which all functions forJapanese-Chinese speech translation are implemented inan Ultra-Mobile PC about the size of a smallpocketbook. From the questionnaires, we have learnedthat there is considerable expectation for the speechtranslation technologies, and that there are many usersimpressed with the advancement of the translationtechnologies who expect the realization of commercialproducts. On the other hand, there were also variousfeedbacks from users requesting improvements in someareas, including: more accurate recognition andtranslation of various expressions in daily conversations,more proper nouns such as restaurant names andsouvenirs should be registered, faster processing so thatthe conversation can be carried out withoutinterruptions, a high usability graphical interface, andbetter speech recognition accuracy of Chinese.

We are now conducting an analysis of the dataobtained from the experiment and applying it toaccelerate the research and system development forrealizing a more user-friendly speech translation system.

The development of the world’s first “cellularphone speech translation service”

Realization of the user-friendly system byanalyzing the usability and accuracy


During the Beijing Olympics, the Beijing OlympicCommittee was operating an official multilingualinformation service as a part of services extended to themany foreign tourists visiting Beijing to watch thegames. The overview of the service is shown in Fig. 1.Many different types of official services were beingprovided within and without the Olympic sites using themultilingual integrated information database as the coreof the information. In these official services, a texttranslation system was used to build the core database;for example, the translation of news from thecommittee. The system was also available at theinformation booth installed within the site(Fig.2) .CapinfoCo., Ltd., was the company that was in charge of thismultilingual database operation. NICT and Capinfo Co.,Ltd., had signed an official memorandum in December2008 on research and development; as one of thecollaborators, we have been offering our technicalassistance to them in the Chinese-Japanese texttranslation system. To express their gratitude for ourassistance, Capinfo Co., Ltd., has presented us with aletter of appreciation on Oct 15 after the Olympicssaying to the effect that the success of the Chinese-Japanese information service during the Beijing

Olympics was dueto the collaborationof NICT.

Since 2006, with the subsidy of the SpecialCoordination Funds for Promoting Science andTechnology, NICT has been engaged in the developmentof Japanese-Chinese/Chinese-Japanese machinetranslation system in cooperation with Kyoto University,the University of Tokyo , Sizuoka University, and JapanScience and Technology Agency. The development ofthe system was based on the translation enginedeveloped by Kyoto University. We created a bilingualcorpus database from the Chinese documents related tosightseeing in Beijing and Olympics; the data wasprovided by Capinfo Co., Ltd. A sophisticated systemspecific to the purpose was developed; namely, theservice to be used during the Beijing Olympics.

For our translation system to be incorporated into theofficial system of the Beijing Olympics Committee, thestandard set by the committee has to be met andverification by an independent organization is required.We have improved the system’s performance byenhancing the alignment technology and creating thebilingual corpus database; we were certified by theChinese Academy of Sciences（Institute of ComputingTechnology）since we met the requirements of theBeijing Olympics Committee standard.

NICT, in cooperation with CapinfoCo., Ltd., has developed a touristinformation system, utilizing thedatabase that was created for thepurpose of improving the translationengine(Fig.3). The system wasinstalled in the area having a largenumber of Japanese audiences throughout the entireperiod of the Beijing Olympics; the reactions of theJapanese audiences were favorable.

Capinfo Co., Ltd., and NICT have agreed to continuewith the collaboration regarding Beijing’s tourism, evenafter the Beijing Olympics; this text translation systemis our first effort. We plan to continue moving forwardwith the collaborative work with Chinese organizationssuch as Capinfo Co., Ltd., centering on Beijing tourism.

Text Translation Servicesused in the BeijingOlympics and forSightseeing in Beijing

Hitoshi IsaharaExecutive Researcher

After completing his master’s course,he joined the ElectrotechnicalLaboratory （ ETL） Ministry ofInternational Trade and Industry（MITI）. In 1995, he joined the CRL(current NICT). He is engaged in theresearch on natural language. Hehas a Ph.D. in engineering.

●P r o f i l e●

Multilingual information service during theBeijing Olympics

Beijing tourist information system

Text translation system

Fig.1:Configuration of the Beijing Olympics’Official Multilingual InformationService System

Fig.2: Information Booth

Fig.3:Beijing TouristInformation Systemwhich Functions inJapanese

The Frontier ofQuantum InformationCommunicationTechnologies


Masahide SasakiGroup Leader, QuantumICT GroupNew GenerationNetwork Research Center

After having completed hisdoctoral course, Dr. Sasaki joinedNKK (current JFE Holdings, Inc.).In 1966, he joined theCommunications ResearchLaboratory, Ministry of Posts and

Telecommunications (current NICT). He is engaged in research anddevelopment of quantum information communication technologies.Ph.D. in science.

●P r o f i l e●

The widespread popularity of broadband services andtechnologies such as optical and IP telephony havefamiliarized us with optical communications. Thoughthe rise of information and communication technologiesseems to be boundless, some limitations of thetechnology have come to light. The expansion ofcommunication capacity by wavelength divisionmultiplexing, which enables the injection of largeamounts of optical powers into optical fibers that are asthin as human hair, has reached the saturation limit.Issues such as ensuring high levels of security inelectronic commerce transactions need to be dealt withurgently. Since the mid 1990s, new technologies (suchas those used in quantum computers) that caninstantaneously break the most sophisticatedcryptography have emerged. Currently, informationcommunication technologies are based on 19th-centuryclassical physics, including electromagnetism andoptics. It is hence necessary to change the manner inwhich information is handled to comply with 20th-century physics, namely, quantum mechanics. Quantuminformation and communications technology (quantumICT) based on quantum mechanics can be used todevelop quantum computers that can perform massivelyparallel computation. Furthermore, quantumcryptography that cannot be broken even by quantumcomputers can be developed. Finally, quantum coding,which can help achieve the maximum transmissionefficiency per photon, can also be developed. Researchand development strategies for quantum ICT weredeveloped early on by the Ministry of Internal Affairsand Communications and the NICT; the R&D of thesetechnologies has been promoted strategically andcomprehensively since 2001.

Quantum ICT is most suited for the direct control ofphotons, atoms, and the quantum mechanicalcharacteristics of electrons. Besides quantum coding,which can help achieve the highest transmissionefficiency, quantum ICT includes quantum cryptographyfor ensuring the absolute security of information andquantum internet by which quantum computers can be

connected via quantum teleportation. Quantum ICT isthe ultimate means for fast information transmissionwherein large amounts of data are protected fromeavesdropping. This technology can be applied tomedical services that are essential to our day-to-daylives. For example, results of medical tests can be safelysent to multiple medical institutions for referral; thiswould make very precise medical diagnosis andappropriate treatment possible. The use of quantum ICTis also expected to become common in areas in which itsuse is not immediately conspicuous at present.

Quantum cryptography is expected to becomeavailable for leased government lines in the next four orfive years. On the other hand, it might take more than adecade for quantum coding, which is the ultimatecommunication system based on quantum mechanics, tobe put to practical use. At first, this technology isexpected to be used as a means for the implementationof high-speed links in communication channels underextreme conditions in which amplifiers cannot be used,such as in optical space communications. The receivingends of such high-speed links have to be capable ofextracting the maximum amount of information from asfew photons as possible. Ground-based optical fibernetworks, on the other hand, are expected to be capableof handling communications up to 2020 because of thepossibility of the extension of the existing optical fiberlines. This is because innovations in optical fibers, high-performance amplifiers, etc., are expected to occur. It isnot easy to predict the course of current technologiesafter the year 2020 , but currently, quantum ICT is theonly known means of rapid and secure transfer ofultrahigh-definition images and other types of contentthe global network that integrates ground and spacenetworks.

Our latest achievements in the development ofquantum coding are briefly discussed below. The key tothe emergence of quantum coding lies in the control ofthe superposition of photon states in which multiphotonsexist simultaneously in different states. NICT hasmastered the generation of superposition states

Quantum information communication tosupersede optical communications as thelimitations of the latter loom large

Controlling photon superposition states — thekey to the emergence of quantum coding


containing multiple photons, say up to five photons ormore. We are one of the three organizations in the worldthat have accomplished this. Fig. 1 outlines thegeneration of this superposition state. Far-red light witha wavelength of 860 nm is generated by exciting aspecial crystal with blue light of a wavelength of 430nm. Inside the crystal, optical parametric downconversion occurs — a single photon with a wavelengthof 430 nm is converted into two photons withwavelengths of 860 nm, and squeezed light containingonly an even number of photons is generated.Subsequently, a part of the squeezed light is guided tothe photon detector by a low-reflection mirror so that thesqueezed light passes to the right side of the detectorwhen photons are detected. By this process, a singlephoton can be subtracted from the squeezed light. Thestate generated by this method can be measured using anoptical tomography method called homodynetomography. An example of the measured data is shownin Fig. 2. The state of a single photon generated by weakpumping is shown in Fig 2-a. The distribution charts atthe bottom reveal that the probability of two or morephotons existing at the same time is almost zero. On theother hand, the probability of zero photon (the vacuumstate) remains finite. The reason for this is the

unavoidable light loss caused by theexperimental setup. The second image fromthe top in Fig. 2 shows the time oscillationof the single photon field. The location ofthe two bands corresponds to the electricfield amplitude of the single photon. Whenenergy is fixed, as in “one photon,” thephases of the wave (peaks and troughs)become random. This can be explained bythe uncertainty principle. Two time-independent bands are formed because ofthis reason. A normal laser beam ischaracterized by a sine-wave oscillationwith peaks and troughs. The topmost imagein Fig. 2 show a specialized expression ofthe optical tomography results called theWigner distribution. The deeper the area ofnegative values shown in this figure, thestronger is the indicated quantum effect.

As the number of photons is increasedwith the intensity of pumping beam, waveswith peaks and troughs begin to emerge, asshown in the central image in Fig 2-b. Fig.2-c shows the state in which the excitationintensity and therefore the number ofphotons is further increased. Note that thepeak and trough appear simultaneously atthe time-reference point on the horizontalaxis, indicating that two waves with 180°phase shifts are superposed. In this case,only an odd numbers of photons are

observed, since even number photons interfere witheach other and disappear (In reality, some even-numberphotons remain because of the light loss). This isimpossible in classical mechanics. The superpositionstate that the NICT successfully generated in Autumn2006, such as the one shown here, has displayed thestrongest quantum effect observed so far (indicated bythe negative values in the Wigner distribution shown inthe topmost images in Fig. 2). This result has not beensurpassed by anyone as of November 2008.

It is not unusual that more than half a century passesbefore a great invention finds practical applications. Theimpact of the introduction of an application is oftenmuch greater than that of the invention itself. Forexample, mobile phones, which are used by almosteveryone today, are based on Marconi’s invention (theradiotelegraph system) in 1899 and the coding theorypresented by the then 32-year-old Shannon in 1948. Weaim to contribute to the area of informationcommunication technology — an area that has greatimportance in today’s society . In order to accomplishthis, we will continue exploring uncharted territories inour field, equipped with a vision that is a few decadesahead of its times.

An advanced vision for thedevelopment of new ICT

a: Single-photon state b: Superposition state of two waves c: Superposition state of two waves

Quadrature amplitude




Probability

Probability

Probability

Quadrature amplitude Quadrature amplitude

Phase [degree] Phase [degree] Phase [degree]

Wigner function distribution Wigner function distribution Wigner function distribution

The number of photons The number of photons The number of photons

Photon number distribution Photon number distribution Photon number distribution

The state of the electric field amplitude The state of the electric field amplitude The state of the electric field amplitude

Fig.1: Superposition of Odd Photon Number States (Conceptual Diagram)

Fig.2: Measurement of the Superposition of Odd Photon Number States (See the text forExplanation.)

PPKTP crystal

Laser beam ofwavelength430nm

Photon detector

Superposition of oddphoton numberstates at awavelength of860nm

Squeezed lightof wavelength860nm

Realization of 3DImaging of ActualMoving Objects byHolography


Makoto OkuiGroup Leader, 3DSpatial Image andSound Group,Universal MediaResearch Center

After completing his master’scourse, he joined the JapanBroadcasting Corporation in 1980and was engaged in research onstereoscopic television systems

etc. Since 2006, he has been involved with research on 3D imagingfor communication, particularly with 3D spatial images and sounds.He also has a Ph.D. in engineering.

●P r o f i l e●

Though everyone is familiar with the idea of three-dimensional imaging, it is still a relatively new concept.We have all seen or read about 3D images in sciencefiction movies and novels depicting the future. Some ofthe basic principles of 3D image displaying methodsdate back to more than a century. Although the basicidea is not new, high-utility 3D video imaging involvesadvanced technology that requires a great deal of noveldisplay and imaging methods; further research anddevelopment is required in order to develop suchmethods.

Three-dimensional images are very popularattractions in theme parks and events. It is also drawingattention as it opens up new business opportunities, fore.g., the production, distribution, and screening of 3Dfilms have now become commonplace, with possibilitiesfor family use as well. In 3D films, a display methodknown as stereoscopic display method is used; thismethod reproduces the effect of binocular parallax*.Special glasses are usually required for viewing suchdisplays.

Though the stereoscopic method is a practicalmethod, the extent to which the display method appearsnatural and realistic is insufficient for some of thepossible future applications. The Ministry of InternalAffairs and Communications is promoting research anddevelopment in 3D image that can be used forcommunication with remote areas, for educationalpurposes, or for use in various interfaces. In order torealize these aims, the extent to which a displayed imageappears realistic should be one that has so far not beenachieved. It is necessary to realize displays that facilitatethe viewing of 3D images from various angles, themanipulation of the images in combination with othersenses (for example, sense of touch), etc.

Besides binocular parallax mentioned earlier, thereare other visual clues used by our eyes for the

perception of depth. Successful reproduction of naturaland realistic images depends on how closely we canreproduce these visual clues. One of the imagingmethods that can reproduce most of the visual cluesrelating to the perception of depth is holography, and itis sufficiently realistic for many applications envisionedfor the future.

Other imaging methods such as photography recordthe intensity of light received from the object, whereasholography, by taking into account the wave nature oflight, creates and records what is known as aninterference fringe. By recording the interference fringe,information on the directions from which light beam isincident in addition to the intensity of light receivedfrom the subject can also be recorded and played back.Using the information thus obtained, 3D images can bereproduced. (Fig. 1a)

Typically, the size of the interference fringe is lessthan or equal to one micrometer. In order to record theinterference fringe, special dry plates (recordingmaterial) are used; these plates are capable of recordingthousands fringe patterns within a space of onemillimeter. Further, because holography is based on thewave nature of light, a special type of light (laser) isused for recording and reproducing images. For theholography of still images, considerable efforts havebeen made for the reproduction of color images andimages using normal light. As a result, the holographic3D image of a still image has attained an ultra-realisticquality that is not present in images created by other 3Dimaging methods.

Since holography is capable of producing ideal 3Dimages, expectations for the use of holography incommunication technology and future informationcommunication possibilities such as 3D television etc.are on the rise. In order to fulfill expectations, werequire an electronic system (electronic holography) thatcan (1) acquire, display, record, and transmit dataelectronically and (2) acquisition of actual movingobject data. However, the reproduction of wavefrontusing interference fringes, which is unique toholography, has remained confined to R&D activities inthe realm of basic research, and it is yet to reach thelevel of full-scale R&D of other 3D imaging methods.Therefore, in order to realize electronic holography,NICT is taking the initiative for the research of

“3D imaging” as a cutting-edge technology

Importance of holography for realistic 3Dimage reproduction is drawing attention

*Binocular parallax: When we view an object with both eyes, there are some differences in the images formed on the retina of our eyes due to the distance of the object. This is called binocularparallax, and it is one of the factors responsible for our perception of depth.


Fig.1:Data Acquisition and Reproduction ofHolographic Image(a) Conventional Holography (Still Image)(b) NICT’s Electronic Holography (ActualMoving Object)

Fig.2: Challenges Facing Electronic Holography

Fig.3:Experimental Apparatusof Holography and theReproduced Image

Subjects Holographic Reproduction

challenging issues that have to be resolved.

As shown in Fig. 2, there are many challenges to beovercome in order to develop a method for electronicholography. A narrow viewing zone and limitations inacquiring the real moving object data are the majorhurdles, and NICT considers these obstacles as subjectsof research. Let me talk about some of our researchachievements so far.

As mentioned earlier, holography requires a means ofrecording and displaying a fringe pattern that is less thanor equal to one micrometer in size. At present, there areno electronic devices with sufficient resolution todisplay such fine patterns, and instead, display deviceswith a pixel size of around 10m are used. Although thisis an order of magnitude coarser than the requiredresolution, 3D images can still be displayed by adjustingthe recording and displaying conditions; however, thereis a decrease in the size of the viewing zone.

The immediate goal of our research at NICT is theachievement of a viewing zone that is sufficiently widefor an image to be viewed as a 3D image with both eyesfrom a close range. The key to achieving this lies in theuse of high-density devices, technology to combinemultiple devices, and various signal processing methods

By using these techniques, we have succeeded increating a 3D image that can be viewed with both eyesfrom a distance of 40 cm.

Another area that we are researching is dataacquisition from actual moving objects. Theconventional method of data acquisition is to illuminatethe subject with the laser beam(in a darkroom with noother light source); however, the acquisition of datafrom various moving objects using this method is verydifficult. In order to solve this difficulty, we havedeveloped an unconstrained method of data acquisitioncalled integral photography that can be employed undernormal lighting conditions. This method uses a lensarray for data acquisition and then converts the data intoholographic data(Fig. 1b). At NICT, we have recentlysucceeded in performing the real-time conversion ofdata into holographic image and the display of the

holographic color image by using this method of dataacquisition. (Fig. 3)

We have to cover more ground before holography isput to practical use, but what we have achieved so fargoes a long way toward realizing its various futureapplications. Many issues such as the expansion of theviewing zone, the improvement of the image size andimage quality, etc., have to be ironed out.

Further, in order to accelerate the widespread use of3D images in society in the coming years, it is importantto promote the advantages of not just holography, but3D imaging as a whole; this will in turn encourageresearch in discovering the technical possibilities of 3Dimaging. The study from the point of view of thecontents is also indispensable. In this regard, we shallalso be making efforts to promote cooperation amongthe government, industry, and academia for the supportof Ultra-Realistic Communications Forum (URCF).

Efforts towards the realization of electronicholography Overcoming the hurdles

(a)Laser

Dry plate

Natural light

Lens array

High-definitioncamera Interference fringe

Conversion tointerferencefringe

◎Data acquisition from the actualmoving object is restricted

◎Unwanted lightobstructs the view

◎Viewing zone (viewable area)is insufficient◎Lack of size and resolution

in the reproduced image

Interference fringe

Image development

Recording ofinterference fringe

(b)

Display devices

Development ofSmall MobileSatellite Terminals


Shinichi YamamotoSenior Researcher, SpaceCommunication Group,New Generation WirelessCommunicationsResearch Center

He joined the Radio ResearchLaboratory, Ministry of Posts andTelecommunications (currentNICT) in 1975. He is currentlyengaged in research on mobile

satellite communications, tracking methods for automotive antenna ,etc., using ETS-V, EVS-VI, COMETS, and ETS-VIII.

●P r o f i l e●

Fig.1:Overview of Engineering Test Satellite ∆(ETS-∆) (Photo: Courtesy of JapanAerospace Exploration Agency)

Fig.2:Overview of the Portable Terminal (Weight:266g; Volume: 264cc; Battery not included)

During serious disasters, cellular phone services andfixed-line phone services may be disrupted, because ofwhich the need for an alternative means ofcommunication becomes necessary for efficient disastermanagement and effective execution of rescueoperations.

Satellite communication is highly effective duringdisasters or in places where no other means ofcommunication is possible, since it is not susceptible todamages caused on the ground and can be used as longas the satellite is visible. The equipment used forsatellite communication must be compact, lightweight,and portable.

The mobile satellite services that are currentlyavailable in Japan are Iridium, Widestar (NTT docomo),and Inmarsat. Iridium uses handheld terminals, whileWidestar and Inmarsat terminal use a rather large flatantenna about the size of a notebook PC.

An Engineering Test Satellite ∆“Kiku No.8” (ETS-∆) was launched from Tanegashima Space Center inDecember 2006. EST-∆ is equipped with two largedeployable antenna reflectors in the S-band (2.5/2.6GHz). Because these antennas could provide an equalperformance of a parabolic antenna of a diameter ofabout 13m on orbit; the miniaturization ofcommunication devices was enabled on the ground. Anoverview of the ETS-∆ is shown in Fig. 1, and that ofthe small mobile satellite terminal ( mobile terminal) is

shown in Fig. 2.During the development of this mobile terminal, we

miniaturized by using a ceramic patch antenna for atransmission and reception antenna and put it in theterminal body. Also, we have developed a new high-efficiency power amplifier to reduce powerconsumption.

The air interface of this mobile terminal is 8kbps,BPSK-SCPC; further, PSI-CELP (5.6 kbps) is used forvoice CODEC, and voice sound quality equivalent to anearly digital cellular phones has been obtained.

Calls can be made directly to the target terminal bykey inputs the ID number(same as dialing a telephonenumber) of the target terminal.

At NICT, we are advancing the communicationexperiment now using ETS-∆ and mobile terminals. Atpresent, we cannot use the large deployable antennareflector for reception in ETS-∆ because of the fault ofthe low-noise amplifier, this fault was detected duringthe initial checkout after the launch of the satellite. As asubstitute for this deployable antenna reflector, we areusing an antenna with a diameter of 1 m for HighAccuracy Clock (HAC).

To compensate for the receiving capability of thesatellite, we use a parabola antenna with a diameter of0.68 m in place of the ceramic patch antenna in themobile terminal.

We have modified the operational parameters of themobile terminals on the basis of the results of thesatellite communication experiments; we are now ableto establish the communication channel about 100% forsatisfactory voice communication.

The Ministry of Internal Affairs and Communicationshas entrusted NICT with research and developmenttoward realization of the Satellite / Terrestrial IntegratedMobile Communication System. Thereby, We started afive-year plan for the examination of practical use ofthis communication system. Please expect futuredevelopment of a mobile satellite phone system.

Mobile satellite phones today

Engineering Test Satellite ∆ (ETS-∆) andmobile terminals

Ongoing satellite communication experiment

Future prospects

Telecommunication services for secure and pleasant society — anytime, anywhere, and with anybody

NICTINTERVIEW


Ionosphere —A Part of Space Nearest the Earth

Continuous Observationsand Prediction ofIonospheric Disturbancesthat Can Affect RadioCommunications

The ionosphere is a layer located roughly between theatmosphere and the outer space, and it extends from aheight of approximately 60km to 1000km above the earth’ssurface. In the ionosphere, some of the atmosphericmolecules / atoms are separated into electrons and ions dueto the extreme ultraviolet radiation from the sun, resultingin the ionization state. The sudden change in the electrondensity in this region can degrade communications with thesatellite. NICT (and its predecessor organization) has beenengaged in observation of the ionosphere for more than 60years. Now we have four ionospheric observatories inJapan (Wakkanai, Kokubunji, Yamagawa, and Okinawa)and some observation instruments in Antarctica. Inaddition, we have installed many instruments in SoutheastAsian countries such as Thailand and Indonesia to observesevere ionospheric disturbances.

The expert researcher, Dr. Takuya Tsugawa, who says,“NICT’s facilities for the ionospheric observation are oneof the best in the world,” started his research on theionosphere during his graduate school days. He wasoriginally interested in the study of the planet Mars butlater shifted his focus to space research, which is moredirectly related to the day-to-day lives of people in theregion close to the earth. His graduate school adviserrecommended him to begin research studies on newly

developed ionospheric observations using GPS.

Since I joined the Space Environment Group of NICT, Ihave been carrying out observations of the equatorialionosphere using the observation facilities in SoutheastAsia countries. We examine the vertical profile of theelectron density by transmitting radio waves into theionosphere in different frequencies and measuring the timelag of echo back. “The observation facilities are located inremote areas, but visiting the actual observation site is aworthwhile experience.”

In the equatorial ionosphere, there is “equatorialanomaly”, a region where the electron density is very high.There also exists a “hole” with a very low electron densitycalled “plasma bubble”, which is generally formed aftersunset. “The equatorial plasma bubble develops along theearth’s arch-shaped magnetic field line. Sometimes it growsup into high altitudes in the equatorial region, resulting inthat the northern edge of the bubble reaches over Japan.This can degrade GPS navigations and cause ground-to-satellite communication failures” says Dr. TakuyaTsugawa. He is also engaged in research on the generationmechanism of these phenomena and the prediction of them.Research on the ionosphere, which plays an important rolein our daily lives, is of significance not only for the citizensof Japan but also for people residing in other countries.

Commencing ionospheric observation researchas a “service to society” Understanding and predicting abnormal

ionospheric phenomena that affect our daily lives

Takuya TsugawaExpert Researcher,Space Environment Group,Applied Electromagnetic Research Center

After having completed his doctoral course at GraduateSchool of Science, Kyoto University, and working as aJSPS Research Fellow, Dr. Takuya Tsugawa joined NICTin December 2007. He is currently engaged in research onprediction of ionospheric disturbances that can affectradio-wave propagation.

●P r o f i l e●


Transferable technology seedsoriginated by NICT were plainlyexplained in the presentation sessionfor technology seeds.

We held One Day Symposiumtitled“Promoting until Here onNew Generation Network andUniversal Communications.”

The autostereoscopic display system wasone of our most popular exhibits. Everydaythere was a long line of people queuing forthe exhibit. Behind the blackout curtain, the3D image emerged and with their nakedeyes, they could see it from different anglesby shifting their bodies.

Report

The stereophonic speaker which let them hear sound coming fromsurrounding directions, as in natural hearing, was another popular exhibit.

NICT Super Event 2008The Latest Research on ICT ― for People, for Earth, and for FutureNICT Super Event 2008The Latest Research on ICT ― for People, for Earth, and for FutureNICT Super Event 2008The Latest Research on ICT ― for People, for Earth, and for FutureNICT Super Event 2008The Latest Research on ICT ― for People, for Earth, and for FutureNICT Super Event 2008The Latest Research on ICT ― for People, for Earth, and for FutureNICT Super Event 2008The Latest Research on ICT ― for People, for Earth, and for FutureNICT Super Event 2008The Latest Research on ICT ― for People, for Earth, and for FutureNICT Super Event 2008The Latest Research on ICT ― for People, for Earth, and for FutureNICT Super Event 2008The Latest Research on ICT ― for People, for Earth, and for FutureNICT Super Event 2008The Latest Research on ICT ― for People, for Earth, and for FutureNICT Super Event 2008The Latest Research on ICT ― for People, for Earth, and for FutureNICT Super Event 2008The Latest Research on ICT ― for People, for Earth, and for FutureNICT Super Event 2008The Latest Research on ICT ― for People, for Earth, and for FutureNICT Super Event 2008The Latest Research on ICT ― for People, for Earth, and for FutureNICT Super Event 2008The Latest Research on ICT ― for People, for Earth, and for FutureNICT Super Event 2008The Latest Research on ICT ― for People, for Earth, and for FutureNICT Super Event 2008The Latest Research on ICT ― for People, for Earth, and for Future


Our inside exhibition booths were also crowded with visitors who werelistening to the researchers’presentation.

A visitor is being given a demonstrationof the brain activity measurement by aninstrument that uses near infrared light.

To promote our overall effort as NICT, we have held““NICT Super Event””again this year from September30 to October 4 at the CEATEC JAPAN 2008 site (Makuhari Messe). We held a One Day Symposium at theInternational Conference Hall, 61 exhibitions at the exhibition hall, and various presentations at the NICTbooth inside the exhibition hall, etc. It was a great success ―― a total of nearly 200,000 people came to theCEATEC during five days, and approximately 48,000 people visited NICT booths. Our booths weresurrounded by booths representing big companies, but the visitors seem to have liked the approach thatour researchers were presenting themselves.

■One Day Symposium:10:30 -16:40, Wednesday, Oct. 1

■Press Interview:14:00 -16:30, Tuesday, Sep. 30

■P r e s e n t a t i o n S e s s i o n o fTechnology Seeds:13:00 -16:30, Thursday, Oct. 2

OurEvents

The overlooked view of the NICT’s exhibition booths is shown.

■Basic Technology Promotion System for Private Sectors :11:00 -12:00, Thursday, Oct. 2 - Friday, Oct. 3

■Announcement of Government -Subsid izedProject Results:13:00 -16:30, Friday, Oct. 3

■Exhibition:10:00 -17:00, Tuesday, Sep. 30 - Saturday, Oct. 4


PRIZEWINNERS'PROFILES

PRIZE WINNER ●Katsumi FujiiSenior Researcher,Electromagnetic Compatibility Group, Applied ElectromagneticResearch Center

◎DATE：9.17.2008◎NAME OF THE WINNING PRIZE：CommunicationsSociety：Distinguished Contributions Award2008◎CONTENTS OF THE WINNING PRIZE：Great contribution to the activity and development of thetechnical study group in IEICE Technical Committee onElectromagnetic Compatibility (EMCJ) as (assistant) secretaryduring 2 years while planning and managing 20 meetings of thetechnical study group, 4 workshops of the second-classtechnical study group, the national convention, the societymeeting and the 30th anniversary project after creating thetechnical study group.

◎NAME OF GROUP：Communications Society,The Institute of Electronics,Information andCommunication Engineers

PRIZE WINNER ●Koumei SugiuraResearch Expert, Spoken Language Communication Group, Knowledge CreatingCommunication Research Center

◎DATE：8.20.2008◎NAME OF THE WINNING PRIZE：Category IncentivePrize of the Meeting for Electronics,Information and Systems Society◎CONTENTS OF THE WINNING PRIZE：A Method of Designing the Morphology of MobileRobots for Learning Multiple Tasks◎NAME OF GROUP：The Institute of ElectricalEngineers of Japan

◎Comment by the winner:I am very honored to have received the award as

incentive prize in the conference of “Electronics,Information and SystemsSociety in Fiscal 2007”. Iherewith, would like tothank everyone who hashelped me. This awardis truly encouraging me,and I want to pushforward with research inNICT.

PRIZE WINNER ●Ferdinand PeperGROUP OF PARTICIPATOR Li Jia

Senior Researcher,Nano ICT Group,Kobe Advanced ICT Research Center

◎DATE：7.1.2008◎NAME OF THE WINNING PRIZE：Best Paper Award◎CONTENTS OF THE WINNING PRIZE：On Brownian Cellular Automata◎NAME OF GROUP：Automata 2008:EPSRCWorkshop on Cellular Automaton Theory

◎Comment by the winner:

I am very happy to have received this award, which is theresult of years of in-depth discussions with my co-researcher,Dr. Lee Jia, on the fundamentals of computing on nanometerscales. We are presenting a novel thinking of probabilisticsearch with selections instead of the step-by-step method ofclassical deterministicapproach. We areplanning to expand ourresearch into thecommunication model ofmicrometer scale.


I was an assistant secretaryfor two years, but I had tospend the last six months ofthe two years in a hospital dueto an accident. So, I’m not sureof my contributions as anassistant secretary. Mr.Shinobu Ishigami (EMCGroup), who was the secretarythen, helped me greatly.

PRIZE WINNER ●

Okinawa Subtropical Environment Remote-Sensing Center

◎DATE：9.12.2008◎NAME OF THE WINNING PRIZE：Award of TheJapan Coast Guard Commandant in the137th Hydrographic Anniversary◎CONTENTS OF THE WINNING PRIZE：Contribution to Marine Information Affairs:Provision of "Marine Investigation" and "MarineInformation"◎NAME OF GROUP：Japan Coast Guard


We, as Okinawa Subtropical Environment Remote-Sensing Center, have set up long range ocean radars atIshigaki Island and Yonakuni Island, and observationand research of oceanic current, etc. are being carriedout. The data obtained are given to the 11th RegionalCoast Guard Headquarters and the Hydrographic andOceanographic Department of Japan Coast Guard to beused for understanding the oceanographic condition, etc.

USB Memory Stick Distributedas a 30-Year Anniversary Token

In the next issue, we will have an interview on the Strategy Headquarters of New Generation Network R&Daiming for the new generation network configuration.

Information for Readers:

A leap second will be inserted between 8:59:59 a.m. and 9:00:00 a.m.at Japan Standard Time (JST) on January 1, 2009 after three years.This leap second insertion will be the 24th after the adjustment by itsinsertion started in 1972.

On January 1, 2009, JST will be as the following

8:59:59 January 1, 2009 8:59:60 January 1, 2009 9:00:00 January 1, 2009

The insertion of a leap second

A leap second is one-second adjustment which keeps the difference between the astronomical

time the universal time based on the rotation and revolution of the heavenly body such as the earth

and the atomic time kept by atomic clock within ±0.9 second. All the leap second adjustments have

been done so far by inserting one second because the universal time has gotten slower in

comparison with the atomic time.

NICT is carefully preparing for the leap second adjustment while engaging in the service of

disseminating Japan Standard Time through transmitting the standard radio wave, etc.

Information on the leap second

In our daily lives, we had been using time called the astronomical time, which is based on the

earth's rotation on its axis and revolution around the sun. However, astronomical time changes in

irregular manner when examined in terms of a millisecond, so that it has not been suitable for today's

leading edjed scientific technology in high precision.

Therefore, about 50 years ago, the second was defined in terms of the time supplied by the atomic

clock, which uses an accurate electromagnetic wave emitted by an atom (the quantum theory). The

time we use today is based on the atomic clock with a high degree of accuracy.

Astronomical Time and Atomic Time

Nov. 2008 No. 374

Published by

Public Relations Office, Strategic Planning DepartmentNational Institute of Information and CommunicationsTechnology

Editorial Cooperation: Japan Space Forum

4-2-1 Nukui-Kitamachi, Koganei, Tokyo 184-8795, JapanTel: +81-42-327-5392 Fax: +81-42-327-7587 E-mail: [email protected]: http://www.nict.go.jp/index.html

tommy_vaio2

タイプライターテキスト

No.384 , Nov 2008

tommy_vaio2

タイプライターテキスト

what are the research topics of mastar project? · the beijing olympics under the mastar project,...

Documents