InterRobot Inc.

Home > Funded Research & Development

Funded Research & Development

We undertake funded development of various systems such as robots, CG characters, interface and communications.　The development includes hardware, software and applications as a package. You can make a one-stop shopping and can expect development of sophisticated system with shorter delivery time with less cost for communications. Shown below are some of the examples we have undertaken thus far.

Communication System

Human-persons make communications not by words alone. Their communication involves nodding and other body actions which enable sharing of rhythm among them and induce each other into the communication. This sharing of body actions produce sense of unity and make the participants feel that interaction with other person is real. iRT®, our patented technology, puts this mechanism into a model and generates body motion based solely on voice.

By introducing this iRT® into a communication system which employs robots or CG characters as an interface, smooth and human-friendly communication is realized.

Communication Robot (Yamagata Museum of Science & Industry)

　The robot explains the articles on display in conjunction with video. When a visitor talks to the robot through a microphone, the robot makes such communication reaction as nodding, blinking or other body language. Through this robot, a visitor can talk with another visitor or an information officer who is in an office in the separate building. Visitors can enjoy conversation with the robot as if it really understand what it says.

Comment:
　In addition to visitors, staff members of the museum also seem to enjoy using this system. Troubles with microphone are reported in relatively large number, which may be an indication of popular use of the system, but the breakdown of the system itself has not been reported. You may enjoy a strange and unusual feeling in having conversations through a robot.

Robot for Kindergartens

　This is a "Communication Robot for Infant" which we developed in collaboration with Kabushiki Kaisha Jakuestsu, one of the top class toy manufacturers in Japan. Information is shown on the display placed in the chest of the robot and the robot can convey the information by body motion which takes place in line with the voice. The display on the chest is allows touch panel operation and children can play games on it.
Through the incorporated microphone, children can talk to the robot. The robot can dance to music.
Comment:
Since this robot is used in kindergarten, we had to pay more attention to its safety features than in other applications. We were plagued by insufficient power of PC for the touch panel placed on the chest as well as its thermo runaway. An interesting finding we made is that children are fond of shaking hands with the robot.

Multiple　Mode Robot

　This is a demonstration of group communication composed of four robots and one CG system. They are modeled on a teacher and three pupils. When visitors talked to the microphone from both sides, the robots make body motion in response to the voice. From the looks of it, you may realize the effectiveness of body motion in communications within the group.(National Museum of Emerging Science & Innovation)

Comment:
　There are many robots in the museum. Visitors enjoy playing with them freely. Faces of the robots were made by Mr. Takashi Nishihira, a famous figurative artist.

Automatic Speech Dialogue

　This is a system to enable speech dialogue using voice recognition and voice synthesis. As for voice recognition engine and voice synthesis engine, we bought a SDK (software development kit) from other companies and used them in the system. Recognition rate has been improving thanks to development of voice recognition technology, but it is not perfect yet. In some places, error in recognition is unavoidable due to the inherent noise. Then, it becomes necessary to select proper voice input devices such as microphone. It is also necessary to exercise some ingenuity in setting vocabulary to be recognized, and this becomes know-how.

　In designing a system, conditions particular to the place of use must be taken into consideration in terms of recognition error or failure of recognition and countermeasures must be adopted. For instance, in a system for business use such as a reception terminal, recognition precision must be enhanced by employing interfaces as touch-panel. In amusement systems like a game, primary consideration is to avoid the situation with no reaction, which is possible when lower recognition precision is adopted.

Voice Synthesis (Text-to-Speech) makes it easier to formulate dialogue scenario. In the case of guidance or explanation, spoken lines can be fixed in advance. In such a case, recorded message can also be used. The use of recorded voice can sometimes enforce particular image of the character.

Reception Terminal (Robot)

　At the entrance hall of Coretech Inc., there is a reception robot system which we produced. It is equipped with voice recognition capability. If you tell the name of the person you want to meet to the microphone, the system shows you the extension number. As the system is equipped with voice synthesis , the extension number can be announced by voice, too. In case the voice recognition does not go well, a search of the extension number is also possible by using the touch-panel.

Comment:
　

Association Game Robot

　The robot listens to you with nodding. The robot can play the part of the speaker in response to the voice of visitors. The visitors can play the word association games with the robot by using voice dialogue function.("Tomarin-Kan").

Comment:
This is located at "Tomarin-Kan", Nuclear PR Center operated by Hokkaido Electric Power Company. It was no easy matter to make fine tuning for prevention of erroneous recognition.

Voice Dialogue Monument (L-chan)

　L-chan responds to specific spoken words and talks to you. It gives body action to the voice input through a microphone.(Osaka City Consumer Center)

Comment:
The voice dialogue system which we made was adopted in Osaka City Consumer Center, located in Asia Pacific Trade Center(ATC). The robot used there is equipped with fewer moving parts compared to robots in other places and accordingly shows simpler motion, but it vividly responds when you talk to it.

Nuclear Power Plant PR Robot

　The robot introduces various information with hand motion and dialogue. It makes an answer to a question by voice recognition. In the communication mode, a visitor positioned in front of the robot and the other visitor placed at the back of the robot can move the robot by dialogue. (Atom Plaza of Eldoland PR Pavilion by Kansai Electric Power Company).

Comment:
In this system, recorded voice was used instead of synthetic voice since peculiar characteristics was essential. The voice is directed either to the right or to the left depending on the position of the Character, thus enabling dialogue between the Characters. The microphone is in contact with the pedestal of the robot, which causes the microphone to catch motion of the robot and often resulted in erroneous motion of the system. Tuning of the voice recognition process was not easy in this system.

Sushi Chef Robot with voice dialogue system

　This is a robot takes orders from children through voice recognition and voice synthesis. The robot nods and talks with body motion when children talk to it. This robot was developed in collaboration with Kokoro Company Ltd. (Nagoya City Science Mseum)

Comment:
This robot was used by Nagoya City Science Museum for the event that was held for a limited time. Visitors can enjoy short conversation with the sushi-chef through voice recognition and voice synthesis technology (The robot does not make sushi, though). The robot itself is S Doll made by Kokoro Company Ltd. We employed some twists including insulation against noise by partitioning the booth for higher voice recognition. Some tuning was also necessary to better catch voice of children. At first, the system hardly stabilized but it suddenly became steady when we changed the OS to Windows 2000.

Information Terminals

InterRobot Inc. has made many CG characters as dialogue-type interface. Various development program breathed life into those CG characters and they have been doing a great job in variety of applications. Please feel free to make your inquiry.

Infotalk

　This is our original browsing software.
　Infotalk is an application software developed asa human-friendly information providing tool. It runs on generic personal computers. Viewing area is divided into web-based information display area, the area for characters which act as agents and characters display area which is placed at the bottom. The synthesized voice which is related to each page or the recorded voice (in WAVE form) comes out of the speaker, and the characters show vivid performance in harmony with the voice.


Comment:

InterCaster (Program Production Tool)

　InterCaster is an application system where CG characters are activated by iRT technology specifically for video production. By adding an image layer produced by this system on a video image, more attractive image contents can be created. To the sound that is input from a microphone or Line-In, the CG character generates real-time motion as if it is actually speaking. In addition, it makes specific actions (bowing and sketching a wave of farewell, etc) by a key-input. These functions enables easy production of animation video employing vivid and attractive CG characters.

Case examples:
RSK Sanyo Broadcasting Co., Ltd.: Program production
KDDI: Program production compatible with broad-band
Others


Comment:

CG characters

Our company thus far has produced many CG characters for dialogue interface. The CG characters we produce are given lives with various development program we provide, and they can act vividly and do a great job in different applications. Please do not hesitate to make an inquiry to us.

Car Navigation System (Demonstration System)

　InterActor ( a CG character activated by iRT) was employed in voice processing part of the next generation car navigation system which Pioneer Corporation announced in "ITS World Conference in Aichi/Nagoya 2004".

Comment:
　We made the display interface with FLASH. It enabled to have experience in operating map handling and searching system.
Comments were affirmative when they were　evaluated in the event site. However, a problem peculiar to vehicle-mounted equipment was pointed out, namely, "Isn’t driving obstructed by displaying the character?" Thus, in its commercialization, consideration to smaller display area and display method other than character is necessary.

Linux-version iRT License

　We delivered Linux-version "iRT SDK" to Kihanna Human Info-Communication Research Center, Communications Research Laboratory of National Institution of Informatics.
We delivered "iRT SDK" to KDDI Corporation for their consideration to employ it in their trial FTTH service using "Set-Top Box(STB)".

Comment:
　Our regular “iRT SDK” is of Windows version, but in this case we developed Linux version SDK upon their specific request.

Others

We are engaged in research and development for various other fields. We undertake funded development as well as joint development.

Interpicture

　Interpicture is an interactive art which was developed by focusing on the healing effect ( in the verification process at present) of iRT. When placed on a wall as a picture, Interpicture can create sober atmosphere by its synchronized motion with the environment.

Comment:
　Sunflower shows bodily movement in response to the nearby sound. The background can be replaced with whatever picture you like.

Kleenex distribution robot(Mospen-Kun2)

This is a kleenex distribution robot, jointly developed with Compy Inc. A monitor is used for the part of face, thus giving facial expression. The robot senses approach of a person via infrared ray sensor,then turns toward the person and passes out tissues.

Comments:
　This is a somewhat different robot in that it passes out tissues. It is employed in events to attract visitors. People tend to refuse receiving tissues if they are passed out by human, but are willing to receive them if handed by a robot. People feel that the robot pays attention to them by such casual action on the part of the robot when it turns toward them.