JP2013114107A - Communication system, utterance content generation device, utterance content generation program, and utterance content generation method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To enable easy generation of utterance content in a confirming action.SOLUTION: A communication system 10 includes a robot 12 and a server 20. The robot specifies an article specified by a user. The server performs a pointing action for confirming the specified article and generates utterance content in response to a request from the robot. In generating the utterance content, the server generates a power set on plural attributes of the specified article and a power set on plural attributes of other articles existing nearby the specified article for each of the aforementioned other articles. The server deletes an element common with the power set on each of the other articles from an element of the power set on the specified article. The server selects elements with the least number of words from among the remaining elements and selects an element with the highest non-similarity with the character string of plural attributes of the other articles existing nearby the specified article from among the remaining elements and generates utterance content.

Description

  The present invention relates to a communication system, an utterance content generation device, an utterance content generation program, and an utterance content generation method. In particular, for example, an article designated by a person is identified by voice recognition, and the identified article is designated by the person. The present invention relates to a communication system, an utterance content generation device, an utterance content generation program, and an utterance content generation method.

  An example of this type of conventional communication system is disclosed in Patent Document 1. In the communication system disclosed in Patent Document 1, when the server specifies an article designated by a person, a “specific word” that identifies the article is used to determine whether the article is an article designated by a person. The robot emits a confirmation voice. At this time, as the “specific word”, a word that is not used to specify another article in the vicinity of a human and has a high recognition rate by voice recognition is selected.

JP 2009-223171 [G10L 15/22, G10L 15/00]

  However, in the communication system of Patent Document 1, it is necessary to obtain a recognition rate of speech recognition in advance for a word used as a specific word, which is troublesome.

  Therefore, a main object of the present invention is to provide a novel communication system, utterance content generation device, utterance content generation program, and utterance content generation method.

  Another object of the present invention is to provide a communication system, an utterance content generation device, an utterance content generation program, and an utterance content generation method that can easily select words to be included in the utterance content.

  The present invention employs the following configuration in order to solve the above problems. The reference numerals in parentheses, supplementary explanations, and the like indicate correspondence relationships with embodiments described later to help understanding of the present invention, and do not limit the present invention in any way.

  A first invention is a communication system for identifying an article designated by a person by voice recognition and confirming by voice whether the identified article is designated by the person, the name of the article and the article A storage unit that stores a plurality of words related to the first word, a first generation unit that reads a plurality of words related to the identified article from the storage unit, and creates a first power set for the read word; A plurality of words related to other articles to be read from the storage means, and a second power set for the read articles is created for each other article, a first power set created by the first creation means And a deletion means for deleting elements common to each of the second power sets created by the second creation means from the first power set. As a result of the deletion, the first selection means for selecting the element with the smallest number of words among the elements included in the first power set, and the plurality of elements related to other articles among the elements selected by the first selection means The second selection means for selecting the element having the highest degree of dissimilarity with the character string of the word, and the utterance content when confirming by voice including the word included in the element selected by the second selection means It is a communication system provided with the utterance content production | generation means to produce | generate.

  In the first invention, the communication system (10) specifies an article designated by a human by voice recognition, and confirms by voice whether the identified article is designated by the human. The storage means (122, 204) stores the name of the article and a plurality of words related to the article. The plurality of words related to the article are words regarding attributes such as the type, color, and thickness of the article, and are words that supplementarily describe the article. The first creation means (200, S33) reads a plurality of words related to the identified article from the storage means, and creates a first power set for the read words. The second creation means (200, S35) reads a plurality of words related to other articles existing in the vicinity of the identified article from the storage means, and creates a second power set for the read words for each of the other articles. To do. The first deleting means (200, S37) determines the common elements between the first power set created by the first creating means and the second power set created by the second creating means. Remove from set. The first selection means (200, S39) selects the element having the smallest number of words among the elements included in the first power set as a result of deletion by the deletion means. A 2nd selection means (200, S41) selects the element with the highest dissimilarity with the character string of the several word relevant to another article among the elements selected by the 1st selection means. The utterance content generation means (200, S13) generates the utterance contents when confirming by voice including the word included in the element selected by the second selection means.

  According to the first invention, unlike the article existing in the vicinity of the specified article, the character string having the highest dissimilarity of the character string is selected, so it is not necessary to obtain the speech recognition rate of the word in advance, and the utterance You can easily select words to include in the content. Therefore, the utterance content can be easily generated.

  The second invention is dependent on the first invention, and each of the character string of the first word included in the element selected by the first selecting means and the character strings of the plurality of second words related to other articles And a second selection unit selects an element including the first word that maximizes the Levenshtein distance calculated by the calculation unit.

  In the second invention, the communication system further includes calculation means (200, S91). The calculating means calculates the Levenshtein distance between the character string of the first word included in the element selected by the first selecting means and each of the character strings of the plurality of second words related to other articles. That is, the similarity of the character string (pronunciation) is calculated. The second selection means selects an element including the first word that maximizes the Levenshtein distance calculated by the calculation means. Therefore, as described above, the element having the highest dissimilarity with the character strings of a plurality of words related to other articles is selected.

  According to the second invention, since only the Levenshtein distance between character strings is calculated, it is easy to select a word.

  The third invention is dependent on the first or second invention, and is a first judging means for judging whether or not another article exists in the vicinity of the specified article, and in the vicinity of the article specified by the first judging means. When it is determined that there are no other articles, candidate determination means for determining each of a plurality of words related to the specified article as candidates to be included in the utterance content, and a human being among the words determined by the candidate determination means Further comprising third selection means for selecting a word having the highest degree of dissimilarity with the character strings of a plurality of words related to other articles existing in the vicinity, and the utterance content generation means is selected by the third selection means The utterance contents when confirming by voice including the words included in the elements are generated.

  In the third invention, the communication system further includes first determination means (200, S31), candidate determination means (200, S43), and third selection means (200, S51). The first determination means determines whether another article exists in the vicinity of the specified article. When it is determined that no other article exists in the vicinity of the article specified by the first determination means, the candidate determining means determines each of a plurality of words related to the specified article as a candidate to be included in the utterance content. The third selection means selects the word having the highest dissimilarity with the character strings of a plurality of words related to other articles existing in the vicinity of the human among the words determined by the candidate determination means. In such a case, the utterance content generation unit generates the utterance content when confirming by voice including the word included in the element selected by the third selection unit.

  According to the third invention, even if there is no other article in the vicinity of the specified article, an appropriate word included in the utterance content can be selected.

  The fourth invention is dependent on the third invention, and when it is determined that there is no other article in the vicinity of the article specified by the first determination means, it is determined whether or not there is another article in the vicinity of the human. When the second judging means for judging and the second judging means judges that there is no other article in the vicinity of the human, one word is selected from the candidates determined by the candidate determining means according to a predetermined rule. The speech content generating means generates speech content when confirming by voice including the word included in the element selected by the fourth selection means.

  In the fourth invention, the communication system further includes second determination means (200, S45) and fourth selection means (200, S53). When it is determined that there is no other article in the vicinity of the article specified by the first determination means, the second determination means determines whether there is another article in the vicinity of the human. The fourth selection unit selects one word from the candidates determined by the candidate determination unit according to a predetermined rule when it is determined by the second determination unit that there is no other article in the vicinity of the person. For example, one word is randomly selected from the candidates, or one related word determined in advance by the system administrator or the like is selected. In such a case, the utterance content generation unit generates the utterance content when confirming by voice including the word included in the element selected by the fourth selection unit.

  In the fourth invention as well, as in the third invention, it is possible to select an appropriate word included in the utterance content even when there is no other article in the vicinity of the specified article.

  The fifth invention is used in a communication system for identifying an article designated by a person by voice recognition and confirming by voice whether the specified article is designated by the person. Utterance content generation apparatus for generating the utterance content of the storage device, the storage means for storing the name of the article and a plurality of words related to the article, the plurality of words related to the specified article is read from the storage means, and the read word First creation means for creating a first power set for each of the other articles, a plurality of words relating to other articles existing in the vicinity of the identified article are read from the storage means, Between the first power set created by the second creation means, the first power set created by the first creation means, and the second power set created by the second creation means A deletion unit that deletes a common element from the first power set; a first selection unit that selects an element having the smallest number of words among elements included in the first power set as a result of deletion by the deletion unit; Of the elements selected by the first selecting means, the second selecting means for selecting the element having the highest degree of dissimilarity with the character strings of a plurality of words related to other articles, and the element selected by the second selecting means It is an utterance content production | generation apparatus provided with the utterance content production | generation means which produces | generates the utterance content at the time of confirming with a sound including the word contained in.

  A sixth invention is used in a communication system for identifying an article designated by a person by voice recognition and confirming by voice whether the identified article is designated by the person. The name of the article and the article An utterance content generation program of an utterance content generation device for generating utterance content for confirmation by voice, comprising storage means for storing a plurality of words related to the item, the article identified by the processor of the utterance content generation device A first creation step for reading a plurality of words related to the item from the storage unit, and generating a first power set for the read word, and reading and reading a plurality of words relating to other articles existing in the vicinity of the specified article from the storage unit Created in the second creation step and the first creation step of creating a second power set for each of the other articles. The first power set as a result of deleting elements common between the first power set and each of the second power sets created in the second creation step from the first power set, and the deletion step A first selection step of selecting an element having the smallest number of words among the elements included in, and dissimilarity to character strings of a plurality of words related to other articles among the elements selected in the first selection step A second selection step for selecting the element having the highest degree, and an utterance content generation step for generating an utterance content for confirming by voice including words included in the element selected in the second selection step This is an utterance content generation program.

  The seventh invention is used in a communication system for identifying an article designated by a person by voice recognition and confirming by voice whether the identified article is designated by the person. The name of the article and the article The utterance content generation method of the utterance content generation device includes a storage means for storing a plurality of words related to the utterance, and generates the utterance content when confirming by speech. The processor of the utterance content generation device includes: A plurality of words related to the specified article are read from the storage means, a first power set is created for the read words, and (b) a plurality of words related to other articles existing in the vicinity of the specified article is read from the storage means. The second power set for the read word is created for each of the other articles, and (c) the first power set created in step (a) and the step ( The elements common to each of the second power sets created in step) are deleted from the first power set, and (d) as a result of the deletion in step (c), the elements included in the first power set are deleted. Among them, the element having the smallest number of words is selected, and (e) the element having the highest dissimilarity with the character strings of a plurality of words related to other articles among the elements selected in step (d). (F) An utterance content generation method for generating an utterance content for confirmation by voice including the word included in the element selected in step (e).

  In the fifth to seventh inventions, as in the first invention, the utterance content can be easily generated.

  According to the present invention, unlike the article existing in the vicinity of the identified article, the character string having the highest dissimilarity of the character string is selected. You can easily select words to include. Therefore, the utterance content can be easily generated.

  The above object, other objects, features and advantages of the present invention will become more apparent from the following detailed description of embodiments with reference to the drawings.

FIG. 1 is an illustrative view showing an outline of a communication system showing an embodiment of the present invention. FIG. 2 is an illustrative view showing the appearance of the robot shown in FIG. 1 from the front. FIG. 3 is a block diagram showing an electrical configuration of the robot shown in FIG. FIG. 4 is a block diagram showing an electrical configuration of the server shown in FIG. FIG. 5 is an illustrative view showing one example of an article dictionary used in the embodiment of FIG. FIG. 6 is an illustrative view showing one example of an article local dictionary used in the embodiment of FIG. FIG. 7 is an illustrative view showing one example of a list of words used in the embodiment of FIG. FIG. 8 is an illustrative view for explaining a method of determining a word used in the confirmation action of the embodiment of FIG. FIG. 9 is an illustrative view showing one example of a memory map of the RAM shown in FIG. FIG. 10 is a flowchart showing the confirmation action determination process of the CPU shown in FIG. FIG. 11 is a flowchart showing the word selection processing of the CPU shown in FIG. FIG. 12 is a flowchart showing a process of generating the attribute set of the article X of the CPU shown in FIG. FIG. 13 is a flowchart showing processing for generating an attribute set of another article in the vicinity of the CPU shown in FIG. FIG. 14 is a flowchart showing processing for reducing the attribute group of the article X of the CPU shown in FIG. FIG. 15 is a flowchart showing the CPU word determination (1) processing shown in FIG. FIG. 16 is a flowchart showing processing for generating a candidate for the name of the article X of the CPU shown in FIG. FIG. 17 is a flowchart showing the CPU word determination (2) process shown in FIG.

  Referring to FIG. 1, a communication system (hereinafter simply referred to as “system”) 10 of this embodiment includes a communication robot (hereinafter simply referred to as “robot”) 12. The robot 12 can communicate with humans and other robots using at least one of voice and body movement (gesture). The robot 12 can access the server 20 via a network 14 such as a wireless LAN. The robot 12 of this embodiment cooperates with the server 20 to identify an article that the human 16 indicates with voice, line of sight, and pointing, and performs an operation such as taking the article to the human 16. .

  A wireless tag 18 is attached to the person 16 and a marker for motion capture is attached, although not shown. The wireless tag 18 transmits a radio signal including identification information. Here, the identification information included in the radio signal is used to individually identify the person 16. The markers are typically set on the top of the human head, both shoulders, both elbows, the tip of the index finger of both hands, etc., and these markers together with the entire human 16 are controlled by the camera 120 controlled by the server 20. Taken. In the embodiment, three cameras 120 are provided, take a picture of the human 16 from three directions, and supply data about the camera video (camera video data) to the server 20.

  The server 20 is connected to the network 14 and executes a motion capture process for detecting the movement of the marker based on the camera video data input as described above. The position of the face can be specified.

  In the system 10, as described above, the robot 12 specifies an article designated by the human 16 as an object. In this embodiment, a book (book) 24 is used as an example of an article that can be an object. The book 24 (indicated as “OBJ” in FIG. 1) is attached with a wireless tag 18 that includes information (identification information) that can identify the book in a radio wave signal and transmits the information.

  However, the articles that can be the object are not only the books of the embodiment, but any household article can be considered if it is a home system. Naturally, it can be used not only for home use but also in any place (company, office, factory, etc.) that works with people. In such a case, various articles existing in any place can be the object.

  Information about all articles (books 24) targeted by the system 10 is registered in an article dictionary database (DB) 122 connected to the server 20. The article dictionary DB 122 will be described later.

  Further, identification information transmitted from the wireless tag 18 attached to the person 16 targeted by the system 10 is read by the wireless tag reader 208 (see FIG. 4) via any one of the plurality of antennas 124. The identification information read by the wireless tag reader 208 is given to the server 20. In response to this, the server 20 identifies the person 116 from the identification information, and grasps (detects) a slightly rough position of the person 16 from the position where the antenna 124 that receives (reads) the identification information is disposed. .

  In FIG. 1, one robot 12 is shown for simplicity, but two or more robots may be used. Further, the human 16 is not necessarily limited to one person, and can be identified by the wireless tag 18 and may be plural.

  In the embodiment shown in FIG. 1, the positions of the robot 12, the human 16, the article 24, and the like are expressed using the world coordinates of the space where the system 10 is installed. Since it is performed in coordinates, details will not be described, but the robot 12 performs a coordinate conversion process between the robot coordinates and the world coordinates as necessary in the process described later.

  The hardware configuration of the robot 12 will be described with reference to FIG. FIG. 2 is a front view showing the appearance of the robot 12 of this embodiment. The robot 12 includes a carriage 30, and two wheels 32 and one slave wheel 34 for autonomously moving the robot 12 are provided on the lower surface of the carriage 30. The two wheels 32 are independently driven by a wheel motor 36 (see FIG. 3), and the carriage 30, that is, the robot 12 can be moved in any direction, front, back, left, and right. The slave wheel 34 is an auxiliary wheel that assists the wheel 32. Therefore, the robot 12 can move in the arranged space by autonomous control.

  A cylindrical sensor attachment panel 38 is provided on the carriage 30, and a large number of infrared distance sensors 40 are attached to the sensor attachment panel 38. These infrared distance sensors 40 measure the distance to the sensor mounting panel 38, that is, the object (human being, obstacle, etc.) around the robot 12.

  In this embodiment, an infrared distance sensor is used as the distance sensor, but an ultrasonic distance sensor, a millimeter wave radar, or the like can be used instead of the infrared distance sensor.

  A body 42 is provided on the sensor mounting panel 38 so as to stand upright. Further, the above-described infrared distance sensor 40 is further provided in the upper front upper portion of the body 42 (a position corresponding to a human chest), and measures the distance mainly to a human in front of the robot 12. Further, the body 42 is provided with a support column 44 extending from substantially the center of the upper end of the side surface, and an omnidirectional camera 46 is provided on the support column 44. The omnidirectional camera 46 photographs the surroundings of the robot 12 and is distinguished from an eye camera 70 described later. As this omnidirectional camera 46, for example, a camera using a solid-state imaging device such as a CCD or a CMOS can be adopted. In addition, the installation positions of the infrared distance sensor 40 and the omnidirectional camera 46 are not limited to the portions, and can be changed as appropriate.

  An upper arm 50R and an upper arm 50L are provided at upper end portions on both sides of the torso 42 (position corresponding to a human shoulder) by a shoulder joint 48R and a shoulder joint 48L, respectively. Although illustration is omitted, each of the shoulder joint 48R and the shoulder joint 48L has three orthogonal degrees of freedom. That is, the shoulder joint 48R can control the angle of the upper arm 50R around each of three orthogonal axes. A certain axis (yaw axis) of the shoulder joint 48R is an axis parallel to the longitudinal direction (or axis) of the upper arm 50R, and the other two axes (pitch axis and roll axis) are orthogonal to the axes from different directions. It is an axis to do. Similarly, the shoulder joint 48L can control the angle of the upper arm 50L around each of three orthogonal axes. A certain axis (yaw axis) of the shoulder joint 48L is an axis parallel to the longitudinal direction (or axis) of the upper arm 50L, and the other two axes (pitch axis and roll axis) are orthogonal to the axes from different directions. It is an axis to do.

  In addition, an elbow joint 52R and an elbow joint 52L are provided at the respective distal ends of the upper arm 50R and the upper arm 50L. Although illustration is omitted, each of the elbow joint 52R and the elbow joint 52L has one degree of freedom, and the angle of the forearm 54R and the forearm 54L can be controlled around the axis (pitch axis).

  At the tip of each of the forearm 54R and the forearm 54L, a hand 56R and a hand 56L corresponding to a human hand are provided. Although the detailed illustration is omitted, these hands 56R and 56L are configured to be openable and closable so that the robot 12 can grip or hold an object using the hands 56R and 56L. However, the shape of the hands 56R and 56L is not limited to the shape of the embodiment, and may have a shape and a function very similar to a human hand.

  Although not shown, the front surface of the carriage 30, the portion corresponding to the shoulder including the shoulder joint 48R and the shoulder joint 48L, the upper arm 50R, the upper arm 50L, the forearm 54R, the forearm 54L, the sphere 56R, and the sphere 56L, A contact sensor 58 (shown generically in FIG. 3) is provided. A contact sensor 58 on the front surface of the carriage 30 detects contact of a person or another obstacle with the carriage 30. Therefore, the robot 12 can detect the contact with the obstacle during its movement and immediately stop the driving of the wheel 32 to suddenly stop the movement of the robot 12. Further, the other contact sensors 58 detect whether or not the respective parts are touched. In addition, the installation position of the contact sensor 58 is not limited to the said site | part, and may be provided in an appropriate position (position corresponding to a person's chest, abdomen, side, back, and waist).

  A neck joint 60 is provided at the upper center of the body 42 (a position corresponding to a person's neck), and a head 62 is further provided thereon. Although illustration is omitted, the neck joint 60 has a degree of freedom of three axes, and the angle can be controlled around each of the three axes. A certain axis (yaw axis) is an axis directed directly above (vertically upward) of the robot 12, and the other two axes (pitch axis and roll axis) are axes orthogonal to each other in different directions.

  The head 62 is provided with a speaker 64 at a position corresponding to a human mouth. The speaker 64 is used for the robot 12 to communicate with humans around it by voice or sound. A microphone 66R and a microphone 66L are provided at a position corresponding to a human ear. Hereinafter, the right microphone 66R and the left microphone 66L may be collectively referred to as a microphone 66. The microphone 66 captures ambient sounds, in particular, the voices of humans who are subjects of communication. Furthermore, an eyeball part 68R and an eyeball part 68L are provided at positions corresponding to human eyes. The eyeball portion 68R and the eyeball portion 68L include an eye camera 70R and an eye camera 70L, respectively. Hereinafter, the right eyeball part 68R and the left eyeball part 68L may be collectively referred to as the eyeball part 68. The right eye camera 70R and the left eye camera 70L may be collectively referred to as an eye camera 70.

  The eye camera 70 captures a human face approaching the robot 12, other parts or objects, and captures a corresponding video signal. In this embodiment, the robot 12 detects the line-of-sight directions (vectors) of the left and right eyes of the human 16 from the video signal from the eye camera 70. Specifically, the line-of-sight detection method is disclosed in Japanese Patent Application Laid-Open No. 2004-255074 as using two cameras, and Japanese Patent Application Laid-Open No. 2006-172209 and Japanese Patent Application Laid-Open No. 2006-285531 as using one camera. However, the details are not important here, so only those publications are cited.

  However, a well-known eye mark recorder or the like may be used for detecting the line-of-sight vector of the human 16.

  The eye camera 70 can be the same camera as the omnidirectional camera 46 described above. For example, the eye camera 70 is fixed in the eyeball unit 68, and the eyeball unit 68 is attached to a predetermined position in the head 62 via an eyeball support unit (not shown). Although illustration is omitted, the eyeball support portion has two degrees of freedom, and the angle can be controlled around each of these axes. For example, one of the two axes is an axis (yaw axis) in a direction toward the top of the head 62, and the other is orthogonal to the one axis and goes straight in a direction in which the front side (face) of the head 62 faces. It is an axis (pitch axis) in the direction to be performed. By rotating the eyeball support portion around each of these two axes, the tip (front) side of the eyeball portion 68 or the eye camera 70 is displaced, and the camera axis, that is, the line-of-sight direction is moved. Note that the installation positions of the speaker 64, the microphone 66, and the eye camera 70 described above are not limited to those portions, and may be provided at appropriate positions.

  As described above, the robot 12 of this embodiment includes independent two-axis driving of the wheels 32, three degrees of freedom of the shoulder joint 48 (6 degrees of freedom on the left and right), one degree of freedom of the elbow joint 52 (2 degrees of freedom on the left and right), It has a total of 17 degrees of freedom, 3 degrees of freedom for the neck joint 60 and 2 degrees of freedom for the eyeball support (4 degrees of freedom on the left and right).

  FIG. 3 is a block diagram showing the electrical configuration of the robot 12. With reference to FIG. 3, the robot 12 includes a CPU 80. The CPU 80 is also called a microcomputer or a processor, and is connected to the memory 84, the motor control board 86, the sensor input / output board 88 and the audio input / output board 90 via the bus 82.

  The memory 84 includes a ROM, an HDD, and a RAM (not shown). In the ROM and the HDD, a control program for controlling the operation of the robot 12 is stored in advance. For example, a detection program for detecting the output (sensor information) of each sensor, a communication program for transmitting / receiving necessary data and commands to / from an external computer, and the like are recorded. The RAM is used as a work memory or a buffer memory.

  Furthermore, in this embodiment, the robot 12 is configured to be able to speak and make a gesture to communicate with the human 16, but the memory 84 has a speech for such a speech and gesture. / Gesture dictionary 85A is set.

  The motor control board 86 is constituted by, for example, a DSP, and controls driving of each axis motor such as each arm, neck joint, and eyeball unit. That is, the motor control board 86 receives control data from the CPU 80, and controls two motors (collectively indicated as “right eyeball motor 92” in FIG. 3) that control the angles of the two axes of the right eyeball portion 68R. Control the rotation angle. Similarly, the motor control board 86 receives control data from the CPU 80, and controls two angles of the two axes of the left eyeball portion 68L (in FIG. 3, collectively referred to as “left eyeball motor 94”). ) To control the rotation angle.

  The motor control board 86 receives control data from the CPU 80, and includes a total of four motors including three motors for controlling the angles of the three orthogonal axes of the shoulder joint 48R and one motor for controlling the angle of the elbow joint 52R. The rotation angle of two motors (collectively indicated as “right arm motor 96” in FIG. 3) is controlled. Similarly, the motor control board 86 receives control data from the CPU 80, and includes three motors for controlling the angles of the three orthogonal axes of the shoulder joint 48L and one motor for controlling the angle of the elbow joint 52L. The rotation angles of a total of four motors (collectively indicated as “left arm motor 98” in FIG. 3) are controlled.

  Further, the motor control board 86 receives control data from the CPU 80, and controls three motors that control the angles of the three orthogonal axes of the neck joint 60 (in FIG. 3, collectively indicated as “head motor 100”). Control the rotation angle. The motor control board 86 receives control data from the CPU 80 and controls the rotation angles of the two motors (collectively indicated as “wheel motor 36” in FIG. 3) that drive the wheels 32.

  A hand actuator 108 is further coupled to the motor control board 86, and the motor control board 86 receives control data from the CPU 80 and controls the opening and closing of the hands 56R and 56L.

  In this embodiment, a motor other than the wheel motor 36 uses a stepping motor (that is, a pulse motor) in order to simplify the control. However, a DC motor may be used similarly to the wheel motor 36. The actuator that drives the body part of the robot 12 is not limited to a motor that uses a current as a power source, and may be changed as appropriate. For example, in another embodiment, an air actuator may be applied.

  Similar to the motor control board 86, the sensor input / output board 88 is configured by a DSP and takes in signals from each sensor and gives them to the CPU 80. That is, data relating to the reflection time from each of the infrared distance sensors 40 is input to the CPU 80 through the sensor input / output board 88. The video signal from the omnidirectional camera 46 is input to the CPU 80 after being subjected to predetermined processing by the sensor input / output board 88 as necessary. Similarly, the video signal from the eye camera 70 is also input to the CPU 80. Further, signals from the plurality of contact sensors 58 described above (collectively indicated as “contact sensors 58” in FIG. 3) are provided to the CPU 80 via the sensor input / output board 88. Similarly, the voice input / output board 90 is also configured by a DSP, and voice or voice in accordance with voice synthesis data provided from the CPU 80 is output from the speaker 64. In addition, voice input from the microphone 66 is given to the CPU 80 via the voice input / output board 90.

  The CPU 80 is connected to the communication LAN board 102 via the bus 82. The communication LAN board 102 is configured by a DSP, for example, and provides transmission data provided from the CPU 80 to the wireless communication device 104, and the wireless communication device 104 transmits the transmission data to the server 20 via the network 14. In addition, the communication LAN board 102 receives data via the wireless communication device 104 and provides the received data to the CPU 80. For example, the transmission data may be a signal (command) from the robot 12 to the server 20, or operation history information (history data) regarding communication performed by the robot 12. The reason why the history data is transmitted as well as the command is to reduce the capacity of the memory 84 and to reduce power consumption. In this embodiment, the history data is transmitted to the server 20 every time communication is performed. However, the history data may be transmitted to the server 20 in units of a fixed time or a fixed amount.

  Further, the wireless tag reader 106 is connected to the CPU 80 via the bus 82. The wireless tag reader 106 receives a radio wave superimposed with identification information transmitted from the wireless tag 18 (RFID tag) via an antenna (not shown). Then, the RFID tag reader 106 amplifies the received radio wave signal, separates the identification signal from the radio wave signal, demodulates (decodes) the identification information, and supplies the identification information to the CPU 80. According to FIG. 1, the wireless tag 18 is attached to a person 16 or an article (in this embodiment, book 24) in the reception of the company where the robot 12 is disposed or in the living room of a general household, and the wireless tag reader 106 is The radio signal transmitted from the wireless tag 18 within the communicable range is detected.

  Note that the wireless tag 18 may be an active type or a passive type that is driven according to a radio wave transmitted from the wireless tag reader 106.

  The hardware configuration of the server 20 will be described with reference to FIG. As shown in FIG. 4, the server 20 includes a CPU 200. The CPU 200, also called a processor, is connected to the memory 204, the camera control board 206, the wireless tag reader 208, the LAN control board 210, the input device control board 212, and the monitor control board 214 via the bus 202.

  The CPU 200 governs overall control of the server 20. The memory 204 comprehensively shows ROM, RAM, HDD, and the like, and records a program for operating the server 20 and functions as a work area when the CPU 200 operates. The camera control board 206 is for controlling the camera 120 connected to the control board 206.

  The wireless tag reader 208 receives a radio wave superimposed with identification information transmitted from the wireless tag 18 attached to the person 16 or the article (book) 24 via the antenna 124 connected to the control board 208. Then, the RFID tag reading device 208 amplifies the received radio wave signal, separates the identification signal from the radio wave signal, demodulates (decodes) the identification information, and gives it to the CPU 200. The antenna 124 is placed throughout the reception of the company where the robot 12 is placed, and in each room of a general household, and receives radio waves from all articles (books) 24 targeted by the system 10 and the wireless tag 18 of the human 16. It can be done. Therefore, although there are a plurality of antennas 124, they are comprehensively shown in FIGS.

  The LAN control board 210 controls the wireless communication device 216 connected to the control board 210 so that the server 20 can access the external network 14 wirelessly. Further, the input device control board 212 controls input by, for example, a keyboard or a mouse as an input device connected to the control board 212. The monitor control board 214 controls output to a monitor connected to the control board 214.

  The server 20 is connected to the article dictionary DB 122 and the voice recognition dictionary DB 126 (see FIG. 1) by an interface (not shown).

  As will be described later, article local dictionary data 504a, speech recognition local dictionary data 504b, utterance dictionary data 504c, and personal correct / incorrect information data 504d are set (stored) in the memory 204 (RAM).

  The article local dictionary corresponding to the article local dictionary data 504a is a dictionary in which contents extracted from the article dictionary DB 122 are registered, as will be described later. When the robot 12 recognizes the person 16, the server 20 extracts only the information of the article (book) 24 existing in the vicinity of the person 16 from the article dictionary DB 122 and registers it in the article local dictionary. The speech recognition local dictionary corresponding to the speech recognition local dictionary data 504b is a dictionary in which contents extracted from the speech recognition dictionary DB 126 are registered, as will be described later. When the robot 12 recognizes the human 16 and creates an article local dictionary, the server 20 extracts information necessary for voice recognition of words registered in the article local dictionary from the voice recognition dictionary DB 126 and performs voice recognition. Register in the local dictionary. Therefore, the article local dictionary and the voice recognition local dictionary are dynamically rewritten in accordance with a change in the position of the person 16. In this manner, a speech recognition local dictionary is created from the speech recognition dictionary stored in the speech recognition dictionary DB 126, and the number of words (phoneme symbol strings) that are subject to speech recognition is reduced by reducing the dictionary used for speech recognition. As a result, the time required for voice recognition processing can be shortened and the rate of correct voice recognition can be increased.

  The utterance dictionary corresponding to the utterance dictionary data 504c stores information necessary for the server 20 to determine the content of speech that the robot 12 causes the human 16 to utter. The personal correct / incorrect information corresponding to the personal correct / incorrect data 504d is information indicating whether or not the system 10 has finally succeeded in specifying the article (book) 24 instructed by the human 16 (successful speech recognition). Rate) for each person's 16 identification information (user ID).

  Next, referring to FIG. 5, the article dictionary stored in the article dictionary DB 122 assigns an ID such as a Ucode to one of the articles, and the name and attribute of each article. Necessary information is registered. The U-code is specifically a 128-bit number, and can individually identify articles that are 1 trillion times more than 1 trillion times 340 trillion. However, the ID used for the article dictionary DB 122 does not necessarily need to be such a U-code, and may consist of a combination of appropriate numbers and symbols.

  Such an article dictionary registers all articles, for example, household articles, to be identified by the system 10 (robot 12 and server 20), for example, with IDs and character strings, which is equivalent to a global dictionary. To do.

  In the article dictionary, information on one article (book) 24 is registered as one record. As described above, for example, “name” and “attribute” are stored in one record in addition to the book ID.

  Although illustration is omitted, “author” and “publisher” may be further stored as information about the book 24.

  “Name” is the title (title) of the corresponding book 24. The “attribute” is information related to the book 24, and in this embodiment, the content is supplementary explanation of the book 24. In this embodiment, information on the type of the book 24 (manga, novel, magazine, etc.), the color and thickness (thick, thin) of the cover is stored in the item “attribute”. In FIG. 5, for easy understanding, both the name and the attribute are described as text strings in text format, but actually, the attribute is described as a character string written in Roman letters.

  Next, the voice recognition dictionary DB 126 will be described. Generally, a speech recognition dictionary includes a word dictionary and a grammar dictionary, but the speech recognition dictionary DB 126 stores data about the word dictionary. A description of the grammar dictionary is omitted. Although illustration is omitted, the speech recognition dictionary DB 126 stores text-format words (or product identification information) corresponding to each of the text-format words (or product identification information) of the product names (names of the book 24 in this embodiment). Table data (speech recognition dictionary data) in which phoneme symbol formats (phoneme symbol strings) corresponding to words are described is stored.

  In the speech recognition process, the input speech is decomposed into phonemes, and a symbol representing the phoneme is generated for each decomposed phoneme. Thus, a phoneme symbol string corresponding to the input speech word is generated. Next, the phoneme symbol string corresponding to the input speech word is compared with the phoneme symbol string stored in the speech recognition dictionary DB 126 (actually, a speech recognition local dictionary described later). The phoneme symbol string closest to the phoneme symbol string corresponding to the input speech word is specified in the speech recognition dictionary DB 126 (speech recognition local dictionary), and is described corresponding to the specified phoneme symbol string. The word is output as a speech recognition result.

  As described above, in this system 10, when the person 16 indicates the article (book) 24 by voice, line of sight, and pointing, the robot 12 and the server 20 cooperate to provide the article (book) indicated by the person 16. 24 is specified, and the robot 12 transports the specified article (book) 24 to the person 16. Hereinafter, the exchange between the human 16 and the system 10 may be referred to as communication.

  More specifically, in the system 10, when the human 16 approaches the robot 12, the robot 12 recognizes the human 16 by the wireless tag 18. The server 20 includes an article dictionary DB 122 in which all articles (books) 24 targeted by the system 10 are registered, and a voice recognition dictionary DB 126 in which words for identifying the articles (books) 24 by voice recognition are registered. It is connected. When the robot 12 recognizes the person 16, the robot 12 transmits the identification information (user ID) of the person 16 to the server 20, and the local dictionary of the article dictionary DB 122 and the voice recognition dictionary DB 126 (article local dictionary, voice) to the server 20. Instructs creation of a recognition local dictionary.

  Upon receiving an instruction to create a local dictionary, the server 20 specifies the position of the person 16 recognized by the robot 12, and the article (book) 24 within a predetermined range from the specified person 16, for example, within a radius of 5 m. Are extracted from the article dictionary DB 122 to create an article local dictionary. Next, only the information necessary for voice recognition of the article (book) 24 registered in the article local dictionary is extracted from the voice recognition dictionary DB 126 to create a voice recognition local dictionary.

  Thereafter, the robot 12 utters, for example, “Let's bring some books?” To the recognized human 16. In response to this utterance, the human 16 replies, for example, “Bring Manga A” while pointing at the article (book) 24 while looking at the article (book) 24 that he / she wants to bring.

  Then, the robot 12 recognizes the voice of the human 16 saying “Please bring the comic A”, estimates the line of sight of the human 16, and estimates the direction in which the pointing finger is pointing. The specified article (book) 24 is specified.

  When the article (book) 24 instructed by the human 16 is identified, the server 20 confirms the content of the voice uttered by the robot 12 to confirm the article (book) 24 identified to the human 16, for example, “red cartoon”. Is determined, and the robot 12 speaks while pointing to the article (book) 24 (name “manga A”). That is, the robot 12 performs an action (confirmation action) for confirming the specified article (book) 24.

  At this time, the server 20 uses features (attributes) different from those of other articles existing in the vicinity of the person 16 recognized by the robot 12 or in the vicinity of the specified article (book) 24 for the content of the utterance, and also for voice recognition. The utterance content for confirming the article (book) 24 is generated with easy-to-use words.

  Here, the words that are easy to recognize are selected based on the knowledge that humans tend to imitate the utterances of robots. This is because it is easy to specify 24. However, if the utterance content is too short, there is a high possibility that the speech recognition will fail. On the other hand, if the utterance content is too long, there is a high possibility that it will not be imitated by humans. The utterance content is generated using words having three attributes.

  However, this is merely an example, and the content of the utterance may be determined using only one attribute word if the attribute word is long, or more if the attribute word is short. By registering the attribute in the article dictionary, it may be determined using four or more.

  Further, when the robot 12 speaks, for example, “is it a red cartoon?” In order to confirm the article (book) 24 identified by the system 10, the human 16 speaks “Yes” or “No”. Then, it responds to the robot 12. The server 20 recognizes the voice in the response of the human 16 and determines whether or not the article (book) 24 specified by the system 10 is the one designated by the human 16. If the article (book) 24 identified by the system 10 is not designated by the person 16, it is confirmed whether or not the next candidate article (book) 24 is designated by the person 16. To do. On the other hand, when the article (book) 24 specified by the system 10 is the one designated by the person 16, the robot 12 carries the article (book) 24 to the person 16.

  Further, the server 20 cumulatively records the result (correct / incorrect information) of confirming the specified article (book) 24. The server 20 refers to the cumulatively recorded correct / incorrect information as a success rate (speech recognition success rate) when generating the utterance content.

  The content disclosed in Japanese Patent Application Laid-Open No. 2009-223171 previously filed by the applicant and already published, except that the method of selecting words (words used for speech) is different when generating the speech content. In this embodiment, only the method for determining the word is described in detail because they are substantially the same and are not essential contents of the present invention.

  First, all the attributes Fc of the identified article (book) 24 (hereinafter referred to as “article X”) are acquired. For example, when the article (book) 24 with the name “manga A” is specified, {manga, red, light} is acquired as the attribute Fc. Next, a set for the attribute (attribute set) is generated. In this embodiment, a power set of attributes Fc is calculated and generated as an attribute set. The same applies hereinafter. However, the set Power (Fc) of the attribute Fc is {{manga}, {red}, {light}, {cartoon, red}, {red, light}, {cartoon, light}, {cartoon, red, light }}.

  Next, other articles existing in the vicinity of the identified article X (for example, within 30 cm) are detected, and attributes of the detected other articles are acquired. When there are a plurality of other articles, an attribute set is generated (a power set is calculated) for each other article.

  For example, a book 24 with the name “Manga B” (hereinafter referred to as “article 1”) and a book 24 with the name “magazine B” (hereinafter referred to as “article 2”) exist in the vicinity of the article X. , Attributes are acquired for each of these other articles 1 and 2, a set for each is calculated, and each attribute set is generated. Here, the attribute of the article 1 is {cartoon, blue, thin}, and the attribute of the article 2 is {magazine, red, thick}. Therefore, the power set Power (F1) of the article 1 is {{manga}, {blue}, {light}, {cartoon, blue}, {blue, light}, {cartoon, light}, {cartoon, blue, light. }}. The power set Power (F2) of the articles 2 is {{magazine}, {red}, {thick}, {magazine, red}, {red, thick}, {magazine, thick}, {magazine, red, thick. }}.

  Next, an element (common element) that is common to the attribute group of the identified article 24 and the attribute group of other neighboring articles 24 is extracted, and the common element is deleted from the attribute group of the identified article 24. Specifically, the common element E1 is extracted from the attribute group of the article X and the attribute group of the article 1, and the common element E2 is extracted from the attribute group of the article X and the attribute group of the article 2.

  In this embodiment, the common element E1 is {{manga}, {thin}, {cartoon, thin}}. The common element E2 is {red}. Therefore, the common element set S is the sum of the common element E1 and the common element E2, and specifically, {{manga}, {red}, {light}, {manga, light}}. Accordingly, when the elements of the set S are deleted (Power (Fc) -S) from the elements of the power set Power (Fc) of the identified article X, {{manga, red}, {red, light}, {cartoon, Red, light}}.

  Next, among the elements of Power (Fc) -S, the element with the smallest number of words is extracted. This operator is defined as, for example, min (). Therefore, min (Power (Fc) -S) becomes {{manga, red}, {red, light}}.

  In this way, the element that minimizes the number of words is selected, as described above, so that the content of the utterance has a length that facilitates speech recognition and that can be imitated by a human being. Because.

  Then, one element of min () elements is selected as a word used for the utterance content. In this embodiment, the element having the lowest similarity to the attribute of the other article 24 existing in the vicinity of the identified article 24 (the element having the highest dissimilarity) is selected. However, the similarity here means the degree to which the words to be compared are similar when they are pronounced by speech.

  Specifically, the sum of the Levenshtein distances for the word (character string) indicating the attribute included in the element is calculated, and the word indicating the attribute included in the element having the maximum sum is selected as the word used for the utterance content. Is done. The Levenshtein distance (edit distance) is a numerical value indicating how different two character strings are. Specifically, it is given as the minimum number of procedures required to transform one character string into another character string by inserting, deleting, or replacing characters.

  However, in this embodiment, in order to correctly recognize the voice, when calculating the Levenshtein distance, a character string shown in Roman letters is used. The character string indicated in Roman letters is registered in the article dictionary (article local dictionary). However, the Levenshtein distance is calculated and totaled for each of the types, colors, and thicknesses included in the attributes (with the same attributes).

  As shown in FIG. 8A, when the Levenshtein distance is calculated for {manga, red} selected as described above, {type, color} among the attributes is compared. Therefore, in the identified article X and article 1, the types are all “manga” and the colors are “aka” and “ao”, and two characters must be replaced, so the Levenshtein distance LD is “ 2 ”. In addition, for the specified item X and item 2, the type must be replaced with four characters for “manga” and “zashishi” and one character must be added, and the color for all items must match with “aka”. The Stein distance LD is “5”. Therefore, the total Levenshtein distance LD when using {manga, red} of {kind, color} as the utterance content is “7”.

  On the other hand, as shown in FIG. 8B, when calculating the Levenshtein distance for {red, thin} selected as described above, {color, thickness} among the attributes are compared. Therefore, in the specified article X and article 1, the characters “aka” and “ao” need to be replaced by two characters, and the thicknesses are all “usui”, so the Levenshtein distance LD is “ 2 ”. In addition, the specified item X and the item 2 all have the same color “aka”, and the thicknesses “usui” and “atsui” need to add one character and replace one character. The Stein distance is “2”. Therefore, the total of the Levenshtein distance LD when {color, thickness} {red, thin} is used as the utterance content is “4”.

  As described above, {manga, red} having a large total Levenshtein distance LD is selected as the utterance content.

  Next, a method for selecting a word to be used for the utterance contents when no other article (book) 24 exists in the vicinity of the identified article (book) 24 will be described.

  If no other article (book) 24 exists in the vicinity of the identified article (book) 24, all attributes of the identified article (book) 24 are acquired. When there is no other article (book) 24 in the vicinity of the identified article (book) 24, basically, any of the character strings of type, color, and thickness among the attributes is used. This is because the article (book) 24 can be confirmed.

  However, when another article (book) 24 exists in the vicinity of the person 16, a word having an attribute that maximizes the sum of the Levenshtein distance with the character string of the attribute of the other article (book) 24 is uttered. Select (determine) the word to use for the content. Here, the Levenshtein distance LD is calculated for each of the attribute type, color, and thickness with respect to another article, and the sum of the Levenshtein distance LD for the type, color, and thickness is calculated. Then, the word with the largest sum (words of type, color, and thickness) is selected as the word used for the utterance content.

  Note that the method for obtaining the Levenshtein distance LD is as described with reference to FIGS. 8A and 8B, and redundant description is omitted.

  FIG. 9 is an illustrative view showing one example of a memory map 500 of the memory 204 (RAM) shown in FIG. As shown in FIG. 9, the RAM includes a program storage area 502 and a data storage area 504.

  An information processing program for executing overall control of the server 20 is stored in the program area 502, and this information processing program includes an operation control program 502a, a speech recognition program 502b, an utterance content generation program 502c, and the like. These programs are read from the HDD all at once or partially as necessary, and stored in the program storage area 502 of the RAM. However, the program may be stored in a ROM (not shown) and read from there.

  The motion control program 502a is a program for calculating control information regarding body motion such as pointing motion of the robot 12 and instructing the robot 12. The voice recognition program 502b is a program for recognizing a voice corresponding to a voice signal transmitted from the robot 12. The utterance content generation program 502c is a program for generating the utterance content in the confirmation action.

  Although illustration is omitted, the program storage area 502 also stores other programs such as a program for detecting a human gaze direction.

  The data storage area 504 stores article local dictionary data 504a, speech recognition local dictionary data 504b, utterance dictionary data 504c, personal correct / incorrect information data 504d, and the like. Further, the data storage area 504 is provided with a dictionary registration flag 504e.

  The article local dictionary data 504a is an article (within a predetermined range centered on the person 16 specified by the user ID transmitted from the robot 12 from the original article dictionary (global dictionary) stored in the article dictionary DB 122 ( Book) Data of some article dictionaries obtained by extracting 24 records. The voice recognition local dictionary data 504b recognizes articles and the like registered in a part of the article dictionary corresponding to the article local dictionary data 504a from the original voice recognition dictionary (global dictionary) stored in the voice recognition dictionary DB 126. This is the data of a part of voice recognition dictionary extracted.

  The utterance dictionary data 504c is data about information necessary for the server 20 to make the robot 12 speak the voice content to the human 16, that is, the utterance content.

  In this embodiment, since the type, color, and thickness are included as the attributes of the article, the utterance content is determined by the following fixed text. For example, when a word about color and type is used, the utterance content is “?? △△△?”. However, a word about the color (red, blue, yellow, brown, white, black, etc.) described in the attribute of the specified article X is entered at “XX”, and “△△” is entered at “△△”. A word of a type (manga, novel, magazine, etc.) described in the attribute of the specified article X is entered. same as below.

  When words about thickness and type are used, the utterance content is “XXΔΔ?”. However, a word having a thickness (thick or thin) described in the attribute of the specified article X is entered at “XX”. same as below.

  Furthermore, when words about color and thickness are used, the utterance content is “XX book of XX?”. However, in this embodiment, since the book 24 is used as an article, the “book” is spoken. However, when other articles are used, the common noun is used.

  Personal correct / incorrect information data 504d is data on a table in which the success rate of voice recognition is described in correspondence with the user ID. For example, the correctness / incorrectness of voice recognition is represented by a ratio (percentage) of the number of times of successful voice recognition (the number of times the specified article (book) 24 is correct) in the number of communication (cumulative).

  The dictionary registration flag 504e is determined to be used in the utterance in the confirmation action, and is a word or a plurality of attributes composed of a word for one attribute for indicating the item X and a word for a common noun of the item X This is a flag for determining whether or not words (hereinafter referred to as “specific words”) composed of the words about “” are registered in the speech recognition dictionary. The dictionary registration flag 504e is composed of a 1-bit register, and if the flag is on, the data value “1” is set in the register, and if the flag is off, the data value “0” is stored in the register. Is set. However, the dictionary registration flag 504e is turned on when a specific word is registered in the speech recognition dictionary, and the dictionary registration flag 504e is turned off when it is not registered in the speech recognition dictionary.

  For example, if the word of the type of article X (here “manga”) and color (here “red”) is used in the utterance of the confirmation action, “red cartoon” is registered as the specific word Is done. Similarly, when a word of the type and thickness (here, “thick”) of the item X is used, “thick cartoon” is registered as the specific word. Further, when the word of the color and thickness of the article X is used in the utterance of the confirmation action, the word of the common noun of the article X is further used, and “red thick book” is registered as the specific word. Although the description is omitted, the same applies to the case where words of other types, colors, and thicknesses are used.

  Although not shown, the data storage area 504 stores other data necessary for executing the information processing program, and is provided with a counter (timer), other flags, and the like as necessary.

  FIG. 10 is a flowchart of the confirmation action determination process of the CPU 200 shown in FIG. As illustrated in FIG. 10, when starting the confirmation action determination process, the CPU 200 generates a pointing operation from the position of the robot 12 and the position of the identified article X in step S1.

  In the next step S3, it is determined whether or not the user is the first meeting. Here, whether or not the CPU 200 is the first to communicate with the human 16 recognized by the robot 12, that is, whether or not it is the first time to specify the article (book) 24 instructed by the human 16. Judging. Specifically, the CPU 200 determines whether or not the user ID transmitted from the robot 12 is registered in the personal correctness / error information data 504d.

  If “YES” in the step S3, that is, if the user is a first-time user, in a step S5, the word included in the utterance content of the confirmation action (used for the utterance) is determined as the name of the specified article X, and the step Proceed to S13. On the other hand, if “NO” in the step S3, that is, if the user is not the first meeting user, in a step S7, it is determined whether or not the success rate of the speech recognition for the user is 70% or more. However, the CPU 200 refers to the personal correctness / error information data 504d and acquires the success rate described in correspondence with the user ID transmitted from the robot 12.

  In step S7, 70% is set as a threshold for determining whether the success rate of speech recognition is high or low. However, the threshold is not limited to this, and the threshold is not limited to the environment in which the system 10 is used or used. It can be freely changed depending on the form.

  If “NO” in the step S7, that is, if the success rate of the speech recognition is less than 70%, the process proceeds to a step S5. On the other hand, if “YES” in the step S7, that is, if the success rate of the speech recognition is 70% or more, a word selection process (see FIG. 11) described later is executed in a step S9. In step S11, the dictionary registration flag 504e is turned on, and the process proceeds to step S13. Although illustration and description are omitted, the CPU 200 turns off the dictionary registration flag 504e when the confirmation action determination process is started.

  In step S13, the utterance content is generated. Here, the CPU 200 refers to the utterance dictionary data 504c and generates the utterance content in the confirmation action. At this time, a word indicating the name of the article X selected in step S5 or a word regarding the attribute selected in step S9 is used.

  In the next step S15, it is determined whether the dictionary registration flag 504e is on. If “NO” in the step S15, that is, if the dictionary registration flag 504e is turned off, the process directly proceeds to a step S19. On the other hand, if “YES” in the step S15, that is, if the dictionary registration flag 504e is turned on, the specific word included in the generated utterance content is voiced together with the user ID corresponding to the article X in a step S17. The data is stored (registered or added) in the speech recognition dictionary data in the recognition DB 126, and the process proceeds to step S19.

  Therefore, if the specific word is used when the user specified by the user ID indicates the article X thereafter, the corresponding article X can be specified by voice recognition of the specific word. . Therefore, smooth communication can be achieved. Such a result is brought about because a person tends to imitate the contents uttered by the robot 12 as described above.

  In step S19, the pointing action generated in step S1 and the utterance content determined in step S13 are transmitted to the robot 12, and the confirmation action determination process is terminated. In response to this, the robot 12 performs a pointing operation instructed from the server 20 and utters (speech outputs) the utterance content instructed from the server 20. That is, the robot 12 performs a confirmation action for the identified article X. Although illustration is omitted, the server 20 recognizes the content of the response by voice according to the user's response (“Yes” or “No”) to the robot 12 thereafter, so that the personal correct / incorrect information data 504d. Update.

  Note that such confirmation determination processing is executed when there is a request from the robot 12, and when there is no request from the robot 12, it is in a standby state.

  FIG. 11 is a flowchart of the word selection process shown in step S9 of FIG. As shown in FIG. 11, when the word selection process is started, the CPU 200 determines whether or not there is another article in the vicinity of the specified article X in step S31. Here, the CPU 200 determines whether or not there is an article 24 that has received radio waves by the same antenna 124 as the antenna 124 that has received radio waves transmitted from the wireless tag 18 attached to the article X.

  If “YES” in the step S31, that is, if there is another article 24 in the vicinity of the article X, a process (see FIG. 12) for generating an attribute set of the article X described later is executed in a step S33. In step S35, a process (see FIG. 13) for generating attribute sets of other articles in the vicinity of the article X, which will be described later, is executed. Subsequently, in step S37, processing for reducing the attribute group of the article X described later (see FIG. 14) is executed, and in step S39, the attribute group having the smallest number of words is extracted from the attribute group of the article X. In step S41, a word selection (1) process (see FIG. 15) described later is executed, and the process returns to the confirmation action determination process.

  If “NO” in the step S31, that is, if there is no other article 24 in the vicinity of the article X, a process for determining a candidate for the name of the article X described later (see FIG. 16) is performed in a step S43. Run. In the next step S <b> 45, it is determined whether there is another article 24 in the vicinity of the user 16. Here, the CPU 200 determines whether or not there is an article 24 that has received radio waves from the same antenna 124 that received the radio waves transmitted from the wireless tag 18 attached to the user 16 indicated by the identified user ID. to decide.

  If “NO” in the step S45, that is, if there is no other article 24 in the vicinity of the user, in a step S53, one word is selected from the candidate names according to a predetermined rule, and the process returns to the confirmation action determining process. . For example, in step S53, the CPU 200 randomly selects one word from candidate names, or selects a word for one attribute determined in advance by the system 10 or the administrator of the server 20.

  On the other hand, if “YES” in the step S45, that is, if there is another article 24 in the vicinity of the user, the attribute of the other article 24 in the vicinity is searched with reference to the article local dictionary data 504a in a step S47. To do. Here, the CPU 200 searches for the article ID of another article 24. In the subsequent step S49, the attribute of the other article 24 is acquired from the record described corresponding to the retrieved article ID. The same applies to the case where the attribute of the article 24 is searched or the attribute is acquired. In step S51, a word selection (2) process (see FIG. 17) described later is executed, and the process returns to the confirmation action determination process.

  As shown in FIG. 12, when the CPU 200 starts the process for generating the attribute set of the article X shown in step S33 of FIG. 11, the CPU 200 searches for the attribute of the article X in step S61. In a succeeding step S63, the attribute of the article X is acquired.

  In step S65, a power set of attributes is calculated. Here, as described above, the power set Power (Fc) for the attribute Fc of the article X is calculated. In step S67, a power set Power (Fc) is generated as an attribute set of the article X, and the process returns to the word selection process.

  As illustrated in FIG. 13, when the CPU 200 starts the process of generating the attribute set of another article in the vicinity shown in step S <b> 35 in FIG. 11, in step S <b> 71, another article K (K identifies another article). The same is true for the following), and the attributes of other articles K are acquired in step S73. In the next step S75, a power set Power (FK) of attributes is calculated. In step S77, a power set Power (FK) is generated as an attribute set of another article K, and the process returns to the word selection process.

  It should be noted that the attribute group generation process for other articles in the vicinity shown in FIG. 13 is executed for each other article.

  As shown in FIG. 14, when the CPU 200 starts the process of reducing the attribute set of the article X shown in step S37 of FIG. 11, in step S81, the set Power (Fc) that should be the attribute set of the article X and the other The common element of the power set Power (K) of the attribute set of the article K is extracted. However, when there are a plurality of other articles K, common elements are extracted for each of the other articles K. In the next step S83, the common element is deleted from the attribute set of the article X, and the process returns to the word selection process.

  As shown in FIG. 15, when the CPU 200 starts the word selection (1) process shown in step S <b> 41 shown in FIG. 11, the character string of the attribute group of the item X and the character string of the attribute group of the other item K are displayed. The sum of the Levenshtein distance LD is calculated. In the next step S91, the attribute group that maximizes the sum of the Levenshtein distances LD is selected as a word to be used in the confirmation action, and the process returns to the word selection process.

  As a result of the process of reducing the attribute group of the item X, if only one attribute group of the item X remains, the attribute group is confirmed without performing the word selection (1) process. Selected as a word to use in action.

  As illustrated in FIG. 16, when the CPU 200 starts the process for generating the candidate name of the article X illustrated in step S <b> 43 illustrated in FIG. 11, the CPU 200 searches for the attribute of the article X in step S <b> 101. In the next step S103, the attribute of the article X (in this embodiment, a word about type, color, and thickness) is acquired. In step S105, each attribute is determined as a word candidate to be used in the confirmation action, and the process returns to the word selection process.

  As shown in FIG. 17, when the CPU 200 starts the word selection (2) process shown in step S51 shown in FIG. 11, in step S111, the character string of the candidate attribute and each other article in the vicinity of the user's position. The sum of the Levenshtein distance LD of the character string of the attribute is calculated. However, the Levenshtein distance LD for each type, color, and thickness is obtained for each of the other articles X, and the sum is calculated. In step S113, the attribute (type, color or thickness) that maximizes the sum of the Levenshtein distances LD is selected as a word to be used in the confirmation action, and the process returns to the word selection process.

  According to this embodiment, an attribute set different from other articles existing in the vicinity of the specified article is extracted, and the attribute set having the maximum Levenshtein distance is used in the confirmation action. Since it selects as a word, it is not necessary to detect a speech recognition rate beforehand, and the word used by confirmation action can be determined easily. Therefore, the utterance content in the confirmation action can be easily generated.

  Further, according to this embodiment, since the attribute group having the smallest number of words is extracted from the extracted attribute groups, words that are easy to imitate for humans can be included in the utterance content.

  Furthermore, according to this embodiment, an attribute set different from other articles is included in the utterance content, and words (specific words) determined by the attribute set are additionally added to the speech recognition dictionary corresponding to the user. Since the registration is performed thereafter, even if the specific word is used, the article indicated by the specific word can be specified by voice recognition. Therefore, communication with the user can be performed smoothly.

  In this embodiment, the server executes the voice recognition process and the confirmation action determination process. However, these processes may be executed by a robot. In such a case, speech recognition dictionary data (speech recognition local dictionary data), article dictionary data (article local dictionary data), utterance dictionary data, and personal errata information data are stored in a robot internal memory or an external memory accessible by the robot. There is a need to.

DESCRIPTION OF SYMBOLS 10 ... Communication system 12 ... Communication robot 14 ... Network 18 ... Wireless tag 20 ... Server 24 ... Goods (book)
80 ... CPU
120 ... Camera 124 ... Antenna 200 ... CPU
208... RFID tag reader

Claims (7)

A communication system for identifying an article designated by a person by voice recognition and confirming by voice whether the identified article is designated by the person,
Storage means for storing a name of an article and a plurality of words related to the article;
First creation means for reading a plurality of words related to the identified article from the storage means and creating a first power set for the read words;
A second creation unit that reads a plurality of words related to other articles existing in the vicinity of the identified article from the storage unit and creates a second power set for the read words for each of the other articles;
Deleting means for deleting elements common between the first power set created by the first creating means and each of the second power sets created by the second creating means from the first power set;
As a result of deletion by the deletion unit, a first selection unit that selects an element having the smallest number of words among elements included in the first power set;
Of the elements selected by the first selection means, second selection means for selecting an element having the highest degree of dissimilarity with a character string of a plurality of words related to the other article, and the second selection means A communication system comprising utterance content generation means for generating utterance content when confirming by voice including a word included in a selected element. A calculating means for calculating a Levenshtein distance between the character string of the first word included in the element selected by the first selecting means and each of the character strings of the plurality of second words related to the other article; Prepared,
2. The communication system according to claim 1, wherein the second selection unit selects an element including a first word having a maximum Levenshtein distance calculated by the calculation unit. First determination means for determining whether or not the other article exists in the vicinity of the identified article;
Candidate determination means for determining each of a plurality of words related to the specified article as a candidate to be included in the utterance content when the first determination means determines that the other article does not exist in the vicinity of the specified article And third selection means for selecting a word having the highest dissimilarity with a character string of a plurality of words related to other articles existing in the vicinity of the human among the words determined by the candidate determination means. In addition,
The communication system according to claim 1, wherein the utterance content generation unit generates utterance content when confirming by voice including a word included in the element selected by the third selection unit. Second determination means for determining whether or not there is another article in the vicinity of the human when the first determination means determines that the other article does not exist in the vicinity of the specified article; and And a fourth selecting means for selecting one word from the candidates determined by the candidate determining means according to a predetermined rule when it is determined by the determining means that no other article exists in the vicinity of the person. Prepared,
4. The communication system according to claim 3, wherein the utterance content generation unit generates utterance content when confirming by voice including a word included in the element selected by the fourth selection unit. 5. It is used for a communication system that identifies an article designated by a person by voice recognition and confirms by voice whether the identified article is designated by the person, and generates utterance contents when confirming by voice. An utterance content generation device,
Storage means for storing a name of an article and a plurality of words related to the article;
First creation means for reading a plurality of words related to the identified article from the storage means and creating a first power set for the read words;
A second creation unit that reads a plurality of words related to other articles existing in the vicinity of the identified article from the storage unit and creates a second power set for the read words for each of the other articles;
Deleting means for deleting elements common between the first power set created by the first creating means and each of the second power sets created by the second creating means from the first power set;
As a result of deletion by the deletion unit, a first selection unit that selects an element having the smallest number of words among elements included in the first power set;
Of the elements selected by the first selection means, second selection means for selecting an element having the highest degree of dissimilarity with a character string of a plurality of words related to the other article, and the second selection means An utterance content generation device comprising utterance content generation means for generating an utterance content for confirmation by voice including a word included in a selected element. Used for a communication system that identifies an article designated by a person by voice recognition and confirms by voice whether the identified article is designated by the person. The name of the article and a plurality of words related to the article An utterance content generation program of an utterance content generation device for generating an utterance content when confirming by voice,
In the processor of the utterance content generation device,
A first creation step of reading a plurality of words related to the identified article from the storage means and creating a first power set for the read words;
A second creation step of reading a plurality of words related to other articles existing in the vicinity of the identified article from the storage means and creating a second power set for the read words for each of the other articles;
Deleting a common element between the first power set created in the first creation step and each of the second power sets created in the second creation step, from the first power set;
A first selection step of selecting an element having the smallest number of words among elements included in the first power set as a result of deletion in the deletion step;
A second selection step of selecting an element having the highest degree of dissimilarity to a plurality of word strings related to the other article among the elements selected in the first selection step; and An utterance content generation program for executing an utterance content generation step for generating an utterance content when confirming by voice including a word included in a selected element. Used for a communication system that identifies an article designated by a person by voice recognition and confirms by voice whether the identified article is designated by the person. The name of the article and a plurality of words related to the article Utterance content generation method of the utterance content generation device for generating the utterance content when confirming with voice,
The processor of the utterance content generation device,
(A) reading a plurality of words related to the specified article from the storage means, creating a first power set for the read words;
(B) A plurality of words related to other articles existing in the vicinity of the identified article are read from the storage unit, and a second power set for the read words is created for each of the other articles,
(C) Delete elements common between the first power set created in step (a) and the second power set created in step (b) from the first power set. ,
(D) selecting an element having the smallest number of words among elements included in the first power set as a result of deletion in the step (c);
(E) Of the elements selected in the step (d), select an element having the highest degree of dissimilarity with a character string of a plurality of words related to the other article; and (f) the step (e The utterance content generation method of generating the utterance content when confirming by voice including the word included in the element selected in (1).

Images