JP6248717B2 - Elevator control device - Google Patents

Description

  The present invention relates to an elevator control device.

  Patent Document 1 listed below describes an elevator control device. The control device includes a voice recognition device and voice recognition state determination means. The voice recognition device performs call registration based on the user's voice. The voice recognition state determination unit determines whether or not the user is speaking. When it is determined by the voice recognition state determining means that the user is speaking, the control device extends the door opening time of the car by a certain time.

JP 2013-52957 A

  In the elevator control device described in Patent Document 1, when it is determined that the user is speaking, the door opening time of the car is extended. For this reason, an extra waiting time occurs when an utterance unrelated to call registration is made.

  The present invention has been made to solve the above problems. An object of the present invention is to provide an elevator control device that can reduce an extra waiting time of an elevator user.

  The elevator control device according to the present invention includes a storage unit that stores a floor name of a building in which the elevator is installed, a voice that is input to a voice input unit that receives voice input, and a floor that is stored in the storage unit. Valid as a word that constitutes a floor name by which the speech recognition means that recognizes the speech in division units by comparing the floor name with division units smaller than words and the recognition result in the division unit by the speech recognition means is effective The calculation means for determining whether the recognition is invalid or not, and if the recognition result is determined to be valid by the calculation means, the door opening time of the car is extended, and the calculation result is determined to be invalid by the calculation means. And a door control means that does not extend the door opening time of the car.

  In the elevator control device according to the present invention, the door control means does not extend the door opening time of the car when the calculation means determines that the recognition result in divided units is invalid. For this reason, the extra waiting time of the elevator user can be shortened.

It is a block diagram of the elevator control apparatus in Embodiment 1 of this invention. It is a block diagram of the memory | storage means 5 in Embodiment 1 of this invention. It is a flowchart of the speech recognition of 1 beat unit in Embodiment 1 of this invention. It is a flowchart of the speech recognition of the word unit in Embodiment 1 of this invention. It is a block diagram of the calculating means 8 in Embodiment 1 of this invention. It is explanatory drawing which shows the 1st example of the state transition 28 in Embodiment 1 of this invention. It is a table | surface which shows an example of the list | wrist L1 of the transition required for reception of the floor name in Embodiment 1 of this invention. It is a table | surface which shows an example of list | wrist L2 of the effective floor name in Embodiment 1 of this invention. It is explanatory drawing which shows the 2nd example of the state transition 28 in Embodiment 1 of this invention. It is a flowchart of the update operation | movement of the state transition 28 in Embodiment 1 of this invention. It is a flowchart of the initialization process of the calculating means 8 in Embodiment 1 of this invention. It is a list of the time which the door control means 11 in Embodiment 1 of this invention hold | maintains. It is a flowchart of the update operation | movement of the elapsed time t4 in Embodiment 1 of this invention. It is a flowchart of the whole operation | movement of the elevator control apparatus in Embodiment 1 of this invention. It is explanatory drawing which shows the 3rd example of the state transition 28 in Embodiment 1 of this invention. It is a block diagram of the elevator system provided with the elevator control apparatus in Embodiment 2 of this invention. It is a flowchart of the whole operation | movement of the elevator control apparatus in Embodiment 2 of this invention.

  The present invention will be described in detail with reference to the accompanying drawings. In each figure, the same or corresponding parts are denoted by the same reference numerals. The overlapping description will be simplified or omitted as appropriate.

Embodiment 1 FIG.
FIG. 1 is a configuration diagram of an elevator system including an elevator control device according to Embodiment 1 of the present invention. Hereinafter, the configuration of the elevator control device will be described with reference to FIG.

  The elevator system includes a car 1 and an elevator control device 2. The car 1 is provided inside a hoistway (not shown). The hoistway is provided in a building having a plurality of floors. Inside the car 1, voice input means 3 and door opening / closing means 4 are provided. The voice input means 3 and the door opening / closing means 4 are connected to the elevator control device 2. The voice input means 3 is, for example, a microphone.

  The elevator control device 2 includes storage means 5, voice recognition means 6, management information acquisition means 7, calculation means 8, switching means 9, time acquisition means 10, door control means 11, and destination floor registration means 12. The elevator control device 2 is provided as a control panel, for example. The storage means 5 is, for example, a memory provided in the control panel.

  The storage unit 5 includes a one-beat unit storage unit 13 and a word unit storage unit 14. The voice recognition unit 6 includes a one-beat unit voice recognition unit 15, a word unit voice recognition unit 16, and an input sound buffer 17. The input sound buffer 17 is, for example, a memory provided in the control panel. The computing means 8 includes a 1-beat unit validity judgment means 18 and a word-unit validity judgment means 19.

  The car 1 is lifted and lowered inside the hoistway. Sound is input to the voice input means 3. The input sound to the voice input means 3 is, for example, the voice of an elevator user. The door opening / closing means 4 opens and closes the door of the car 1.

  The memory | storage means 5 has memorize | stored the information which shows the floor name etc. of the building in which the hoistway was provided. The voice recognition unit 6 recognizes the voice input to the voice input unit 3 based on the information stored in the storage unit 5. The voice recognition unit 6 stores the voice input to the voice input unit 3 in the input sound buffer 17 as an input sound. The management information acquisition means 7 acquires elevator management information. The speech recognition result is input from the speech recognition means 6 to the calculation means 8. Based on the management information acquired by the management information acquisition unit 7, the calculation unit 8 determines whether or not the voice recognized by the voice recognition unit 6 indicates an effective floor name of the building. A valid floor name is a floor on which a call can be registered as a destination floor. The switching unit 9 switches between the storage unit 5, the voice recognition unit 6, and the calculation unit 8. The time acquisition unit 10 acquires the current time. The door control means 11 holds information indicating a plurality of times. Further, the door control means 11 controls the door opening / closing means 4 based on the recognition result by the voice recognition means 6, the determination result by the calculation means 8 and the time acquired by the time acquisition means 10. The destination floor registration unit 12 registers a call to the destination floor based on the determination result in the calculation unit 8.

  The management information is information used for determining a valid floor name. In the first embodiment, the management information includes the floor name of the service floor of the elevator and the floor name of the floor existing in the traveling direction of the car 1. The service floor is a floor on which elevator users can get on and off the car 1.

  The storage unit 5, the speech recognition unit 6 and the calculation unit 8 process the building floor name in units of words and units of division smaller than words. In Embodiment 1, the division unit smaller than the word is one beat. For example, a floor name indicating the first floor is processed as one word “Ikkai” in word units. Further, the floor name indicating the first floor is processed by being divided into four words “i”, “t”, “k” and “i” in units of one beat.

  The 1-beat unit storage means 13 stores the floor name divided into 1-beat units. The one-beat unit voice recognition unit 15 performs voice recognition in units of one beat on the voice input to the voice input unit 3. The one-beat-unit validity determining unit 18 determines whether or not the one-beat unit word recognized by the one-beat unit speech recognition unit 15 is valid as a word constituting an effective floor name.

  The word unit storage unit 14 stores floor names in units of words. The word unit speech recognition means 16 performs speech recognition on a word basis using the input sound stored in the input sound buffer 17. The word unit validity determination unit 19 determines whether or not the word recognized by the word unit speech recognition unit 16 is a valid floor name.

  FIG. 2 is a configuration diagram of the storage unit 5 according to the first embodiment of the present invention. FIG. 2 shows a case where the floors existing in the building are the first to fifth floors as an example. Hereinafter, the storage means 5 will be described in detail with reference to FIG.

  The word unit storage unit 14 stores a word unit dictionary 20 and a threshold value 21 for speech recognition in units of words. The word unit dictionary 20 includes an acoustic model 22 and a label 23. The acoustic model 22 is information corresponding to the pronunciation of each floor name in units of words. The label 23 is information indicating which floor name word each acoustic model 22 corresponds to. Each acoustic model 22 is associated with each label 23 on a one-to-one basis. In FIG. 2, the acoustic model 22 and the label 23 are stored for each word indicating each floor name, such as “Ikkai”, “Nikai”, “Sankai”, “Yonkai”, and “Gokai”. The threshold value 21 for speech recognition in units of words is a preset value. The threshold value 21 for speech recognition in units of words is used when the speech recognition means 6 performs speech recognition in units of words.

  The 1-beat unit storage means 13 stores a 1-beat unit dictionary 24 and a threshold value 25 for speech recognition in 1-beat units. The one-beat unit dictionary 24 includes an acoustic model 26 and a label 27. The acoustic model 26 is information corresponding to the pronunciation of a word in units of one beat obtained by dividing each floor name into units of one beat. The label 27 is information indicating which one beat unit word each acoustic model 26 corresponds to. Each acoustic model 26 is associated with each label 27 on a one-to-one basis. In FIG. 2, the acoustic model 26 and the label 27 are labeled as “i”, “tsu”, “ka”, “ni”, “sa”, “n”, “yo”, and “go”. Is stored for each word of one beat constituting the word indicating. The threshold value 25 for voice recognition in units of one beat is a preset value. The threshold value 25 for voice recognition in units of one beat is used when the voice recognition means 6 performs voice recognition in units of one beat.

  One-beat unit speech recognition by the one-beat unit speech recognition means 15 will be described. The one-beat unit voice recognition unit 15 stores the voice input to the voice input unit 3 in the input sound buffer 17 for each beat. The 1-beat unit speech recognition means 15 compares the sound for 1 beat with the acoustic model 26 stored in the 1-beat unit dictionary 24 to calculate the similarity in 1-beat units. The degree of similarity in units of one beat is calculated as a numerical value. The 1-beat unit voice recognition means 15 compares the sound for one beat with all the acoustic models 26 to calculate the same number of similarities in units of one beat as the number of acoustic models 26. That is, the degree of similarity in units of one beat, which is the same as the number of acoustic models 26, is calculated for one beat of voice. The 1-beat unit speech recognition means 15 creates a set of calculated 1-beat similarity and 1-beat word based on the label 27 associated with the acoustic model 26. The 1-beat unit speech recognition means 15 selects all pairs in which the similarity in 1-beat units exceeds the threshold value 25 for speech recognition in 1-beat units among the created sets. At this time, if the selected set exists, the 1-beat unit speech recognition means 15 outputs all the words in 1-beat units included in the selected set as the recognition result. On the other hand, if the selected set does not exist, the 1-beat unit speech recognition means 15 outputs “0” as a signal indicating a speech recognition failure. The one-beat unit voice recognition unit 15 performs the above-described process every time one beat of voice is input to the voice input unit 3.

  FIG. 3 is a flowchart of voice recognition in units of one beat in Embodiment 1 of the present invention. In the following, speech recognition in units of one beat will be described by exemplifying the case where one beat of voice “yo” is input to the voice input means 3. Here, it is assumed that the voice recognition threshold 25 in units of one beat is set to 0.8.

  The one-beat unit voice recognition means 15 stores the voice “yo” for one beat in the input sound buffer 17 as an input sound (step S101). The 1-beat unit speech recognition means 15 compares the sound “Y” for 1 beat with the acoustic model 26 in the 1-beat unit dictionary 24 to calculate the similarity in 1-beat units. Then, the 1-beat unit speech recognition means 15 creates a set of similarity in 1-beat units and words in 1-beat units based on the label 27 (step S102). Here, the similarity between the sound “Yo” for one beat and each acoustic model 26 is 0.63 for “I”, 0.32 for “I”, 0.68 for “K”, and “D” for Assume that 0.55, “sa” is 0.73, “n” is 0.21, “yo” is 0.87, and “go” is 0.81. The 1-beat unit speech recognition means 15 selects a pair having a similarity higher than 0.8, which is the value of the threshold value 25 for speech recognition in 1-beat units (step S103). Since the degree of similarity higher than 0.8 is 0.87 paired with “yo” and 0.81 paired with “go”, these two pairs are selected in step S103. The one-beat unit speech recognition means 15 determines whether or not the group selected in step S103 exists (step S104). If it is determined in step S104 that the group selected in step S103 exists, the 1-beat unit speech recognition means 15 outputs a word in 1-beat unit included in the group selected in step S103 as a recognition result (step S103). S105). On the other hand, if it is determined in step S104 that the group selected in step S103 does not exist, the 1-beat unit speech recognition unit 15 outputs a signal indicating failure of speech recognition (step S106). Here, since the group selected in step S103 exists, the process proceeds from step S104 to step S105. As a result, the one-beat unit speech recognition means 15 outputs the words “yo” and “go” in units of one beat as recognition results.

  The word unit speech recognition by the word unit speech recognition means 16 will be described. The word unit voice recognition means 16 processes the voices of all the beats stored in the input sound buffer 17 as input sounds into one integrated input sound. The word unit speech recognition means 16 compares the integrated input sound with the acoustic model 22 stored in the word unit dictionary 20 to calculate the word unit similarity. The word unit similarity is calculated as a numerical value. The word unit speech recognition means 16 compares the integrated input sound with all the acoustic models 22 to calculate the same number of word units as the number of acoustic models 22. That is, for the integrated input sound, the same number of word units as the number of acoustic models 22 are calculated. The word unit speech recognition means 16 creates a calculated word unit similarity and word pair based on the label 23 associated with the acoustic model 22. The word unit speech recognition means 16 selects a pair having the highest similarity in word units among the created groups. At this time, when the similarity of the word unit included in the selected group exceeds the threshold value 21 of the speech recognition of the word unit, the word unit speech recognition means 16 uses the word included in the selected group as a recognition result. Output. On the other hand, when the similarity of the word unit included in the selected group does not exceed the threshold value 21 for the speech recognition of the word unit, the word unit speech recognition means 16 outputs a signal indicating the speech recognition failure.

  FIG. 4 is a flowchart of word-by-word speech recognition in Embodiment 1 of the present invention. Hereinafter, speech recognition in units of words will be described by exemplifying a case where “Yongkai” is stored as an integrated input sound in the input sound buffer 17. Here, it is assumed that the threshold value 21 for speech recognition in units of words is set to 0.8.

  The word unit speech recognition means 16 reads the integrated input sound “Yongkai” from the input sound buffer 17. The word unit speech recognition means 16 compares the integrated input sound “Yongkai” with the acoustic model 22 in the word unit dictionary 20 to calculate the similarity in word units. Then, the word unit speech recognition means 16 creates a word unit similarity and word pair based on the label 23 (step S201). Here, the similarity between the integrated input sound “Yongkai” and each acoustic model 22 is 0.62 for “Ikkai”, 0.38 for “Nikai”, 0.80 for “Sankai”, and 0 for “Yongkai”. 91, “Gokai” is 0.43. The word unit speech recognition means 16 selects a set of “Yongkai” having the highest similarity 0.91 among the similarities calculated in step S201 (step S202). The word unit speech recognition means 16 determines whether or not the similarity of the group selected in step S202 is higher than the threshold value 21 for speech recognition in word units (step S203). In step S203, when it is determined that the similarity of the group selected in step S202 exceeds the threshold value 21 for word-based speech recognition, the word-unit speech recognition means 16 uses the group selected in step S202. The included word is output as a recognition result (step S204). On the other hand, in step S203, when it is determined that the similarity of the group selected in step S202 does not exceed the word-unit speech recognition threshold value 21, the word-unit speech recognition means 16 indicates a speech recognition failure. A signal is output (step S205). Here, since the similarity 0.91 paired with “Yong Kai” exceeds 0.8, which is the value of the speech recognition threshold value 21 in units of words, the process proceeds from step S203 to step S204. As a result, the word unit speech recognition means 16 outputs the word “Yongkai” as a recognition result.

  FIG. 5 is a configuration diagram of the calculation means 8 in the first embodiment of the present invention. Hereinafter, the calculation means 8 will be described in detail with reference to FIG.

  The one-beat-unit validity determining means 18 holds a state transition 28, a list L1 of transitions necessary for accepting a floor name, and a completion determination pending flag FL1. The word unit validity determination means 19 holds a list L2 of valid floor names. The state transition 28 is information indicating the state transition of the floor name voice input. The state transition 28 is updated based on the recognition result of the voice recognition in units of one beat output from the voice recognition unit 15 in units of one beat. The transition list L1 necessary for accepting floor names is information indicating transitions necessary for accepting the floor names of each floor existing in the building. “Accepting a floor name” means that words in units of one beat constituting a word indicating the floor name are input in the correct order. The completion determination pending flag FL1 is a variable indicating whether or not the determination as to whether voice input of a valid floor name has been completed is pending. If the completion determination hold flag FL1 is 1, it is determined whether or not the floor name voice input has been completed. If the completion determination pending flag FL1 is 0, it is not pending whether the floor name voice input has been completed. Hereinafter, the transition list L1 and the valid floor name list L2 necessary for accepting the floor name are abbreviated as “list L1” and “list L2”. Further, the completion determination pending flag FL1 is abbreviated as “flag FL1”.

  FIG. 6 is an explanatory diagram showing a first example of the state transition 28 according to Embodiment 1 of the present invention. FIG. 6 shows a case where the floors existing in the building are the first to fifth floors as an example. Hereinafter, the state transition 28 will be described with reference to FIG.

  FIG. 6 shows a state 1 to a state 6 and a transition T201 to a transition T209. State 1 is an initial state. State 6 is an acceptance state. Each transition is associated with one word in units of one beat. The word of one beat unit associated with each transition is the same as that contained in the one beat unit dictionary 24. When the recognition result in units of one beat input to the calculation means 8 matches the word in units of one beat associated with the transition, the transition is performed. Each transition holds a flag FL2. Numbers such as FL2 [1], FL2 [2],... Are assigned to the flag FL2. The number of the flag FL2 is associated with the number of the transition that holds it. For example, FL2 [3] is associated with transition T203. The flag FL2 is a variable indicating whether or not a transition that holds the flag FL2 is possible. If the flag FL2 is 1, a transition to hold it is possible. If the flag FL2 is 0, no transition to hold it is possible. That is, the possible transition is changed by setting the flag FL2. The flag FL2 is set according to a valid floor name based on the list L1 and the list L2. In FIG. 6, since all the flags FL2 are 1, all transitions are possible. If all transitions are possible, all floor names can be accepted. In order to change the accepted floor name, it is necessary to optimally set the flag FL2 held by each transition.

  FIG. 7 is a table showing an example of a transition list L1 necessary for accepting a floor name according to the first embodiment of the present invention. FIG. 7 shows a case where the floors present in the building are the first to fifth floors as an example. Therefore, the list L1 in FIG. 7 shows necessary transitions for each word indicating the floor names of the first to fifth floors.

  FIG. 8 is a table showing an example of a list L2 of valid floor names in the first embodiment of the present invention. FIG. 8 shows a case where the floors existing in the building are from the first floor to the fifth floor, and the car 1 whose traveling direction is upward is stopped on the second floor. Therefore, the list L2 in FIG. 8 shows words indicating floor names from the third floor to the fifth floor.

  FIG. 9 is an explanatory diagram illustrating a second example of the state transition 28 according to Embodiment 1 of the present invention. FIG. 9 shows a case where the flag FL2 of the state transition 28 shown in FIG. 8 is set based on the list L1 shown in FIG. 7 and the list L2 shown in FIG. In FIG. 9, FL2 [1], FL2 [2], and FL2 [6] are set to 0. For this reason, transition T201, transition T202, and transition T206 are impossible transitions. Transition T201, transition T202, and transition T206 are transitions necessary only for accepting invalid floor names.

  The update of the state transition 28 in the computing means 8 will be described. In the update of the state transition 28 in the calculation means 8, the 1-beat unit validity determination means 18 performs the selection process three times. When the recognition result in units of one beat is input from the voice input unit 3 to the calculation unit 8, the one-beat unit validity determination unit 18 performs a first selection process. In the first selection process, the one-beat-unit validity determination unit 18 selects all the transitions of the current state in the state transition 28. In the second selection process, the one-beat-unit validity determination unit 18 selects all transitions in which the flag FL2 held is “1” among the transitions selected in the first selection process. When even one transition selected in the second selection process exists, the one-beat-unit validity determination unit 18 performs the third selection process. In the third selection process, the one-beat-unit validity determination unit 18 uses the same one-beat unit word as the recognition result in one-beat unit input to the calculation unit 8 among the transitions selected in the second selection process. Select all associated transitions. When at least one transition selected in the third selection process exists, the one-beat-unit validity determination unit 18 updates the state transition 28. Specifically, the current state of the state transition 28 is set as a transition destination state by the transition selected in the third selection process. Thereafter, the calculation means 8 outputs information indicating that a transition has been made. Note that if there is no transition selected in the second or third selection process, the computing means 8 outputs that no transition has been performed.

  FIG. 10 is a flowchart of the update operation of the state transition 28 in the first embodiment of the present invention. Hereinafter, the update operation of the state transition 28 will be described by exemplifying a case where the recognition results in units of one beat first input to the computing unit 8 are “yo” and “go”. Here, it is assumed that the floors existing in the building are from the first floor to the fifth floor, and the car 1 whose traveling direction is upward is stopped on the second floor. In addition, it is assumed that all floors of the elevator are service floors. At this time, the list L1 is as shown in FIG. 7, the list L2 is as shown in FIG. 8, and the flag FL2 is set as shown in FIG. In addition, since the input to the calculation means 8 has not been performed yet, the current state in the state transition 28 is the state 1.

  The one-beat-unit validity determining unit 18 selects the transition T201, transition T202, transition T203, transition T204, and transition T205 that the current state 1 has (step S301). The one-beat-unit validity determination means 18 selects a transition having a flag FL2 that is “1” among the transitions selected in step S301 (step S302). Here, since FL2 [1] = 0, FL2 [2] = 0, FL2 [3] = 1, FL2 [4] = 1, FL2 [5] = 1, in step S302, transition T203 and transition T204 are performed. And transition T205 is selected. The one-beat-unit validity determining unit 18 determines whether or not the transition selected in step S302 exists (step S302). If it is determined in step S303 that the transition selected in step S302 exists, the one-beat-unit validity determination unit 18 transitions associated with “yo” and “go” among the transitions selected in step S302. Is selected (step S304). In step S304, a transition T204 associated with “yo” and a transition T205 associated with “go” are selected. Subsequently, the one-beat-unit validity determination unit 18 determines whether or not the transition selected in step S304 exists (step S305). If it is determined in step S305 that the transition selected in step S304 exists, the one-beat-unit validity determination unit 18 sets the transition destination by the transition selected in step S304 as the current state (step S306). In step S306, state 3 which is the transition destination by transition T204 and state 4 which is the transition destination by transition T205 become the current state. In this way, the state transition 28 is updated. Subsequent to step S306, the one-beat-unit validity determination unit 18 outputs a message indicating that the transition has been performed (step S307). If it is determined in step S303 that the transition selected in step S302 does not exist, the process proceeds to step S308. If it is determined in step S305 that the transition selected in step S304 does not exist, the process proceeds to step S308. In step S308, the one-beat-unit validity determination unit 18 outputs that the transition has not been performed.

  The initialization process of the calculation means 8 will be described. The calculation means 8 initializes the state transition 28, derives a valid floor name, and sets the flag FL2 as initialization processing. Specifically, in the initialization process, the calculation means 8 performs the following operation. Set the flag FL1 to 0. The current state of state transition 28 is set to the initial state. All the flags FL2 held by each transition of the state transition 28 are set to 0. Management information is acquired from the management information acquisition means 7. The floor name of the service floor is acquired as management information. At this time, the floor name of the service floor is given as a character string. As the management information, the floor name of the floor existing in the traveling direction of the car 1 is acquired. At this time, the floor name of the floor existing in the traveling direction of the car 1 is given as a character string. The common name of the acquired floor name of the service floor and the floor name of the floor existing in the traveling direction of the car 1 is registered in the list L2. Among the transitions in the list L1, all transitions corresponding to the floor names registered in the list L2 are selected. The flag FL2 held by the selected transition is set to 1. This allows only valid floor names to be accepted.

  FIG. 11 is a flowchart of the initialization process of the computing means 8 in the first embodiment of the present invention. Hereinafter, the initialization process of the computing means 8 will be described by exemplifying a case where the floors existing in the building are the first floor to the fifth floor and the car 1 whose traveling direction is upward is stopped on the second floor. In addition, it is assumed that all floors of the elevator are service floors. At this time, the list L1 is as shown in FIG. 7, and the list L2 is as shown in FIG.

  The calculating means 8 sets the flag FL1 held by the one-beat unit validity determining means 18 to 0 (step S401). The computing unit 8 sets the current state of the state transition 28 held by the one-beat unit validity determining unit 18 to the initial state. The calculation means 8 sets all the flags FL2 [1] to FL2 [9] of the state transition 28 held by the one-beat unit validity judgment means 18 to 0 (step S403). The computing unit 8 acquires the floor names “Ikkai”, “Nikai”, “Sankai”, “Yongkai”, and “Gokai” of the service floor from the management information acquisition unit 7 (step S404). The calculation means 8 acquires the floor names “Sankai”, “Yonkai” and “Gokai” of the floors existing in the traveling direction of the car 1 from the management information acquisition means 7 (step S405). The calculation means 8 registers the common floor name among the floor names of the service floor and the floor names of the floors existing in the traveling direction of the car 1 in the list L2 (step S406). In step S406, “Sankai”, “Yongkai” and “Gokai” are registered in the list L2. The computing means 8 selects all transitions corresponding to “Sankai”, “Yongkai”, and “Gokai” registered in the list L2 among the transitions in the list L1 (step S407). In step S407, as shown in FIG. 7, transitions T203, T204, T205, T207, T208, and T209 are selected by combining the transitions necessary for accepting “Sankai”, “Yongkai”, and “Gokai”. The calculating means 8 sets the flag FL2 held by the transition selected in step S407 to 1 (step S408). In step S408, flags FL2 [3], FL2 [4], FL2 [5], FL2 [7], FL2 [8] and FL2 [9] corresponding to the transitions T203, T204, T205, T207, T208 and T209 are set. Set to 1. As a result, as shown in FIG. 9, only “Sankai”, “Yongkai”, and “Gokai” among the floor names are accepted.

  FIG. 12 is a list of times held by the door control unit 11 according to the first embodiment of the present invention. The door control means 11 holds an initial door opening time t1, a door opening time t2, a door opening extension time t3, and an elapsed time t4. t1, t2, t3, and t4 are all expressed in seconds.

  The initial door opening time t1 is a time set in advance as a time from when the door of the car 1 is completely opened until the door starts to close. The initial door opening time t1 is set, for example, as the number of seconds longer than the time it takes for the user to start inputting voice for one beat after the door of the car 1 has been opened. The initial door opening time t1 is set to 4 seconds, for example. The door opening time t2 is a time stored during the elevator operation as a time from when the door of the car 1 is completely opened until the door starts to close. That is, the door opening time t2 is the time for actually opening the door of the car 1 during the elevator operation. The door opening time t2 is updated when it is extended by the door control means 11. The door opening extension time t3 is an upper limit of the time during which the door opening time t2 can be extended at a time. The door opening extension time t3 is set, for example, as the number of seconds that is equal to or longer than the time required to input one beat of sound. For example, the time required to input one beat of sound is obtained from sound data. The door opening extension time t3 is set to 1 second, for example. The elapsed time t4 is the time elapsed from the completion of the door opening of the car 1 to the present time. The elapsed time t4 is stored in the door control means 11 during elevator operation. The elapsed time t4 is sequentially updated while the car 1 is open.

  FIG. 13 is a flowchart of the update operation for the elapsed time t4 in the first embodiment of the present invention. Hereinafter, the update operation of the elapsed time t4 by the door control means 11 will be described with reference to FIG.

  The update operation of the elapsed time t4 is started simultaneously with the completion of the door opening of the car 1. The door control means 11 sets the number of seconds of the initial door opening time t1 as the door opening time t2 (step S501). The initial door opening time t1 is, for example, 4 seconds. Subsequent to step S501, the door control unit 11 acquires the current time by the time acquisition unit 10 (step S502). The current time is expressed in hours, minutes and seconds. Subsequent to step S502, the door control means 11 sets the current time acquired in step S502 as the door opening completion time (step S503). The door opening completion time is expressed in hours, minutes and seconds. Subsequent to step S503, the door control unit 11 acquires the current time by the time acquisition unit 10 (step S504). Subsequent to step S504, the door control means 11 obtains the difference between the current time obtained in step S504 and the door opening completion time in seconds. And the door control means 11 sets the said difference as elapsed time t4 (step S505). Subsequent to step S505, the door control means 11 determines whether or not the elapsed time t4 is equal to or longer than the door opening time t2 (step S506). In step S506, when it is determined that the elapsed time t4 is equal to or longer than the door opening time t2, the door control unit 11 ends the process. On the other hand, if it is determined in step S506 that the elapsed time t4 is less than the door opening time t2, the processes in steps S504 to S506 are repeated. In this way, the door control means 11 updates the elapsed time t4.

  FIG. 14 is a flowchart of the overall operation of the elevator control apparatus according to Embodiment 1 of the present invention. FIG. 14 shows an operation after the car 1 stops in response to a call. The operation shown in FIG. 14 is started simultaneously with the completion of the opening of the door of the car 1.

  First, the calculation means 8 performs an initialization process (step S601). Subsequent to step S601, the switching means 9 causes the storage means 5 to move to the 1-beat unit storage means 13, the speech recognition means 6 to the 1-beat unit speech recognition means 15, and the computing means 8 to the 1-beat-unit validity determination means 18. It is switched (step S602). Subsequent to step S602, the one-beat unit voice recognition means 15 performs voice recognition in units of one beat (step S603). Subsequent to step S603, it is determined whether or not a recognition result is output from the one-beat unit voice recognition unit 15 (step S604).

  If it is determined in step S604 that a word in units of one beat has been output as the recognition result, the recognition result is input to the calculation means 8 (step S605). Subsequent to step S605, it is determined whether or not a transition has been made in the state transition 28 held by the one-beat-unit validity determination unit 18 based on the input of the recognition result (step S606). The determination in step S606 is a determination of whether the recognition result in units of one beat is valid or invalid as a word constituting an effective floor name.

  If it is determined in step S606 that a transition has been made based on the input of the recognition result, it is determined in the state transition 28 whether or not the current state is an accepting state (step S607). If it is determined in step S607 that at least one of the current states is in the accepting state, it means that the voice input of the floor name may have been completed. On the other hand, when it is determined in step S607 that the current state is not the accepting state, it means that there is no possibility that the voice input of the floor name has been completed. That is, by confirming whether or not the current state has transitioned to the accepting state in step S607, it is detected that a set of words in units of one beat input so far is an effective floor name.

  If it is determined in step S607 that at least one of the current states is in the accepting state, the process proceeds to step S608. On the other hand, if it is determined in step S607 that the current state is not an accepting state, the process proceeds to step S610. In step S608, it is determined whether or not there is a state having a transition in which the flag FL2 to be held is “1” in the current state. That is, in step S608, it is determined whether there is a possible transition from the current state.

  If it is determined in step S608 that there is a possible transition from the current state, the process proceeds to step S609 in order to suspend the determination of whether the floor name voice input has been completed. In step S609, the flag FL1 is set to 1. By setting FL1 to 1, it can be determined whether or not the voice input of the floor name has been completed after confirming whether or not there is a voice input during the door opening extension time t3. Subsequent to step S609, the door control unit 11 determines whether the held door opening time t2 is less than the sum of the elapsed time t4 and the door opening extension time t3 (step S610).

  If it is determined in step S610 that the door opening time t2 is less than the sum of the elapsed time t4 and the door opening extension time t3, the door control means 11 extends the door opening time t2 (step S611). In step S611, the sum of the elapsed time t4 and the door opening extension time t3 is set as the door opening time t2. After step S611, the process proceeds to step S602. On the other hand, if it is determined in step S610 that the door opening time t2 is equal to or greater than the sum of the elapsed time t4 and the door opening extension time t3, the process proceeds to step S602. In this case, even if the door opening time t2 is not extended, the door opening is continued for the door opening extended time t3 or more from the present time, so the door control means 11 does not extend the door opening time t2.

  If it is determined in step S608 that there is no possible transition from the current state, it means that the voice input of the floor name has been completed. If it is determined in step S608 that there is no possible transition from the current state, the process proceeds to step S612. In step S612, the switching means 9 switches the storage means 5 to the word unit storage means 14, the speech recognition means 6 to the word unit speech recognition means 16, and the calculation means 8 to the word unit validity determination means 19. Subsequent to step S612, the word unit speech recognition means 16 performs word unit speech recognition (step S613). Subsequent to step S613, it is determined whether or not a recognition result is output from the word unit speech recognition means 16 (step S614).

  If it is determined in step S614 that a word has been output as the recognition result, the word unit speech recognition means 16 compares the recognition result with the list L2 (step S615). Then, the word unit speech recognition means 16 determines whether or not the word that is the recognition result is included in the list L2 (step S616). If it is determined in step S616 that the word that is the recognition result is included in the list L2, it means that the word is a valid floor name. If it is determined in step S616 that the word as a recognition result is not included in the list L2, it means that the word is an invalid floor name. That is, in step S616, it is determined whether or not the recognition result in units of words is a word indicating an effective floor name.

  If it is determined in step S616 that the word that is the recognition result is included in the list L2, the process proceeds to step S617. In step S617, the destination floor registration unit 12 registers a call with the floor indicated by the word determined to be a valid floor name in step S616 as the destination floor. Subsequent to step S617, the door control means 11 sends a door closing command to the door opening / closing means 4 (step S621). When the door opening / closing means 4 receives a door closing command from the door control means 11, the door opening / closing means 4 closes the door of the car 1.

  In step S604, if a signal indicating failure in speech recognition is output, the process proceeds to step S618. If no transition has been made in step S606, the process proceeds to step S618. In step S614, if a signal indicating failure in speech recognition is output, the process proceeds to step S618. If the word that is the recognition result is not included in the list L2 in step S616, the process proceeds to step S618.

  In step S618, the door control unit 11 determines whether or not the elapsed time t4 being held is equal to or longer than the door opening time t2. That the elapsed time t4 is equal to or greater than the door opening time t2 means that the door opening has been performed for the time period of the door opening time t2. If it is determined in step S618 that the elapsed time t4 is equal to or longer than the door opening time t2, the process proceeds to step S619. On the other hand, if it is determined in step S618 that the elapsed time t4 is less than the door opening time t2, the process proceeds to step S602.

  In step S619, it is determined whether or not the flag FL1 is 1. In step S619, the flag FL1 being 1 means that there is no new voice input during the pending determination of whether the voice input of the floor name has been completed. In this case, since it can be determined that the voice input of the floor name has been completed, the flag FL1 is set to 0 (step S620). After step S620, the process proceeds to step S612.

  In step S619, the fact that the flag FL1 is not 1 means that the door opening time t2 has elapsed without completing the voice input of the floor name. In this case, since it can be determined that the voice input of the floor name has been completed, the process proceeds to step S621.

  With reference to FIG. 14, the specific example of operation | movement of an elevator control apparatus is demonstrated. Hereinafter, a case where the user speaks “Yong Kai” and registers a call will be described. Here, the case where the floor existing in the building is from the first floor to the fifth floor, and the car 1 whose traveling direction is upward is stopped on the second floor is illustrated. In addition, it is assumed that all floors of the elevator are service floors. The initial door opening time t1 is 4 seconds and the door opening extension time t3 is 1 second. The first door opening time t2 is set to 4 seconds, and the elapsed time t4 is set to 0 seconds.

  The list L2 is set as shown in FIG. 8 and the flag FL2 is set as shown in FIG. 9 by the initialization process of the calculation means 8 in step S601. In step S601, the flag FL1 is set to 0.

  In step S602, the switching unit 9 switches the storage unit 5 to the one-beat unit storage unit 13, the voice recognition unit 6 to the one-beat unit voice recognition unit 15, and the calculation unit 8 to the one-beat unit validity determination unit 18. . If voice input by the user has not been performed at the time of step S603, the input sound is silent. In this case, the one-beat unit voice recognition means 15 performs voice recognition in units of one beat for silence. As a result, the voice recognition means 15 does not output the recognition result. Therefore, after step S604, the process proceeds to step S618. If the elapsed time t4 at the time of step S618 is 0.5 seconds, the elapsed time t4 is smaller than the door opening time t2 set to 4 seconds. In this case, after step S618, the process proceeds to step S602. Steps S602 to S618 are repeated until a voice input is performed by the user.

  The one-beat unit voice recognition means 15 performs voice recognition in units of one beat every time one voice is input. For example, the one-beat unit voice recognition unit 15 performs the process of step S603 four times between the start and end of the utterance “Yongkai” by the user.

  At the time when the user utters “Yo” out of “Yongkai”, the one-beat unit speech recognition means 15 performs voice recognition in units of one beat for the first input sound “Yo” (step S603). Here, it is assumed that “yo” and “go” are output from the one-beat unit voice recognition means 15 as recognition results. In this case, after step S604, the process proceeds to step S605.

  In step S 605, the recognition results “yo” and “go” are input to the calculation means 8. As shown in FIG. 9, the flag FL2 [3] of transition T204 corresponding to “Y” is 1. Further, the flag FL2 [5] of the transition T205 corresponding to “Go” is 1. That is, transition T204 and transition T205 are possible transitions. For this reason, the one-beat-unit validity determination unit 18 performs state transition and updates the current state to state 3 and state 4. Since the transition has been performed in step S605, the process proceeds to step S607 after step S606.

  Since both the current state, state 3 and state 4, are not in the accepting state, step S607 is followed by step S610. If the elapsed time t4 at the time of step S610 is 2.5 seconds, 3.5 seconds, which is the sum of the elapsed time t4 and the door opening extension time t3, is smaller than the door opening time t2 set to 4 seconds. . Therefore, after step S610, the process proceeds to step S602. In this case, the door opening time t2 is not extended.

  When the user speaks up to “Yong” of “Yongkai”, the one-beat-unit speech recognition means 15 performs voice recognition in units of one beat for the second input sound “N” (step S603). Here, it is assumed that “n” is output as a recognition result from the one-beat unit voice recognition means 15. In this case, after step S604, the process proceeds to step S605.

  In step S 605, the recognition result “n” is input to the calculation means 8. As shown in FIG. 9, the flag FL2 [7] of the transition T207 corresponding to “N” is 1. That is, the transition T207 is a possible transition. For this reason, the one-beat-unit validity determination unit 18 performs state transition and updates the current state to state 4. Since the transition has been performed in step S605, the process proceeds to step S607 after step S606.

  Since state 4 which is the current state is not an accepting state, step S607 follows step S610. If the elapsed time t4 at the time of step S610 is 3.1 seconds, 4.1 seconds, which is the sum of the elapsed time t4 and the door opening extension time t3, is larger than the door opening time t2 set to 4 seconds. . For this reason, after step S610, the process proceeds to step S611. In step S611, the door opening time t2 is extended by setting 4.1 seconds as the door opening time t2.

  When the user speaks up to “Yonka” of “Yonkai”, the one-beat-unit speech recognition means 15 performs voice recognition in units of one beat for the third input sound “F” (step S603). Here, it is assumed that “K” is output from the one-beat unit voice recognition means 15 as a recognition result. In this case, after step S604, the process proceeds to step S605.

  In step S <b> 605, the recognition result “K” is input to the calculation means 8. As shown in FIG. 9, the flag FL <b> 2 [8] of the transition T <b> 208 corresponding to “K” is 1. That is, transition T208 is a possible transition. For this reason, the one-beat-unit validity determination unit 18 performs state transition and updates the current state to state 5. Since the transition has been performed in step S605, the process proceeds to step S607 after step S606.

  Since the current state 5 is not an acceptance state, the process proceeds to step S610 after step S607. If the elapsed time t4 at the time of step S610 is 3.6 seconds, 4.6 seconds, which is the sum of the elapsed time t4 and the door opening extension time t3, is larger than the door opening time t2 set to 4 seconds. . For this reason, after step S610, the process proceeds to step S611. In step S611, the door opening time t2 is extended by setting 4.6 seconds as the door opening time t2.

  When the user utters “Yongkai”, the 1-beat unit speech recognition means 15 performs speech recognition in 1-beat units for the fourth input sound “I” (step S603). Here, it is assumed that “i” is output as a recognition result from the one-beat unit voice recognition means 15. In this case, after step S604, the process proceeds to step S605.

In step S 605, the recognition result “I” is input to the calculation means 8. As shown in FIG. 9, the flag FL <b> 2 [9] of the transition T <b> 209 corresponding to “I” is “1”. That is, transition T209 is a possible transition. For this reason, the 1-beat unit validity determination means 18 performs a state transition and updates the current state to the state 6. Since the transition has been performed in step S605, the process proceeds to step S607 after step S606. Since the current state 6 is an acceptance state, the process proceeds to step S608 after step S607. Since the transition is not possible from state 6, the next step S608, the process proceeds to step S 6 12.

  In step S612, the switching means 9 switches the storage means 5 to the word unit storage means 14, the speech recognition means 6 to the word unit speech recognition means 16, and the calculation means 8 to the word unit validity determination means 19. In step S <b> 613, the word unit speech recognition means 16 performs speech recognition on a word basis for the integrated input sound “Yongkai” stored in the input sound buffer 17. Here, it is assumed that “Yongkai” is output from the word unit speech recognition means 16 as a recognition result. In this case, after step S614, the process proceeds to step S615.

  Since the recognition result “Yongkai” is included in the list L2, the process advances from step S616 to step S617. Then, the destination floor registration means 12 registers “Yongkai” as the destination floor. In step S621, the door control means 11 sends a door closing command to the door opening / closing means 4.

  FIG. 15 is an explanatory diagram showing a third example of the state transition 28 according to Embodiment 1 of the present invention. FIG. 15 shows a case where the floors existing in the building are from the first floor to the third floor, and the first floor and the second floor exist in the traveling direction of the car 1. In addition, it is assumed that all floors of the elevator are service floors. The floor names of the first floor, the second floor, and the third floor are represented as “ichi”, “ni”, and “sun”, respectively. At this time, valid floor names are “1” and “d”. Hereinafter, the operation of the elevator control apparatus when “i” and “d” are output as the first recognition result by the one-beat unit speech recognition means 15 will be described with reference to FIG.

  In step S <b> 605, the recognition results “I” and “D” are input to the calculation means 8. As shown in FIG. 15, the flag FL2 [81] of the transition T801 corresponding to “I” and the flag FL2 [82] of the transition T802 corresponding to “D” are “1”. That is, transition T801 and transition T804 are possible transitions. For this reason, the one-beat-unit validity determination unit 18 performs state transition and updates the current state to state 2 and state 4. Since the transition has been performed in step S605, the process proceeds to step S607 after step S606.

  In step S607, it is determined whether or not the current state is an accepting state. Although state 2 is not an acceptance state, since state 4 is an acceptance state, step S607 is followed by step S608. In step S608, it is determined whether there is a possible transition from the current state. Since the state 2 has the possible transition T804, the process proceeds to step S609 after step S608. In step S609, FL1 is set to 1.

  Thereafter, “chi” is input within the door opening time t2, and if it is determined in step S608 that the voice input of the floor name has been completed, the input “chi” after the flag FL1 is set to 1 is also included. Word unit speech recognition and word unit validity determination are performed.

  On the other hand, if it is not determined in step S608 that the voice input of the floor name has been completed within the door opening time t2, it is determined in step S618 that the door opening time t2 has elapsed, and then the flag FL1 is checked in step S619. Is done. Since FL1 = 1, it is determined in step S619 that the determination as to whether the voice input of the floor name has been completed is suspended. Then, after FL1 is set to 0 in step S620, word unit processing is performed using the input sound that has already been input.

  In the first embodiment, the elevator control device includes a storage unit 5, a voice recognition unit 6, a calculation unit 8, and a door control unit 11. The memory | storage means 5 has memorize | stored the floor name of the building in which the elevator was installed. The voice recognition means 6 recognizes the voice in units of one beat by comparing the voice input to the voice input means 3 that accepts voice input with the floor name stored in the storage means 5 in units of one beat. To do. The calculation means 8 determines whether the recognition result in units of one beat by the voice recognition means 6 is valid or invalid as a word constituting the floor name. The door control means 11 extends the door opening time of the car 1 when the calculation means determines that the recognition result in units of one beat is valid. The door control means 11 does not extend the door opening time of the car 1 when the calculation means determines that the recognition result in units of one beat is invalid. That is, the elevator control device determines whether or not to extend the door opening time for each beat of the voice spoken by the elevator user. As a result, the door opening time of the car 1 can be minimized. As a result, the extra waiting time of the elevator user can be shortened.

  In the first embodiment, the elevator control device includes management information acquisition means 7 and destination floor registration means 12. The management information acquisition means 7 acquires elevator management information. The voice recognition unit 6 compares the voice already input to the voice input unit 3 with the floor name stored in the storage unit 5 in units of words after input of the voice is completed. Recognize with. Based on the management information acquired by the management information acquisition unit 7, the calculation unit 8 determines whether the recognition result in units of words by the voice recognition unit 6 is a word indicating an effective floor name. The destination floor registration unit 12 registers the destination floor of the car 1 based on the determination on the recognition result in units of words by the calculation unit 8. For this reason, according to the elevator control apparatus in Embodiment 1, it is possible to accurately register the destination floor by voice recognition while reducing the extra waiting time of the user.

  In the first embodiment, the elevator control device includes a one-beat unit speech recognition unit 15 included in the speech recognition unit 6, a word unit speech recognition unit 16 included in the speech recognition unit 6, and a switching unit 9. The one-beat unit voice recognition means 15 recognizes voice in units of one beat. The word unit speech recognition means 16 recognizes speech in units of words. The switching means 9 switches the operation of the voice recognition means 6 from that by the one-beat unit voice recognition means 15 to that by the word unit voice recognition means 16 after completing the voice input of the floor name. For this reason, the elevator control apparatus can perform the operation for registering the destination floor accurately after performing the operation for reducing the extra waiting time of the user.

Embodiment 2. FIG.
FIG. 16 is a configuration diagram of an elevator system including the elevator control device according to Embodiment 2 of the present invention. Hereinafter, the configuration of the elevator control device will be described focusing on the differences from the first embodiment. The same parts as those in the first embodiment are denoted by the same reference numerals, and a part of the description is omitted.

  The elevator control device includes a syllable unit storage unit 53, a syllable speech recognition unit 55, and a syllable unit validity determination unit 58. The one syllable unit storage means 53, the one syllable speech recognition means 55, and the one syllable unit validity determination means 58 process the building floor name in word units and division units smaller than words. In Embodiment 2, the division unit smaller than the word is one syllable. For example, the floor name indicating the first floor is processed by being divided into three words “I”, “K” and “I” in units of one syllable.

  FIG. 17 is a flowchart of the overall operation of the elevator control apparatus according to Embodiment 2 of the present invention. The operation shown in FIG. 17 is the same as that in FIG. 14 except that the processing is performed in units of one syllable instead of in units of one beat.

  In Embodiment 2, one syllable is used as a division unit smaller than a word. Also in this case, the same effect as in the first embodiment can be obtained.

  In the second embodiment, the elevator control device includes a syllable unit speech recognition unit 55 included in the speech recognition unit 6, a word unit speech recognition unit 16 included in the speech recognition unit 6, and a switching unit 9. The one syllable unit speech recognition means 55 recognizes speech in units of one syllable. The word unit speech recognition means 16 recognizes speech in units of words. The switching means 9 switches the operation of the voice recognition means 6 from the one by the syllable unit voice recognition means 55 to the one by the word unit voice recognition means 16 after completing the voice input of the floor name. For this reason, the elevator control apparatus can perform the operation for registering the destination floor accurately after performing the operation for reducing the extra waiting time of the user.

  In the first and second embodiments, the voice recognition means 6 includes an input sound buffer 17. The input sound buffer 17 stores the sound input to the sound input means 3 in units of division smaller than words. The voice recognizing unit 6 integrates the divided unit voices stored in the input sound buffer 17 after the voice input to the voice input unit 3 is completed, and the floor name stored in the storage unit 5. Is recognized in units of words, thereby performing speech recognition in units of words. For this reason, the speech recognition means 6 can perform both speech recognition in units of division and speech recognition in units of words for a single speech input.

  In the first and second embodiments, the speech recognition means 6 can output a plurality of words as recognition results in units smaller than words. Hereinafter, the reason why the voice recognition unit 6 can output a plurality of words as a recognition result will be described. In the case of speech recognition for each word, even if words that are not clear are mixed in the speech uttered by the user, if the utterances of the remaining words are clear, the overall similarity may be high. In this case, words as intended by the user are output as recognition results. However, in speech recognition in units of 1 beat or 1 syllable, if unclear speech is input, the similarity with the word different from the word intended by the user is the highest and it is considered that erroneous recognition is caused. . Then, it is determined to be invalid in the determination in step S606 or step S706, and there is a high possibility that the subsequent sound cannot be input. For this reason, if a plurality of words are output as a recognition result in a division unit smaller than a word and used for determining whether or not the word is valid as a word constituting a floor name, it is impossible to input a subsequent sound. Can be suppressed.

  In the first and second embodiments, the elevator management information includes the floor name existing in the traveling direction of the car 1 and the floor name of the service floor of the elevator. For this reason, the elevator control apparatus can appropriately determine an effective floor name.

  The management information may include only one of the floor name existing in the traveling direction of the car 1 or the floor name of the elevator service floor. Also in this case, a valid floor name can be determined appropriately. When the management information includes only the floor name of the floor existing in the traveling direction of the elevator, the floor name of the floor existing in the traveling direction of the elevator is used as an effective floor name as it is in step S406 of FIG. It will be registered in the list L2. If the management information includes only the floor name of the service floor, the floor name of the service floor is registered in the list L2 as an effective floor name as it is in step S406 of FIG.

  The management information may include a floor name of a floor that does not exist in the traveling direction of the car 1 and a floor name that is not a service floor of the elevator. In this case, if a list of floor names of existing floors is stored in the computing means 8 or the like, an effective floor name can be appropriately determined. Specifically, in step S406 of FIG. 11, among floor names of existing floors, floor names that are neither floor names that are not in the elevator traveling direction nor floor names that are not service floors. Only the floor name may be selected and registered in the list L2 as an effective floor.

  The management information may include either one of a floor name that does not exist in the traveling direction of the car 1 or a floor name that is not a service floor of the elevator. Also in this case, a valid floor name can be determined appropriately. When the management information includes only floor names of floors that do not exist in the traveling direction of the elevator, floors that do not exist in the traveling direction of the car 1 among the floor names of existing floors in step S406 of FIG. A floor name not included in the floor name may be registered in the list L2 as an effective floor name. If the management information is only the floor name of the floor that is not serviced, it is not included in the floor names of the floors that are not service floors among the floor names of existing floors in step S406 of FIG. The floor name may be registered in the list L2 as an effective floor name.

  The present invention is not limited to the embodiment described above. For example, floor names that do not use numbers such as “Lobby” and “Tenbokai” may be used instead of floor names using numbers. Also in this case, the same effect as the above-described embodiment can be obtained.

  1 car, 2 elevator control device, 3 voice input means, 4 door opening / closing means, 5 storage means, 6 voice recognition means, 7 management information acquisition means, 8 calculation means, 9 switching means, 10 time acquisition means, 11 door control means , 12 destination floor registration means, 13 1 beat unit storage means, 14 word unit storage means, 15 1 beat unit speech recognition means, 16 word unit speech recognition means, 17 input sound buffer, 18 1 beat unit validity judgment means, 19 words Unit validity determination means, 20 word unit dictionary, 21 word unit voice recognition threshold, 22, 26 acoustic model, 23, 27 label, 24 one beat unit dictionary, 25 one beat unit voice recognition threshold, 28 state transition, 53 1 syllable unit storage means, 55 1 syllable unit speech recognition means, 58 1 syllable unit validity determination means, FL1 completion determination hold flag, FL2 flag, List of transitions required for acceptance of 1 floor name, L2 list of valid floor name, t1 initial door-open time period, t2 door-open time, t3 door-open extension time, t4 elapsed time

Claims (10)

Storage means for storing the floor name of the building where the elevator is installed;
Voice recognition means for recognizing the voice in units of division by comparing the voice input to the voice input means for accepting voice input and the floor name stored in the storage means in units of division smaller than words. When,
A calculation means for determining whether the recognition result in the division unit by the voice recognition means is valid or invalid as a word constituting an effective floor name;
When the calculation means determines that the recognition result in the division unit is valid, it extends the door opening time of the car, and when the calculation means determines that the recognition result in the division unit is invalid, Door control means that does not extend the door opening time;
Elevator control device. Management information acquisition means for acquiring the management information of the elevator;
A destination floor registration means for registering the destination floor of the car;
The voice recognition means recognizes the voice in word units by comparing the voice input to the voice input means and the floor name stored in the storage means in word units,
The calculation means determines whether the recognition result in units of words by the voice recognition means is a word indicating an effective floor name based on the management information acquired by the management information acquisition means,
The elevator control device according to claim 1, wherein the destination floor registration unit registers the destination floor of the car based on a determination on a recognition result in units of words by the calculation unit.   The voice recognition means includes an input sound buffer that saves the voice input to the voice input means for each of the divided units, and is stored in the input sound buffer after completion of voice input to the voice input means. The elevator control device according to claim 2, wherein a unit of the divided unit voices and a floor name stored in the storage unit are compared in word units.   The elevator control device according to claim 2 or 3, wherein the management information includes a floor name of a floor existing in the traveling direction of the car or a floor name of a floor not existing in the traveling direction of the car.   The elevator control device according to any one of claims 2 to 4, wherein the management information includes a floor name of an elevator service floor or a floor name that is not an elevator service floor.   The elevator control device according to any one of claims 1 to 5, wherein the division unit is one beat.   The elevator control device according to claim 1, wherein the division unit is one syllable. The voice recognition means has a one-beat unit voice recognition means for recognizing a voice in one-beat units;
The speech recognition means has a word unit speech recognition means for recognizing speech in units of words;
A switching means for switching the operation of the voice recognition means from that by the one-beat unit voice recognition means to that by the word unit voice recognition means after the voice input of the floor name is completed;
The elevator control device according to claim 6 provided with. The speech recognition means has one syllable unit speech recognition means for recognizing speech in units of one syllable;
The speech recognition means has a word unit speech recognition means for recognizing speech in units of words;
Switching means for switching the operation of the voice recognition means from that by the one syllable unit voice recognition means to that by the word unit voice recognition means after the voice input of the floor name is completed;
The elevator control device according to claim 7 provided with.   The elevator control device according to any one of claims 1 to 9, wherein the voice recognition unit is capable of outputting a plurality of words as a recognition result in the division unit.

Images