• 专利附录
  • 专利说明
  • 权利要求
  • 类似技术
  • 同族专利
  • 引用文献
  • 法律状态
  • 优先权
    • 专利摘要:
    The invention relates to a method and a radio system for the identification and testing of radio messages (M1 ... M3) as well as for the assignment of radio messages (M1 ... M3) to vehicles (F1 ... F3), wherein in each case a speaker at a predetermined Place each voice message (M1 ... M3) indicating the identifier (K). According to the invention, it is provided that a) a number of delivered radio messages (M1 ... M3) is recorded, - wherein in each case the identifier (K) contained in the radio message (M1 ... M3) by means of speech recognition (0) in a digital identifier (Kd) is transformed, - wherein from those voice radio messages (M1 ... M3), each of which the same digital identifier has been assigned, a biometric data record (B1 ... B3) is extracted, and - said biometric data set (B1 ... B3) is assigned to the respective digital identification (Kd), and b) a further speech message (M4) is subsequently recorded, - wherein a further biometric data record (B4) is extracted from the further speech message (M4), wherein among the stored biometric data sets (B1 ... B3) is searched for that biometric data record (B1), which coincides with the further biometric data set (B4) best and the radio message (M4) with that vehicle (F1) the biometric record (B1) assigned identifier (Kd) is assigned.
    公开号:AT516219A1
    申请号:T50621/2014
    申请日:2014-09-09
    公开日:2016-03-15
    发明作者:
    申请人:Frequentis Ag；
    IPC主号:
    专利说明:

    The invention relates to a method for the identification and testing of radio messages according to the preamble of claim 1.
    Furthermore, the invention relates to a device for identification and testing of radio messages according to the preamble of claim 7.
    Finally, the invention relates to a data carrier according to claim 12.
    In control centers with communication systems, automated detection of the respective communicating subscriber is desirable. Particularly in communication systems where an operator, e.g. Air traffic controller, in a control center receives the respective position of communication participants displayed on a monitor, it is desirable that when building a
    Speech radio communication simultaneously on the monitor, on which the position of the respective radio subscriber is displayed, the respective symbol of
    Communication subscriber can be highlighted. This facilitates the work of the operator and, through better situational awareness, can avoid errors and inconsistencies which, in extreme cases, can cause severe damage depending on the application.
    Typical known communication systems would be used, for example, in the area of rescue, fire, police, military, maritime or aviation communications. A typical application in the field of air traffic is shown in Fig. 1. The radio communication between a center Z and the individual vehicles F ^ F2, F3, in the present case aircraft, is handled via a radio data link. In particular, speakers located in the vehicles F ^ F2, F3 are radio messages M ^ M2, M3, M4 via radio to a receiving device S assigned to the center Z, which receives the radio messages M1; M2, M3 forwarded to the control panel. In the center Z, a voice message M arrives, which has an identifier K standing at its beginning or alternatively also at the end, as well as an information part or speech content I. (Fig. 2)
    In conventional methods from the prior art, the identifier, as shown in Fig. 3darisiert digitized and a speech recognition unit 0 is supplied, which determines from the identifier K a digitized identifier Kd. In addition, methods known in the art also provide a probability value Ps indicating the probability with which the digitized identifier Kd was detected.
    From the prior art it is known to subject voice messages to speech recognition. In this case, an identifier, which is typically at the beginning of each voice message, is used, and from this identifier a digital identifier is created which contains the content of this identifier in digitized form.
    Thus, as shown in Fig. 2, the identifier K, which prior to its detection is merely in the form of a broadcast signal, can be written to a related digital identifier Kd, in the present case the value " LH 430 " has to be transformed.
    The known prior art will be explained in more detail with reference to FIGS. 1 to 3. Fig.1 shows a known radio system. Fig. 2 shows schematically a voice message arriving at the central station. Fig. 3 shows the determination of a digital identifier according to the prior art.
    A major problem with this approach is that the speakers mispronounce their identifier with a certain probability, and that speech recognition, especially in the case of blurred pronunciation or poor speech connection, leads to misrecognition of the digital identifier. In this case, the operator in the center Z is incorrectly informed about his respective communication partner, which can sometimes lead to dangerous situations or even accidents. The operator can avoid errors and inconsistencies by better situational awareness.
    The object of the invention is therefore to remedy these problems and to provide a method and a system for identifying and checking voice communications, as well as for the allocation of radio messages from vehicles, in which the probability of mis-assignment of radio messages to vehicles is significantly reduced, and the situational perception of the respective Improve operator.
    The invention solves this problem in a method of the type mentioned above with the characterizing feature of claim 1.
    According to the invention, in a method for the identification and testing of radio messages and for the assignment of radio messages to vehicles, in each case an in-vehicle speaker at a given location, in particular at the beginning of each radio message, indicates the identification of its vehicle, provided that a) Number of delivered radio messages is recorded, - in each case the spoken identification (K) of the respective vehicle contained in the radio message is transformed into a digital identifier by means of speech recognition, and this digital identifier is assigned to the respective voice message, - of which those radio messages each having the same digital Identifier is assigned, a biometric data record is extracted which characterizes the voice biometry of the respective speaker, and - this biometric data record is assigned to the respective digital identifier and in a data file nbank is stored, and b) after another voice message is recorded, - is extracted from the further voice message, another biometric record that characterizes the vocal biometry of the respective speaker of the further speech message, - wherein biometrics records stored in the database after that biometrics Data record which is best matched and determined with the further biometric data record, and the radio message (M4) is assigned to the vehicle with the identifier assigned to this biometric data record.
    In order to give the operator a better idea of the reliability of the identification and checking of the radio message as well as the assignment of the speech message to a vehicle, it may be provided that the biometric data record is compared with the further biometric data record and the probability of the biometrics Record and the further biometrics record from the same speaker.
    In order to obtain a further improvement and confirmation of the assignment, provision may be made for each further identification of the respective vehicle, which is included in the further speech message, to be transformed into another digital identifier by means of speech recognition, and this further digital identifier is assigned to the respective further radiotelephone message; This confirms the accuracy of the detection by reducing the probability of false positives on the one hand and the recognition rate on the other hand overall is improved.
    In order to further improve the assignment during ongoing operation of the system, it can be provided that upon detection of a radio message with a digital identifier for which a biometric data record has already been determined, the biometric data record associated with the identifier of the radio message is adapted to the newly added radio message , The number of voice messages associated with the tag and / or the total duration of the radio messages is assigned to the biometric record. The likelihood of the detected match is considered to be greater the greater the number and voice messages have been determined with the digital identifier and / or the greater the total duration of the respective biometrics record associated radio messages with the digital identifier.
    In order to provide an advantageous solution in the event of discrepancies in the speech recognition used in the prior art and in the speech recognition proposed according to the invention, different or alternative approaches may be taken. It can be provided that, in the determination of the digital identifier from a voice message by means of voice recognition, the probability for the correctness of the respectively determined identifier by voice recognition is made.
    It is alternatively or additionally possible for this purpose to check whether the identifier determined by speech recognition and the identifier determined by means of comparison of the biometric data records differ from one another and, in the case of a difference, the determined probability with that of the biometric data record and the further biometric data record from the same speaker, and the likelihood of correct speech recognition being compared.
    Finally, it may alternatively or additionally be provided for the same purpose that the vehicle is assigned the identifier for which the greater likelihood speaks.
    In order to be able to manage new vehicles in operation with the communication center, it can be provided that for speech messages, the spoken identification of the respective vehicle means contained in the speech message is transformed into a digital identifier, and this digital identifier is assigned to the respective speech message, from those voice messages each having the same digital identifier, extracting a biometric record characterizing the vocal biometry of the respective speaker, and assigning that biometric record to the respective digital identifier and storing it in a database.
    The invention solves this problem in a radio communication system of the type mentioned above with the characterizing feature of patent claim 7.
    The invention relates to a radio communication system for identifying and checking voice messages and for allocating radio messages to vehicles. The radiotelephone system comprises a central station which is or can be brought into radio communication with a number of vehicles, wherein a respective speaker located in the vehicle in the course of the radio communication at a predetermined
    Place, especially at the beginning, each radio message indicating the identification of his vehicle.
    According to the invention, it is provided that a) the control center has a voice recognition unit which records incoming voice messages and which transforms the recognized vehicle identification into a digital identifier by means of voice recognition, the control center comprising a biometrics unit consisting of those radio messages each the same digital identifier has been assigned, a biometric data record which characterizes the voice biometry of the respective speaker, extracts it and delivers it at its output, and the central unit has an allocation unit which determines the digital identifier produced by the voice recognition unit and the biometrics created by the biometric unit. Record assigned and assigns the created digital identifier to the created biometrics record, - the central has a the allocation unit downstream database containing a biometrics record and the associated digital identification in and b) the biometric unit extracts from the further voice message, another biometric record characterizing the voice biometry of the respective speaker of the further radio message on its Output is available, - the database is downstream of the biometrics unit and looking at the next biometric data records at the entrance of the database for the biometrics record that most closely matches the other biometric data set, the database that biometric data set together with the identifier associated with this biometric dataset is available.
    In order to give the operator a better impression of the reliability of the identification and testing of the radiotelephone message as well as the assignment of the radiotelephone message to a vehicle, it may be provided that a biometric comparison unit, the biometric data set created by the biometrics unit and that determined by the database the biometric comparison unit compares the biometric data record and the further biometric data record and evaluates differences of the two biometric data records and wherein the biometrics comparison unit evaluates the probability with the biometric data record and the further one Biometric dataset originating from the same speaker, identified and available.
    In order to obtain a further improvement and confirmation of the assignment, it can be provided that, in the presence of a biometric data record and a digital identifier for which a biometric data record is already stored in the database, the allocation unit stores the data record stored in the database and associated with the digital identifier newly added speech radio message as well as their biometric record adapts.
    In order to further improve the allocation during the ongoing operation of the system, it may be provided that the allocation unit determines the number of intercom messages and / or the total duration of the intercom messages assigned to the biometric data record stored in the database, and that the database (D) the number and / or total duration assigned to the biometric data record is transmitted to the comparison unit, and that the comparison unit increases the probability of the recognized agreement, in particular by statistical methods, the greater the number and / or the greater the total duration of the respective one Biometric data record associated with radio communication messages with the digital identifier.
    In order to provide an advantageous solution in the event of discrepancies in the speech recognition and speech recognition proposed by the prior art, it may be provided that the speech recognition unit, upon receipt of a speech message together with the digital identifier, indicates a probability of correctness of the respectively determined identifier and that a probability comparison unit is provided which is supplied with the probability determined by the speech recognition unit for the correctness of the respectively determined digital identifier and the probability determined by the biometric comparison unit, wherein the comparison unit determines which of the two probabilities fed to it is larger and outputs a signal to its output, and that a selection unit is provided, the signal generated by the probability comparison unit, the digi detected by the speech recognition unit tale identifier as well as the data derived from the database digital identifier are fed, and due to the incoming
    Signals each selects that digital identifier and at its disposal disposal, which is associated with the respective larger probabilities.
    Advantageously, a program for carrying out a method according to the invention is stored on a data carrier and is executed on one of the computers.
    Several advantageous embodiments of the invention will be illustrated with reference to FIGS. 4 to 7.
    Fig. 4 shows the procedure in the further processing of Sprechfunkmeldung.Fig. 5 shows the data structure stored in the database. Fig. 6 shows the determination of the probability of detection of a biometric data record due to an incoming voice message. Fig. 7 shows a combination of the types of determination of a digital identifier with improved accuracy shown in Figs. 3 and 6.
    In the embodiment of the invention shown in FIG. 4, a voice message Mi, M2 is read in and fed to a voice recognition unit 0 and to a biometrics unit 1. The speech recognition unit 0 supplies a digital identification Kd of the respective vehicle Fi,... F3 on the basis of its speech recognition. The biometric unit 1 provides a bioemetry record B which contains the biometric information essential to the voice biometry of the speaker of the radio message M. This assignment is submitted to the respective operator in the control center for confirmation.
    The operator thus receives the biometric data record B and the digital identifier Kd. Furthermore, the voice message is supplied to an allocation unit 5. If the operator confirms the assignment, the digital identifier Kd and the biometric data set Bi,... B3 are assigned to one another by the allocation unit 5 and stored in the database D.
    A possible embodiment of a structure of such a database D is shown in FIG. In each case, data records di,... D3 are combined in each case with a digital recognition Kd, a biometric data record Bd, a total speaking time td and a recognition counter nd.
    The vehicles Fi and F2, as shown in Fig. 1, respectively transmit voice messages M1; M2 to the communication center Z. As shown in Fig. 4, the two messages M ^ M2 of the speech recognition unit 0 are supplied, which the respective recognized text of the identifier K of the messages M1; M2 converts it into a digital identifier Kd1, Kd2 and keeps it at its input. Furthermore, the identifiers are each directed to a unit 1 for determining a biometric data record, which provides a biometric data record Bi, B2 at its output. The operator in the central office receives the request to confirm the affiliation of the speaker to the respective vehicle F with the recognized digital identifier Kd. For this purpose, in a particular embodiment of the invention, the identifier is displayed in each case and the operator must confirm this identifier. If the operator confirms the identifier Kd, then a data record di,... D3 is created by the allocation unit 5. The data record di,... D3 comprises the digital identifier Kd and the biometric data record Bd and is stored in the database D. In addition, the data set di,... D3 is also stored in the database D for the length of the respective radio message Mi, M2 and the number td of the radio messages IVL, M2 previously provided with the respective identification Kd.
    If a further radio message M4 is received, which has been sent by a vehicle F! Of which there is already a stored biometric data record Βί in the database D, the biometric data record Bi is adapted to the newly added radio message M4. The respective entries in the database D, in particular the record Bi associated with the identifier Kd, as well as the total duration td of the voice radio messages IVL, M4 and the number nd of the determined radio messages M1; M4 are adjusted and increased accordingly, in particular by the duration of the now received voice message M4 or by the value 1.
    The voice radio message M4 is fed to a biometric unit 1 which extracts from the further radio message M4 a further biometric data record B4 according to the same specification. The biometric data record B4 is present at the output of the biometric unit 1 and characterizes the voice biometry of the respective speaker of the other
    Voice message M4. Subsequently, a database query Q (B4) is created to the database D, which searches for the biometric data set Bi which best matches the further biometric data set B4. On the basis of the query Q (B4), the database D transmits the biometrics data set Βί of the radio message. The output of the database contains the biometric data record B 1 and the associated identifier Kd 1.
    The further biometric data record B4 and the biometric data record B: are fed to a biometric comparison unit 2, which determines and quantifies the differences between the biometric data record Bi and the further biometric data record B4. Based on this result, the biometric comparison unit 2 determines a probability value indicating the probability with the biometrics data set B! and the further record B4 originate from the same speaker.
    Such a comparison is disclosed, for example, in U.S. Patent No. US6411933, as well as at http://www.research.ibm.com/hlt/html/body_patents.html, and is well known to one of ordinary skill in the art. Further similar methods are disclosed, for example, in the following publications:
    Jain, A.K. ; Dept. of Comput. Be. &Amp; Eng., Michigan State Univ., USA; Ross, A.; Prabhakar, S., " An introduction to biometric recognition ", Circuits and Systems for Video Technology, IEEE Transactions on (Volume 14, Issue: 1), pages 4-20,
    Biometrics: Personal Identification in Networked Society, published by AnilK. Jain, Ruud Bolle, Sharath Pankanti, Kluwer Academic Press, ISBN 0-7923-8345-1
    With such a system illustrated in FIG. 6, it is only possible for an operator to easily recognize with what probability PB the recognized digital identifier Kd actually originates from the speaker of the vehicle F to which the identifier Kd is assigned should the speaker have his identifier K due to the determination of the digital identification Kd on the basis of the voice biometry of the speaker, this has no influence on the obtained identification Kd. Despite incorrect indication of the identifier k, the correct digital identifier Kd is determined.
    A particular embodiment of the invention, shown in Fig. 7, combines the advantages of the prior art method with the advantage of the invention. As known from the prior art arrangement of FIG. 3 also has the arrangement shown in FIG. 7 via a speech recognition unit 0 and via a unit 1 for determining a biometric data record B.
    In the following, two different case constellations are shown together, namely the case in which a speaker of the vehicle Fi with the identifier Ki indicates the correct identifier i at the beginning of his voice message M4, on the other hand, the case in which the speaker an incorrect identifier K2 at the beginning of his radio message M4 indicates.
    The voice radio message M4 arriving at the input of the system is supplied both to the speech recognition unit 0 and to the biometric data record B detection unit 1. At the output of the speech recognition unit 0 there is a digital identification Kd2 which corresponds to the identifier indicated by the speaker. In the event that the speaker has correctly named the identifier, the speech recognition unit detects the actual identifier K! In addition, the speech recognition unit 0 also outputs a probability value Ps indicating how likely it is that the respective speaker has actually given a voice message Mab containing the respective digital identifier Kd3 contains in spoken form. Depending on the nature of the radio channel and on the clarity of the pronunciation of the speaker, different probability values are obtained for this value. However, the possibility of recognizing intentional or unintentional visual mispronunciation of the respective identifier Kd by the speaker is not given when the speech recognition unit 0 is used solely to determine the identifier. With a clear and pronounced pronunciation and good transmission ratios, the output of the speech recognition unit 0 has a high probability value Ps. In the embodiment shown in Fig. 3, in the absence of correct indication of the identification of the vehicle due to the high probability value of Ps, the vehicle F2 would instead be marked on the screen of the operator instead of the vehicle F · in the case of correctly indicating the identification of the vehicle would the correct Identification - also with a high probability value Ps - be recognized, without further review would be possible.
    The embodiment of the invention shown in Fig. 7 further has, like the system shown in Fig. 6, a unit 1 for obtaining a biometric data set. At the output of the unit 1 for determining the biometric data set is the further biometric data B4. 4, a database query Q (B4) is sent to the database D, which searches for the biometrics data set B4 that best matches the further biometrics data set B4. The database D supplies at its output, on the one hand, the biometric record B-1 which best matches the identical memory and, on the other hand, the digital identifier Kd1 of the first vehicle Fi assigned to the biometric record B-1. The biometric data record Bi and the further biometric data record B4 are fed to the biometric comparison unit 2. This determines differences between the biometric data set B ^ and the further biometric data set B4 and consequently a probability PB, with which the biometric data set B · and the further biometry data set B4 originate from the same speaker.
    If the respective speaker has given the respective identifier Kd correctly, an association is created on the basis of the association between the correctly specified digital identifier Kd1 and the biometric data set B ^, which is independent of the specific, incorrect identification of the identifier by the speaker. Thus, a likelihood value PB is determined which indicates whether the speaker who has issued a digital ID Kd1 to a vehicle F! was assigned, continue in the same
    Vehicle is. The two probability values, namely the probability Ps of the correctness of the speech recognition and the probability value PB, with which the biometric data set Bi and the further biometric data set B4 derive from the same speaker, are compared with each other.
    The comparison unit 2, which is used to determine the correspondence of the biometric data set Bi and the further biometric data set B4, in this particular embodiment of the invention, furthermore, also the total time tds and the total number nd of the radio communications supplied. The greater the number nd of the radio messages M and the greater the total duration td of the speech messages M assigned to a respective biometric data record B, the probability P B becomes greater.
    If the two recognition methods arrive at the same result, the respective digital identifier Kd1 can be verified. Moreover, it can be stated that the recognized digital identifier Kd1 is very likely to be correct since both recognition methods have come to the same result. This probability can be assigned to the recognized digital identifier. Upon further processing, records created based on the particular identifier may be assigned that circumstance or probability. It is also possible that in the representation of the respective identifier, the probability and / or this circumstance, possibly symbolically represented.
    If in each case a different digital identification Kd1, Kd2 is determined by the two recognition methods, then that of the two identifiers Kd1, Kd2 is ultimately considered correct, for which the greater probability Ps, Pb respectively speaks. The two probabilities Ps, Pb are supplied to a probability comparison unit 3 which compares the two probabilities Ps, Pb with each other and a signal Se indicates at its output which of the two probabilities is larger. Further, a selection unit 4 is provided to which the signal Sei output from the probability comparison unit 3 is supplied is. The selection unit 4 is supplied with the identifier Kd2 detected by the voice recognition unit 0, and the identifier Kd1 supplied from the database D is supplied. The selection unit 4 selects those of the determined identifiers Kd1, Kd2 whose respective associated probability Ps, Pb is larger. Due to the greater likelihood PB, the correct identifier Kd1 of the vehicle F-t is determined in the present case despite false indication of the identifier Kd2 by the speaker in the vehicle Fi. At the output of the selection unit 4 is the digital identifier Kd1 of the vehicle F ·, on.
    权利要求:
    Claims (12)
    [1]
    Claims 1. A method for the identification and testing of radio messages (1 ^ ... 1 ^ 3) and for the assignment of radio messages (M1 ... M3) to vehicles ^ .. ^ 3), wherein in each case one of the vehicles (F ^ -Fs ) located in the course of radiotelephone traffic at a predetermined location, in particular at the beginning, each radiotelephone message (ΜΊ ... Μ3) each indicating the identifier (K) of its vehicle, characterized in that a) a number of delivered radio messages (1 ^ .. .1 ^ 3) is recorded, - wherein in each case the spoken identifier (K) of the respective vehicle (F1... F3) contained in the radiotelephone message (1 ^ ... 1 ^ 3) is converted into a digital identification (0) by means of speech recognition (0). Kd) is assigned, and this digital identification (Kd) is assigned to the respective speech message (M1 ... M3), - of which those radio messages (1 ^ ... 1 ^ 3) to which the same digital identifier was assigned in each case, a biometric Record (B ^ .- Bs) is extracted, the voice biometry of jew in which the biometric record (B ^ -. Ba) is assigned to the respective digital identifier (Kd) and stored in a database, and b) thereafter another voice message (M4) is recorded; another voice message (M4), another Biometrie record (B4) is extracted, which characterizes the vocal biometry of the respective speaker of the further speech message (M4), - wherein among in the database (D) stored biometric data sets ^ ... B3) after that Biometric data record (Bi) is found that best matches the other biometric data set (B4), and the radio message (M4) is assigned to the vehicle (F ^) with the identifier (Kd) assigned to this biometric data record (B ^).
    [2]
    2. The method according to claim 1, characterized in that the biometric data set (B-,) are compared with the further biometric data set (B4) and the probability (PB) is determined with the biometric data set (B ^ and the further biometry Dataset (B4) from the same speaker.
    [3]
    A method according to claim 2, characterized in that in each case a spoken further identifier (K4) of the respective vehicle contained in the further speech message (B4) is transformed into another digital identifier (Kd4) by speech recognition, and this further digital identifier (Kd4) of the respective one in the case of a correspondence of the biometric data record (B) with the further biometric data record (B4), the identification of the assignment of the speech radio message to the respective digital identification (Kd4; Kdi) is confirmed.
    [4]
    Method according to one of the preceding claims, characterized in that, upon detection of a radio communication message with a digital identification (Kd) for which a biometric data set (B ^ has already been determined which corresponds to the identification (Kd) of the radio message (M4) Biometric data record (B ^ is adapted to the newly arrived radiotelephone message (M4), wherein the number of the identification (Kd) associated detected radio communications (M1; M4) and / or the total duration of the radio messages to the biometric data set (B ^ is assigned the greater the number and radio messages with the correct digital identifier (Kd), and / or the greater the total duration (td) of the respective biometric data set (B ^), the greater the likelihood (PB) of the recognized match Speech radio messages (M1, M2) with the digital identifier (Kd) is.
    [5]
    5. Method according to one of the preceding claims, characterized in that, in the determination of the digital identification (Kd) from a voice message by means of speech recognition, the probability (Ps) of the correctness of the respective identifier (Kd) is determined by speech recognition (0); checking whether the identifier determined by speech recognition (1) and the identifier determined by means of comparison (2) of the biometric data records (B) differ from one another and a) in the case of a difference between the identifiers, the determined probability (PB) with the biometric data record (B ^ and the other biometric data set (B4) originate from the same speaker and the likelihood (Ps) of the correctness of speech recognition (O) are compared, and the vehicle (Fi) is assigned the identifier for which the greater probability (Ps Pb) speaks, and / or b) in the case of identical or identical identifiers the vehicle (F ^ this Ke and, if appropriate, this circumstance is displayed or communicated.
    [6]
    6. Method according to one of the preceding claims, characterized in that for voice radio messages (M3), in each case the identifier (K3) of the respective vehicle (F3) included in the radiotelephone message (M3) is transformed into a digital identifier (Kd3) by voice recognition (1) , and this digital identification number is assigned to the respective radiotelephone message, a biometric data set (B) is then extracted from those intercom messages (M) to which the same digital identifier (Kd3) has been assigned, which characterizes the vocal biometry of the respective speaker, and this Biometric data record (B3) of the respective digital identifier (Kd3) assigned and stored in a database (D).
    [7]
    7. radio system for the identification and testing of radio communications (Μί., Μ ^ zur ^ and for the assignment of voice radio messages (Μί., Μ ^ Fahrzeugen to vehicles (Fd ... F3), comprising a center (Z), with a number of Vehicles (Fi ... F3) inSprechfunkkommunikation is or can be brought, each one in the vehicle (Fd ... F3) located speaker in the course of radiotelephone traffic at a predetermined location, especially at the beginning of each voice message (M1 ... M3) respectively indicating the identifier (K) of its vehicle, characterized in that a) the center (Z) has a speech recognition unit (0) which records incoming speech messages (1 ^ ... 1 ^ 3) and which in the radio message (M1. .. M3) contained spoken identifier (K) of the respective vehicle (F ^ .- F ^ transformed by means of speech recognition in a digital identifier (Kd), - the control center (Z) has a biometry unit (1), which from those voice radio messages which each assign the same digital identifier (Kd) sen, a biometric data set (Β ^. ^ ^ which characterizes the voice biometry of the respective speaker is extracted and output at its output, and - the control center (Z) has an allocation unit (5) which receives the information from the voice recognition unit (0). created digital identifier (Kd) and the biometrics unit (1) created biometric data set (Bi .-. B ^ are fed and assigns the created digital identifier (Kd) to the created biometric data set (Bi.-.Ba), the central station (Z) has a database (D) connected downstream of the allocation unit (5), which jointly stores a biometric data record (B-1... B3) and the digital identifier (Kd) assigned to it in a database and provides the biometrics Data set (Βϊ .-. Β ^ with specification of the respective digital identifier (Kd) for retrieval available, and b) wherein the biometrics unit (1) from the further radio message (M4), another biometric data record (B4) extracts which determines the voice biometry of the respective speaker de r further radio message (M4) is provided at its output, - wherein the database (D) of the biometric unit (1) is connected downstream and at the arrival of the further biometric data record (B4) at the input of the database (D) after that biometrics Data set (B ^ which matches best with the further biometric data set (B4), the database (D) this biometric data set (B ^ together with the this biometric data set (B ^ associated identifier (Kd) available.
    [8]
    8. radio system according to claim 7, characterized by a biometric comparison unit (2) of the biometrics unit (1) created biometric data set (B ^ and of the database (D) determined further biometric data set (B4) the biometric comparison unit (2) compares the biometric data record (Bi) and the further biometric data record (B4) with one another and evaluates differences in the two biometric data sets (B1; B4), and - the biometric comparison unit (2 ) determines and maintains the probability (PB) with which the biometric data set (Bi) and the further biometric data set (B4) originate from the same speaker.
    [9]
    9. radio system according to claim 7 or 8, characterized in that the assignment unit (5) in the presence of a biometric data set (B4) and a digital identifier (Kd), already stored for a biometric data set (B ^ in the database (D) is adapted to the data stored at the database and the digital identifier (KD) associated record to the newly added radio message (M4) and biometrics record (B4).
    [10]
    10. Radio system according to one of claims 7 to 9, characterized in that - the allocation unit (5) determines the number (n ^ of the identifier (Kd) of determined voice communications (M1; M4) and / or the total duration (ti) of the communications (M1 ; M4) and assigns it to the biometric record (B ^) stored in the database (D), and - that the database (D) assigns the number (n ^ and / or total duration (t ^ to the biometric record (B) to the biometric record (B) Comparing unit (2) is transmitted, and - that the comparison unit (2), the probability (PB) of the detected agreement, in particular by statistical methods, the larger the larger the number (n ^ and / or the greater the total duration (td) of the respective biometric record (B ^ associated radio messages (M1; M2) with the digital identifier (Kd).
    [11]
    11 radio station system according to one of the preceding claims 7 to 10, characterized in that the voice recognition unit (0) upon receipt of a radio message (M ^ M2) together with the digital identifier (Kd) a probability (Ps) for the correctness of the respectively determined identifier (Kd ) and that a probability comparison unit (3) is provided which determines the probability (Ps) determined by the speech recognition unit (0) for the correctness of the respectively determined digital identifier (Kd2) and the probability (PB) determined by the biometric comparison unit (2). wherein the comparison unit (3) determines which of the two probabilities (PB, Ps) supplied to it are greater and outputs a signal (Sei) at its output, and that a selection unit (4) is provided which receives the information from the probability comparison unit (3) generated signal (Sei), the digit determined by the speech recognition unit (0) ale identifier (Kd2) as well as the digital identifier (Kd1) determined by the database (D) and which, on the basis of the signal (Sei) arriving at it, selects the digital identifier (Kd1, Kd2) and makes it available at its output each associated with larger probabilities (PB, Ps).
    [12]
    12. A data carrier on which a program for carrying out a method according to one of claims 1 to 6 is stored.
    类似技术:
    公开号 | 公开日 | 专利标题
    EP2909833B1|2016-07-13|Speech recognition in a motor vehicle
    DE102009017176A1|2009-10-29|Navigation arrangement for a motor vehicle
    EP3430615A1|2019-01-23|Transportation means, and system and method for adapting the length of a permissible speech pause in the context of a speech input
    DE102011107934A1|2013-01-24|assistance system
    DE102016100065A1|2017-07-06|System and method for detecting entrained objects for a vehicle
    AT516219A1|2016-03-15|Method for identifying and checking voice radio messages
    DE102017220266B3|2018-12-13|Method for checking an onboard speech recognizer of a motor vehicle and control device and motor vehicle
    DE102017202176A1|2018-08-16|Method for detecting a manipulation on a respective data network of at least one motor vehicle and server device
    EP3152753B1|2019-03-13|Assistance system that can be controlled by means of voice inputs, having a functional device and a plurality of voice recognition modules
    WO2015106930A1|2015-07-23|Method and system for generating a control command
    DE102015212650B4|2020-02-13|Method and system for computer-assisted processing of a speech input
    EP3335199A1|2018-06-20|Method and device for carrying out a screening of persons
    DE102016003903A1|2016-09-29|Method for speech recognition in a motor vehicle
    DE102015210117A1|2016-12-08|Driver assistance device for traffic sign recognition for a motor vehicle and a method for checking a possible misrecognition of a recognized with the driver assistance device traffic sign
    DE102020107775A1|2021-09-23|Detection and interpretation of acoustic signals and events in the vehicle exterior and / or interior
    DE102018207906A1|2019-11-21|Apparatus, system and method for automatically configuring a vehicle
    DE102018126056A1|2020-04-23|Method and computer program for transcribing a recorded voice communication
    DE102018212657A1|2020-01-30|Method and device for detecting irregularities in a computer network
    DE102017206281A1|2018-10-18|Processing a voice input
    DE102016009702A1|2017-02-02|Method and device for object recognition for a vehicle
    DE102017206876A1|2018-10-25|Method and device for outputting a status message in a motor vehicle with voice control system
    DE102012217130A1|2014-03-27|Client device e.g. mobile telephone has control device which receives quick start command for transmitting mobile report of event into security alarm network, according to environment data received at preset time
    DE102017205728A1|2018-10-04|A method for issuing an emergency call, for confirming an emergency call signal for a vehicle, methods for detecting an emergency situation, corresponding devices, computer program and machine-readable storage media
    DE102015211627A1|2017-01-12|Method and device for informing a user when approaching a destination
    DE102019001008A1|2019-08-01|Device for the collaborative input of acoustic speech signals
    同族专利:
    公开号 | 公开日
    EP3117428A1|2017-01-18|
    EP3117428B1|2018-11-28|
    ES2704948T3|2019-03-20|
    WO2016037204A1|2016-03-17|
    AT516219B1|2017-06-15|
    引用文献:
    公开号 | 申请日 | 公开日 | 申请人 | 专利标题
    DE19619015A1|1996-05-10|1997-11-13|Deutsche Forsch Luft Raumfahrt|Traffic monitoring method and arrangement|
    US20030193409A1|2001-10-10|2003-10-16|Crank Kelly C.|Method and apparatus for tracking aircraft and securing against unauthorized access|
    US20070106510A1|2005-09-29|2007-05-10|Ivras Inc.|Voice based data capturing system|
    US20080221886A1|2007-01-03|2008-09-11|Airbus France|Method and system for the entry of flight data for an aircraft, transmitted between a crew on board the aircraft and ground staff|
    US7809405B1|2007-01-31|2010-10-05|Rockwell Collins, Inc.|System and method for reducing aviation voice communication confusion|
    KR20130070345A|2011-12-19|2013-06-27|세종대학교산학협력단|Apparatus and method for recognizing of speaker using vocal signal|
    US6411933B1|1999-11-22|2002-06-25|International Business Machines Corporation|Methods and apparatus for correlating biometric attributes and biometric attribute production features|
    US8190437B2|2008-10-24|2012-05-29|Nuance Communications, Inc.|Speaker verification methods and apparatus|US10157616B2|2016-12-19|2018-12-18|Honeywell International Inc.|Methods and apparatus for post-processing speech recognition results of received radio voice messages onboard an aircraft|
    法律状态:
    优先权:
    申请号 | 申请日 | 专利标题
    ATA50621/2014A|AT516219B1|2014-09-09|2014-09-09|Method for identifying and checking voice radio messages|ATA50621/2014A| AT516219B1|2014-09-09|2014-09-09|Method for identifying and checking voice radio messages|
    ES15781848T| ES2704948T3|2014-09-09|2015-09-08|Procedure for the identification and verification of radiotelephone messages|
    PCT/AT2015/050217| WO2016037204A1|2014-09-09|2015-09-08|Method for identifying and verifying radiotelephony messages|
    EP15781848.5A| EP3117428B1|2014-09-09|2015-09-08|Method for identifying and verifying radiotelephony messages|
    [返回顶部]