JP2008067015A - Secure communication system, secure transmission device, secure reception device, and secure communication method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a secure communication system, a secure transmission device, a secure reception device, and a secure communication method which are capable of suppressing the increase of the information amount of communication. <P>SOLUTION: On the transmission side, voice information is coded to digital voice signals in accordance with a prescribed system, and preliminarily set partial voice signals out of the voice signals are reconfigurably dispersed into a plurality of pieces of information by a secret sharing scheme to generate a plurality of pieces of shared information, and respective pieces of shared information and the other voice signals are transmitted so that respective pieces of shared information pass a plurality of different communication paths. In the reception side, respective pieces of shared information and the other voice signals are received through the plurality of communication paths, and respective pieces of shared information are reconfigured by the secret sharing scheme to generate a reconfigured voice signal, and a decoded voice signal is generated on the basis of the reconfigured voice signal and the other voice signals and is converted to voice information in accordance with a prescribed system. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a secret communication system, a secret transmission device, a secret reception device, and a secret communication method.

  In general, voice communication such as VoIP using a network such as the Internet is TCP / IP communication performed using a one-to-one gateway. Conventionally, voice communication has been encrypted by IPsec, SSL, etc. to conceal communication contents. However, these encryption technologies have a problem that they are vulnerable to leakage of secret keys.

  In order to solve such a problem, a secret communication method using a secret sharing method has been proposed. That is, voice information is distributed to a plurality of pieces of shared information by the secret sharing method, and each piece of shared information is communicated via different communication paths. As a result, unless wiretapping is performed on all communication paths, the content of communication is not wiretapped, and higher secrecy can be realized compared to conventional encrypted communication (see, for example, Non-Patent Document 1). ).

  Note that the secret sharing method (threshold secret sharing method; k-out-of-n distribution method) is a kind of encryption method, and uses an arbitrarily defined constant n and k (n ≧ k) to Encryption is performed by distributing information into n pieces of distributed information. The decoding of the original information can be performed by using any k or more pieces of shared information, and no information related to the original information can be obtained from less than k pieces of shared information (for example, see Non-Patent Document 2). ).

Fujita et al., "Proposal of an acoustic secret sharing method for 1-bit audio", Sound theory, 2004, Vol.1, pp.579-580. Adi Shamir, “How to share a secret”, Communications of the ACM, 1979, Vol. 22, No. 11, pp.612-613

In the secret communication method described in Non-Patent Document 1, the amount of information is approximately n · 2 ⁽ⁿ⁻¹⁾ times when the information is distributed to n pieces of distributed information before and after information distribution by the secret sharing method. Therefore, there is a problem that the communication band is consumed greatly.

  The present invention has been made paying attention to such problems, and an object thereof is to provide a secret communication system, a secret transmission device, a secret reception device, and a secret communication method capable of suppressing an increase in the amount of information to be communicated. It is said.

  In order to achieve the above object, the secret communication system according to the present invention includes an encoding unit that encodes audio information into a digital audio signal according to a predetermined method, and an audio signal encoded by the encoding unit. Among them, a dispersion generating unit that generates a plurality of pieces of shared information by distributing a plurality of preset audio signals to a plurality of reconfigurable by a secret sharing method, and the plurality of pieces of shared information generated by the distribution generating unit And transmitting means for transmitting the remaining audio signal among the audio signals encoded by the encoding means, and receiving the plurality of shared information and the remaining audio signal transmitted by the transmitting means. Receiving means, reconstructing means for reconstructing the plurality of shared information received by the receiving means by the secret sharing method to generate a reconstructed audio signal, and the reconfiguration generated by the reconfiguring means A decoding means for generating a decoded voice signal based on a voice signal and the remaining voice signal received by the receiving means; and the decoded voice signal generated by the decoding means is converted into voice information according to the predetermined method. Voice conversion means for converting.

  Since the secret communication system according to the present invention encodes voice information into a digital voice signal according to a predetermined method by the encoding means, the amount of information can be reduced as compared with the case where the voice information is used as it is. In addition, since the distributed generation means generates a plurality of pieces of distributed information by distributing a plurality of preset audio signals among the encoded audio signals so as to be reconfigurable by the secret sharing method. Compared with the case where a plurality of pieces of shared information are generated from the entirety, the amount of information that increases can be reduced. Thereby, the secret communication system which concerns on this invention can suppress the increase in the information content which communicates.

  Since a part of the audio signal is distributed to a plurality of shared information by the secret sharing method and communication is performed by the transmitting means and the receiving means, unless all the shared information is acquired, the communication contents are not wiretapped, and it is confidential High nature. Since the information to be kept secret is not the audio signal but the conversation content, it is not necessary to distribute the entire audio signal, and the conversation content can be kept secret by distributing a part of the audio signal.

  In the secret communication system according to the present invention, the transmission unit can transmit the plurality of shared information through a plurality of different communication paths, and the reception unit transmits the plurality of shared information transmitted by the transmission unit. Can be received via the plurality of communication paths. In this case, since a plurality of pieces of shared information can be communicated through different communication paths, communication contents are not wiretapped unless wiretapping is performed on all communication paths, and higher confidentiality is achieved. Can be realized.

  In the secret communication system according to the present invention, it is preferable that the encoding unit encodes audio information into a digital audio signal according to CELP, and the audio conversion unit converts the decoded audio signal into audio information according to the CELP. In particular, in the secret communication system according to the present invention, the encoding means encodes speech information into a digital speech signal according to LD-CELP, and the speech conversion means converts the decoded speech signal into speech information according to the LD-CELP. It is preferable to do. Further, in the secret communication system according to the present invention, the dispersion generating means is a part of a vector code set in advance for each cycle of the speech signal encoded with 10 bits per cycle according to the LD-CELP. It is preferable to generate a plurality of pieces of shared information by distributing a plurality of pieces of information in a reconfigurable manner by a secret sharing method. In particular, in the secret communication system according to the present invention, the variance generating means includes a vector code set in advance for each cycle of the speech signal encoded with 10 bits per cycle according to the LD-CELP. It is preferable to generate a plurality of pieces of shared information by distributing three bits into a plurality of reconfigurable pieces by a secret sharing method.

CELP (Code Excited Linear Prediction) is one of speech coding techniques and uses speech basis vector quantization and linear prediction. LD-CELP (Low-Delay Code Excited Linear Prediction: ITU-T G.728) has a relatively simple structure among CELP. The LD-CELP encoder outputs a 10-bit code for every 5 samples of the PCM signal, and the 10-bit information includes a 3-bit gain code and a 7-bit vector code. The encoder and the decoder have a common codebook having 8 (= 2 ³ ) types of gain information corresponding to them and 128 (= 2 ⁷ ) types of vector information.

  When encoding is performed according to LD-CELP, the efficiency of secrecy is higher by dispersing part of the vector code by the secret sharing method than when part of the gain code is distributed. In particular, by distributing 3 bits of the vector code, communication can be performed with a smaller amount of information while keeping the secret.

  The secret transmission device according to the present invention includes an encoding unit that encodes audio information into a digital audio signal in accordance with a predetermined method, and a preset part of the audio signal encoded by the encoding unit. And a plurality of communication paths in which the plurality of pieces of distributed information generated by the distribution generation unit are different from each other. And transmitting means for transmitting the plurality of pieces of shared information and the remaining audio signal among the audio signals encoded by the encoding means.

  The secret message receiving apparatus according to the present invention includes a coding unit that codes voice information into a digital voice signal according to a predetermined method, and a preset voice signal encoded by the coding unit. A plurality of pieces of distributed information generated by distributing a part of audio signals in a reconfigurable manner by a secret sharing method to generate a plurality of pieces of shared information; A secret message receiving apparatus used together with a secret message transmission apparatus having a transmission means for transmitting the plurality of pieces of shared information and the remaining voice signal among the voice signals encoded by the encoding means so as to pass through a communication path A receiving unit configured to receive the plurality of shared information and the remaining audio signal transmitted by the secret transmission device via the plurality of communication paths; and Reconstructing means for reconstructing the received shared information by the secret sharing method to generate a reconstructed audio signal, the reconstructed audio signal generated by the reconfiguring means, and receiving by the receiving means Decoding means for generating a decoded audio signal based on the remaining audio signal, and audio conversion means for converting the decoded audio signal generated by the decoding means into audio information according to the predetermined method. It is characterized by that.

  The secret communication method according to the present invention includes an encoding step for encoding audio information into a digital audio signal according to a predetermined method, and a preset part of the audio signal encoded in the encoding step. And a plurality of communication paths in which the plurality of pieces of shared information generated in the distributed generation step are different from each other. A transmission step for transmitting the plurality of shared information and the remaining audio signal among the audio signals encoded in the encoding step; and the plurality of shared information transmitted in the transmission step; A reception step of receiving the remaining audio signal via the plurality of communication paths; and the plurality of pieces of shared information received in the reception step. Based on a reconstruction step of reconstructing by a distributed method to generate a reconstructed speech signal, the reconstructed speech signal generated in the reconstruction step, and the remaining speech signal received in the reception step A decoding step for generating a decoded audio signal; and an audio conversion step for converting the decoded audio signal generated in the decoding step into audio information in accordance with the predetermined method.

  Since the secret transmission device, the secret reception device, and the secret communication method according to the present invention encode the voice information into a digital voice signal according to a predetermined method, the amount of information is reduced as compared with the case where the voice information is used as it is. be able to. In addition, among the encoded audio signals, a part of the preset audio signals are dispersed into a plurality of reconfigurable information by the secret sharing method to generate a plurality of pieces of distributed information. Compared with the case where information is generated, the amount of information to be increased can be reduced. Thereby, the increase in the amount of information to communicate can be suppressed.

  A part of the audio signal is distributed to multiple pieces of shared information by the secret sharing method, and each piece of shared information is communicated through different communication paths. Is not eavesdropped, and high confidentiality can be realized. Since the information to be kept secret is not the audio signal but the conversation content, it is not necessary to distribute the entire audio signal, and the conversation content can be kept secret by distributing a part of the audio signal.

  According to the present invention, it is possible to provide a secret communication system, a secret transmission device, a secret reception device, and a secret communication method that can suppress an increase in the amount of information to be communicated.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.
1 to 4 show a secret communication system, a secret transmission device, a secret reception device, and a secret communication method according to an embodiment of the present invention. The secret communication method according to the embodiment of the present invention is a method executed by the secret communication system, the secret transmission device, and the secret reception device according to the embodiment of the present invention.
As shown in FIG. 1 and FIG. 2, the secret communication system includes a secret transmission device 1 and a secret reception device 2.

  As shown in FIG. 1, the secret story transmission device 1 includes an encoding unit 11, a variance generation unit 12, and a transmission unit 13. The encoding unit 11 includes a CELP encoder 14 and a code buffer 15. The CELP encoder 14 receives the audio information converted into the PCM signal, and encodes the audio information into a digital audio signal at 10 bits per cycle for every 5 samples of the PCM signal according to the LD-CELP. . The code buffer 15 is connected to the CELP encoder 14, acquires the audio signal encoded by the CELP encoder 14, and temporarily stores it.

  The distribution generation means 12 includes a bit splitter 16, a secret sharing module 17, and a split module 18. The bit splitter 16 is connected to the code buffer 15, the secret sharing module 17, and the dividing module 18. The bit splitter 16 acquires the audio signal stored in the code buffer 15 and divides it into 3 bits of a vector code set in advance every 10 bits per cycle and the remaining 7 bits. Yes. Further, the bit splitter 16 transmits 3 bits of the divided vector code to the secret sharing module 17 and transmits the remaining 7 bits to the dividing module 18.

  The secret sharing module 17 acquires 3 bits of the vector code transmitted from the bit splitter 16 and distributes it to n pieces so as to be reconfigurable by the secret sharing method to generate n pieces of shared information. The dividing module 18 obtains the remaining 7 bits transmitted from the bit splitter 16 and converts them into n non-secret information according to the band in each communication path based on the information (Load balancing Information) from the load balancer. It comes to divide.

  The transmission means 13 includes a packet constructor 19 and a transmission-side multi-homing router 20. The packet constructor 19 is connected to the secret sharing module 17 and the dividing module 18. The packet constructor 19 obtains n pieces of shared information generated by the secret sharing module 17 and n pieces of non-secret information generated by the division module 18 and combines them to generate n packets. It is supposed to be. The transmission side multihoming router 20 is connected to the packet constructor 19 and the Internet. The transmission-side multihoming router 20 acquires n packets generated by the packet constructor 19 and transmits them to the Internet so that each packet passes through a plurality of different communication paths.

  As shown in FIG. 2, the secret message receiving apparatus 2 includes a receiving unit 21, a reconstruction unit 22, a decoding unit 23, and a voice conversion unit 24. The receiving means 21 includes a receiving side multi-homing router 25. The reception side multihoming router 25 is connected to the Internet and receives n packets transmitted by the transmission side multihoming router 20 via a plurality of communication paths.

  The reconfiguration unit 22 includes a data splitter 26 and a reconfiguration module 27. The data splitter 26 is connected to the receiving side multihoming router 25, the reconfiguration module 27, and the decoding unit 23. The data splitter 26 acquires n packets received by the receiving multihoming router 25, divides it into n pieces of shared information and n pieces of non-secret information, and transmits each piece of shared information to the reconfiguration module 27. Each non-secret information is transmitted to the decryption means 23. The reconfiguration module 27 acquires each piece of distributed information transmitted from the data splitter 26 and reconfigures it by the secret sharing method to generate a reconfigured audio signal.

  The decoding means 23 has a reconstructor module 28. The reconstructor module 28 is connected to the reconstruction module 27 and the data splitter 26. The reconstructor module 28 acquires the reconstructed audio signal generated by the reconfiguration module 27 and each non-secret information transmitted from the data splitter 26, and converts the reconfigured audio signal into a vector code as set in advance. The non-secret information is converted to the remaining 7 bits and combined to generate a decoded speech signal.

  The voice conversion means 24 has a CELP decoder 29. The CELP decoder 29 is connected to the reconstructor module 28. The CELP decoder 29 converts the decoded audio signal generated by the reconstructor module 28 into audio information including a PCM signal in accordance with LD-CELP, and outputs the audio information as a PCM signal.

Next, the operation will be described.
In the secret communication system according to the embodiment of the present invention, first, an audio signal converted into a PCM signal is input to the CELP encoder 14 by a secret transmission device. The CELP encoder 14 encodes the audio information into a digital audio signal according to LD-CELP, and stores the audio signal in the code buffer 15. At this time, since the audio information is encoded into a digital audio signal according to LD-CELP, the amount of information can be reduced as compared with the case where the audio information is used as it is.

  The bit splitter 16 divides the audio signal into 3 bits of a vector code set in advance every 10 bits per cycle and the remaining 7 bits. The secret sharing module 17 generates 3 pieces of distributed information by distributing 3 bits of the vector code into n pieces that can be reconfigured by the secret sharing method. At this time, since the n pieces of shared information are generated from a part of the preset audio signals among the encoded audio signals, the number of pieces of distributed information is increased as compared with the case where n pieces of shared information are generated from the entire audio signals. The amount of information to be reduced can be reduced. Thereby, the increase in the amount of information to communicate can be suppressed.

  The dividing module 18 divides the remaining 7 bits into n pieces of non-secret information. The packet constructor 19 combines n pieces of distributed information and n pieces of non-secret information to generate n packets. The transmitting multihoming router 20 transmits the n packets to the Internet so that each packet passes through a plurality of different communication paths.

  In the secret talk receiver 2, the receiving multihoming router 25 receives the transmitted n packets via the plurality of communication paths on the Internet. The data splitter 26 divides the n packets into n pieces of distributed information and n pieces of non-secret information. The reconstruction module 27 reconstructs each shared information by the secret sharing method to generate a reconstructed audio signal. As previously set, the reconstructor module 28 converts the reconstructed speech signal into 3 bits of a vector code, converts each non-secret information into the remaining 7 bits and combines them to generate a decoded speech signal. CELP decoder 29 converts the decoded audio signal into audio information composed of a PCM signal in accordance with LD-CELP, and outputs the audio information as a PCM signal.

  In this way, a part of the audio signal is distributed into n pieces of shared information by the secret sharing method, and each shared information is communicated through a different communication path, so that wiretapping is performed on all communication paths. As long as there is no communication content, eavesdropping is not performed and high secrecy can be realized. Since the information to be kept secret is not the audio signal but the conversation content, it is not necessary to distribute the entire audio signal, and the conversation content can be kept secret by distributing a part of the audio signal.

[Evaluation experiment of secrecy]
The relationship between the amount of secret sharing and the secrecy of the contents was evaluated by sentence comprehension by four listeners whose mother tongue was Japanese. As shown in FIG. 3, as an encryption bit (Random bit) to be distributed by the secret sharing method, a gain code (MSB (Most Significant Bit)) is used for every 10 bits of a voice signal encoded according to LD-CELP. Select sg (= 0,1,2,3) bits for Gain index) and sv (= 0,1,2,3,4,5,6,7) bits for vector code (Vector index) To do. Thus, the experiment was performed under the experimental conditions of 32 (= 4 × 8) patterns in total.

  As input speech, 320 types of female speaker's speech (sampling frequency 16kHz, quantization bit 16bit) of "phoneme balance 1000 sentence" (manufactured by NTT-AT) composed of about 5 to 8 words. used. The experiment was performed in a soundproof room, and the sound was presented to both ears using headphones with a sound presentation level of 60 dBA. Each listener was allowed to listen to the presentation sound multiple times, and the contents of the presentation sound were entered on a sheet. The number of trials was 320 times per person, and 320 types of presentation sounds were all created from different original sounds.

  For each answer of each sentence, the correct answer rate was obtained for independent words (nouns, verbs, adjectives, adverbs), and the degree of sentence comprehension was obtained. For each of the 32 patterns of experimental conditions, the sentence comprehension for 40 trials by each listener and each original speech was added and averaged, and the results are shown in Table 1. Further, the sentence intelligibility when only the gain code (Gain code) is an encrypted bit (sv = 0), and the sentence intelligibility when only the vector code (Vector code) is an encrypted bit (sg = 0), And the standard deviation in each 40 trials was calculated | required and the result is shown in FIG.

  Based on the results of FIG. 4, analysis of variance of two factors (2 × 4 levels) was performed. As a result, the interaction was significant. In addition, as a result of the simple main effect test for each code, when the number of encrypted bits is 1, 2, and 3, there is a significant difference (p <.01) between the vector code and the gain code, respectively. Admitted.

  From this experimental result, in order to communicate with a smaller amount of information without reducing confidentiality as much as possible, it is more effective to distribute part of the vector code by the secret sharing method than to the gain code. confirmed. In particular, by distributing 3 bits of the vector code, communication can be performed with a smaller amount of information while maintaining high confidentiality.

It is a block diagram which shows the secret transmission apparatus of the secret communication system of embodiment of this invention. It is a block diagram which shows the secret message receiver of the secret message communication system of embodiment of this invention. It is a conceptual diagram which shows an example (sg = 1, sv = 3) of the encryption pattern of the audio | voice signal of the secret communication system of embodiment of this invention. It is a graph which shows the relationship between the encryption bit number of the audio | voice signal (sg = 0, sv = 0) and sentence intelligibility of the secret communication system of embodiment of this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Secret transmission apparatus 2 Secret reception apparatus 11 Encoding means 12 Distribution | distribution production | generation means 13 Transmission means 14 CELP encoder 15 Code buffer 16 Bit splitter 17 Secret distribution module 18 Dividing module 19 Packet constructor 20 Transmission side multihoming router 21 Reception means 22 Reconfiguration means 23 Decoding means 24 Voice conversion means 25 Receiving side multihoming router 26 Data splitter 27 Reconfiguration module 28 Reconstructor module 29 CELP decoder

Claims (9)

Encoding means for encoding audio information into a digital audio signal according to a predetermined method;
Dispersion generating means for generating a plurality of pieces of distributed information by distributing a plurality of predetermined voice signals among the voice signals encoded by the encoding means to be reconfigurable by a secret sharing method;
Transmitting means for transmitting the plurality of shared information generated by the dispersion generating means and the remaining audio signal among the audio signals encoded by the encoding means;
Receiving means for receiving the plurality of shared information and the remaining audio signal transmitted by the transmitting means;
Reconstructing means for reconstructing the plurality of shared information received by the receiving means by the secret sharing method to generate a reconstructed audio signal;
Decoding means for generating a decoded speech signal based on the reconstructed speech signal generated by the reconstruction means and the remaining speech signal received by the receiving means;
Audio conversion means for converting the decoded audio signal generated by the decoding means into audio information according to the predetermined method;
A secret communication system characterized by comprising: The transmission means can transmit the plurality of shared information through a plurality of different communication paths, respectively.
The receiving means is capable of receiving the plurality of shared information transmitted by the transmitting means via the plurality of communication paths;
The secret communication system according to claim 1. The encoding means encodes audio information into a digital audio signal according to CELP,
The voice converting means converts the decoded voice signal into voice information according to the CELP;
The secret communication system according to claim 1 or 2, characterized in that: The encoding means encodes audio information into a digital audio signal according to LD-CELP,
The voice converting means converts the decoded voice signal into voice information according to the LD-CELP;
The secret communication system according to claim 1 or 2, characterized in that:   The variance generating means includes a plurality of vector codes preset for each cycle of the speech signal encoded with 10 bits per cycle according to the LD-CELP so as to be reconfigurable by a secret sharing method. The secret communication system according to claim 4, wherein a plurality of pieces of shared information are generated by being distributed.   The variance generating means can reconstruct 3 bits of the vector code preset for each cycle of the speech signal encoded with 10 bits per cycle according to the LD-CELP by a secret sharing method. The secret communication system according to claim 4, wherein a plurality of pieces of shared information are generated by being distributed in a plurality. Encoding means for encoding audio information into a digital audio signal according to a predetermined method;
Dispersion generating means for generating a plurality of pieces of distributed information by distributing a plurality of predetermined voice signals among the voice signals encoded by the encoding means to be reconfigurable by a secret sharing method;
The plurality of shared information and the remaining audio signal among the audio signals encoded by the encoding unit so that the plurality of shared information generated by the distribution generating unit pass through a plurality of different communication paths. Transmitting means for transmitting
A secret story transmitting device comprising: Encoding means for encoding audio information into a digital audio signal according to a predetermined method, and a part of the preset audio signal among the audio signals encoded by the encoding means is re-recorded by a secret sharing method. A plurality of pieces of distributed information that are configured to be distributed in a plurality of configurable ways to generate a plurality of pieces of shared information and the plurality of pieces of shared information generated by the pieces of distributed generation means pass through a plurality of different communication paths. And a secret message receiving device used together with a secret message transmitting device having a transmitting means for transmitting the remaining voice signal among the voice signals encoded by the encoding means,
Receiving means for receiving the plurality of shared information and the remaining audio signal transmitted by the secret transmission device via the plurality of communication paths;
Reconstructing means for reconstructing the plurality of shared information received by the receiving means by the secret sharing method to generate a reconstructed audio signal;
Decoding means for generating a decoded speech signal based on the reconstructed speech signal generated by the reconstruction means and the remaining speech signal received by the receiving means;
Audio conversion means for converting the decoded audio signal generated by the decoding means into audio information according to the predetermined method;
An apparatus for receiving a secret story, comprising: An encoding step of encoding audio information into a digital audio signal according to a predetermined scheme;
A distributed generation step of generating a plurality of pieces of distributed information by distributing a plurality of preset voice signals among the voice signals encoded in the encoding step so as to be reconfigurable by a secret sharing method;
The plurality of shared information and the remaining audio signal among the audio signals encoded in the encoding step so that the plurality of shared information generated in the distributed generation step pass through a plurality of different communication paths. Sending step for sending and
A reception step of receiving the plurality of shared information and the remaining audio signal transmitted in the transmission step via the plurality of communication paths;
A reconstructing step of reconstructing the plurality of shared information received in the receiving step by the secret sharing method to generate a reconstructed audio signal;
A decoding step for generating a decoded audio signal based on the reconstructed audio signal generated in the reconfiguration step and the remaining audio signal received in the reception step;
A voice conversion step of converting the decoded voice signal generated in the decoding step into voice information according to the predetermined method;
A secret communication method characterized by comprising:

Images