Timing synchronization method for end-to-end voice encryption in mobile communication network

The proposed timing synchronization method for end-to-end voice encryption in mobile networks uses dual training sequences to enhance synchronization accuracy and reduce complexity by employing linear frequency modulation and OFDM signals, addressing vocoder penetration and device diversity issues.

CN119071895APending Publication Date: 2024-12-03HOHAI UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411259218.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-09-09
Publication Date
2024-12-03

AI Technical Summary

Technical Problem

Existing synchronization algorithms in end-to-end voice encryption systems for mobile communication networks face challenges such as inability to penetrate vocoders, poor synchronization accuracy, and high computational complexity, particularly due to non-integer resampling and diverse mobile devices causing synchronization errors.

Method used

A timing synchronization method involving two training sequence frames, A and B, is employed. Frame A uses a linear frequency modulation signal for initial synchronization, and frame B, an OFDM-modulated signal, for precise synchronization, utilizing cross-correlation and frame energy calculations to determine synchronization points effectively.

Benefits of technology

The method achieves high synchronization accuracy with reduced computational complexity by using optimized training sequences that penetrate vocoders and adapt to diverse mobile devices, ensuring robust and efficient encryption.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119071895A_ABST
    Figure CN119071895A_ABST
Patent Text Reader

Abstract

The invention discloses a timing synchronization method for end-to-end voice encryption in a mobile communication network. The timing synchronization method mainly comprises the following two steps. The method comprises the steps of 1, performing primary synchronization by using a training sequence frame A to obtain a primary timing detection range; and step 2, in a timing detection range of initial synchronization, performing accurate synchronization by using the training sequence frame B to obtain an accurate initial position of the encrypted signal. The timing synchronization scheme is provided for mobile communication network call encryption, the calculation complexity is low, the accuracy is high, and good robustness is achieved under the real-time call condition.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention belongs to the field of communication technology, and in particular relates to a timing synchronization method for end-to-end voice encryption in a mobile communication network. Background Art

[0002] With the popularization of mobile communication services, mobile terminals such as mobile phones have penetrated into every aspect of people's lives. Communication data is generated all the time, but people's communication security is not guaranteed. In the early days, most encryption algorithms were designed with operators as the main body. Considering the cost of additional channel load overhead, operators will not design too complex encryption algorithms. In addition, the user's data encryption end is in the hands of operators, which still poses security risks. In order to solve this problem, end-to-end voice encryption technology has been continuously proposed in recent years, making it possible to realize encrypted calls. End-to-end voice encryption is encrypted at the user's sending end, and only the modulated voice-like signal passes through the communication system. After receiving the modulated signal, it is decrypted at the user end. Voice signals are transmitted and processed in frames. End-to-end voice encryption communication mainly faces two technical problems: encryption and synchronization. There are relatively few studies on the solution of synchronization problems at home and abroad.

[0003] In communication systems, the autocorrelation method is generally used to determine the synchronization of frame signals. By comparing the correlation coefficient between the two signals, if the calculated value is greater than the threshold, it can be determined to be synchronized. However, because the signal at the transmitting end must pass through the vocoder in the voice channel, the synchronization algorithms in general communication systems, such as the Schmidl & Cox algorithm, the Minn algorithm, and the Park algorithm, are not suitable for end-to-end voice encryption systems. Teng Guangchao, Lang Jianjun, and others proposed a key synchronization method based on dual-tone multi-frequency (Teng Guangchao et al., Design and Implementation of a Key Synchronization Scheme in a Voice Encryptor, 2014), using two sinusoidal signals of different frequencies to represent digital signals. Although this scheme can achieve preliminary synchronization, if it is used for a long time, the utilization rate of the channel will be greatly reduced; at the same time, if it only relies on the correlation of single-frequency sinusoidal signals, the accuracy of synchronization cannot be guaranteed. Liu Guangjian and Zhang Yajun also proposed a similar method (Liu Guangjian et al., A Method for Realizing Speech Signal Synchronization, 2007). However, this algorithm requires a large number of discrete Fourier transforms, and the encryption system frequently performs synchronization operations to ensure the correctness of decryption. Therefore, its computational complexity is very large and cannot meet the needs of practical applications.

[0004] Moreover, due to the complex composition of the mobile call system and the wide variety of mobile devices involved, such as mobile phones, Bluetooth headsets, tablets, etc., the voice signal will be resampled multiple times during the transmission process, especially the non-integer resampling, which will cause several points of error in the synchronization sequence when encoding. Therefore, the synchronization point found by only one correlation calculation plus threshold judgment is not reliable, and secondary precise synchronization is required.

[0005] In summary, the current synchronization algorithm in the end-to-end voice encryption system designed for mobile communication networks has the problems of being unable to penetrate the vocoder, poor synchronization accuracy, and large amount of calculation. Summary of the invention

[0006] The invention provides a timing synchronization method for end-to-end voice encryption in a mobile communication network, which is used to solve the problems that the existing synchronization algorithm cannot penetrate a vocoder, has poor synchronization accuracy and large calculation amount.

[0007] In order to solve the above technical problems, the present invention provides a timing synchronization method for end-to-end voice encryption in a mobile communication network, comprising:

[0008] Step 1: Use training sequence frame A for initial synchronization to determine the initial timing detection range;

[0009] Step 2: Within the timing detection range of the initial synchronization, use the training sequence frame B for accurate synchronization to obtain the precise starting position of the encrypted signal.

[0010] Compared with the prior art, the present invention has the following beneficial effects:

[0011] 1. The present invention is a synchronization scheme proposed for an end-to-end voice encryption system in a mobile communication network; the synchronization accuracy is high, the computational complexity is relatively low, and the robustness is good.

[0012] 2. The present invention reasonably utilizes the inter-frame structure of the encryption system, embeds two training sequence frames A and B into the encrypted signal, and performs preliminary and precise detection twice at the receiving end to complete the timing synchronization of the encrypted signal, thereby improving the accuracy of synchronization.

[0013] 3. The present invention uses the sliding frame signal of the receiving end buffer and the training sequence frame A to perform a series of operations such as cross-correlation and frame energy, and determines the initial synchronization range through the result of the threshold judgment. Then, the training sequence frame B is used to demodulate and judge the error of the signal within the initial synchronization range according to the sliding window frame, and finally determine the timing synchronization point. The synchronization performance is better and the calculation complexity is lower.

[0014] 4. The training sequence used in the present invention is designed in combination with the optimal carrier frequency band of the vocoder and encryption system in the mobile communication system. The training sequence frame A used for the initial synchronization is a linear frequency modulation signal (LFM) with minimal distortion after the communication equipment test, and the training sequence B is a fixed data sequence or the last data frame of the previous communication modulated by orthogonal frequency division multiplexing (OFDM) technology. Both training sequences can penetrate the vocoder in the mobile communication system and have high synchronization accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 This is a flowchart of end-to-end voice encryption

[0016] Figure 2 Schematic diagram of the frame structure of the present invention

[0017] Figure 3 It is a structural flow chart of the present invention

[0018] Figure 4 It is a specific step flow chart of the present invention

[0019] Figure 5 This is the waveform of the original linear frequency modulation signal and the signal after passing through the actual channel.

[0020] Figure 6 This is the calculation result of the peak value in the peak judgment formula DETAILED DESCRIPTION

[0021] In order to make the technical solution, purpose and effect of the present invention more clearly presented, the technical solution will be described in detail below in conjunction with the embodiments of the present invention and the accompanying drawings.

[0022] The encryption method used in this case is a voice-like encryption method based on OFDM, such as Figure 1 As shown. The sending terminal can be divided into four main steps: First, the voice signal is obtained at the receiving end of the system, and the analog-to-digital conversion is performed to obtain the voice parameter information. Then, the voice parameters are compressed through the voice coding algorithm, and the corresponding coding parameters are generated. Secondly, the best carrier frequency band is verified in the experiment, and the coding parameters are modulated into the subcarrier of the carrier frequency band through OFDM, and the modulated signal is generated to realize the encryption of the voice. The synchronization sequence is embedded in the encrypted signal at equal intervals after the encryption is completed.

[0023] The receiving terminal uses the embedded synchronization sequence to synchronize the signal and demodulates after synchronization. The OFDM demodulation process is the inverse process of the OFDM modulation process. The received voice signal is subjected to a fast Fourier transform (FFT), and then bit-mapped according to the modulation method of the transmitter to obtain a parallel bit stream, which is then converted to serial to restore the voice coding parameters.

[0024] Figure 2 This is a diagram of the physical signal frame structure. The frame structure contains information about training sequence frames A and B. After the voice signal is encrypted by the encryption system, the two training sequence frames A and B are encrypted as follows: Figure 2 The structure shown is embedded in the encrypted signal. At the sending end, the data frame is composed in the following order: protection interval PD1, training sequence A, protection interval PD2, training sequence B, encrypted data cyclic prefix CP Data , encrypted data Data, protection interval PD3. The training sequence frame A is a linear frequency modulation signal with minimum distortion selected through the call channel test, with a length of 80 data points; the training sequence frame B is an encrypted signal generated by a fixed data frame or the last data frame of the previous communication after passing through the encryption system, with a length of 360 data points, the length of the protection interval PD1 is 320 data points, the length of the protection interval PD2 is 360 data points, and the encrypted data cyclic prefix CP Data The length of is 40 data points, the length of the encrypted data Data is 320 data points, and the length of the protection interval PD3 is 320 data points.

[0025] Figure 3 The synchronization method provided by the present invention mainly comprises the following steps:

[0026] Step 1: Use training sequence frame A for initial synchronization to determine the initial timing detection range;

[0027] Step 2: Within the timing detection range of the initial synchronization, use the training sequence frame B for accurate synchronization to obtain the precise starting position of the encrypted signal.

[0028] The specific steps are:

[0029] Training sequence frame setting preparation: The transmitting and receiving ends pre-agreed on the training sequence frame A, which is a fixed data sequence, and the length in this embodiment is 80 points; the transmitting and receiving ends pre-agreed on the training sequence frame B, which can be an orthogonal frequency division multiplexing (OFDM) frame generated by modulation corresponding to a fixed data sequence, or an OFDM frame generated by modulation corresponding to the last data frame of the previous communication, and the length in this embodiment is 360 points;

[0030] In the step 1: in the initial synchronization, the transmitting end sends the training sequence frame A, and a first-in-first-out buffer is first set at the receiving end to store the received signal, and the received signal is slidably intercepted with 80 points as a frame length unit as a signal sliding frame, and the energy of each sliding frame signal and the sliding cross-correlation value with the local training sequence A are calculated, and finally a threshold judgment is made to determine the preliminary timing detection range.

[0031] In step 2: after the transmitting end sends the training sequence frame A, it continues to send the training sequence frame B, and the receiving end buffer receives the signal; taking the timing detection range of the initial synchronization as the starting time, taking 360 points as a frame length unit, slidingly intercepting the received signal as a signal sliding frame, performing OFDM demodulation on each signal sliding frame, calculating the residual square sum of the demodulated data and the original data, finding the minimum error point to complete accurate synchronization, and starting from this point plus one as the correct starting point, the data after the starting point is the correct encrypted data, that is, Figure 2 Cyclic prefix CP in the encrypted data frame Data And the encrypted data frame Data.

[0032] The specific process of this algorithm is described in detail below.

[0033] 1. First, the initial synchronization in step 1 is explained.

[0034] In conventional communication systems, an autocorrelation operation is generally used at the receiving end to detect the training sequence to determine the initial synchronization range. However, in the mobile communication network system, there is a vocoder in the channel, and ordinary training sequences are difficult to penetrate the vocoder, which will produce large distortion. Therefore, it is impossible to directly use the autocorrelation operation at the receiving end. The present invention uses a linear frequency modulation signal that can penetrate the vocoder as the training sequence A. The signal generation expression is shown in (1), and a series of operations such as cross-correlation are used at the receiving end to determine the initial data frame range.

[0035]

[0036] In formula (1), LFM(n) represents the value of the linear frequency modulation signal at point n, f0 represents the initial frequency, T represents the sampling period, and β represents the modulation frequency.

[0037] Specific steps are as follows Figure 4 As shown, the length of the receiving end buffer is set to 2M+K. In this embodiment, M=360, K=10. The length K=10 is to ensure that the buffer can receive data of 2 frames in length and have 10 redundant data bits to ensure robust reception. Then, the local training sequence frame A and the signal in the receiving end buffer are cross-correlated, so that the range of the synchronization position can be preliminarily determined, reducing the amount of calculation for precise synchronization.

[0038] It should be noted that although the linear frequency modulation signal can penetrate the vocoder, due to the diversity of mobile devices, the waveform change degree of the linear frequency modulation signal received by the receiver during the call is different after encoding and decoding. In order to improve the accuracy of synchronization, the training sequence frame A used for cross-correlation calculation at the receiver uses the average value of the linear frequency modulation signal after testing on different devices. The original signal and the final average signal are shown in Figure 5 shown.

[0039] according to Figure 2 The frame structure shown in FIG. 1 shows, in this embodiment, the length of the training sequence frame A is 80 points. The cross-correlation operation shown in expression (2) is performed using the local training sequence frame A of the receiving end and the signal in the receiving end buffer to obtain the cross-correlation value C(x). In expression (2), C(x) represents the cross-correlation value between the sliding frame x with a starting time of time t and a data frame length of N and the local training sequence A, represents the signal value of the sliding frame x at time t+k-1, and A(k) represents the signal value of the local training sequence A at time k;

[0040]

[0041] Using only the cross-correlation value to determine the initial synchronization range will result in a large synchronization error, which will increase the amount of calculation required for subsequent accurate synchronization determination. Therefore, it is necessary to add a constraint condition to reduce the error range of the initial synchronization. While calculating the cross-correlation value of the sliding frame x, we further calculate the frame energy of this frame, as shown in expression (3), where E(x) represents the frame energy of the sliding frame x.

[0042]

[0043] In this embodiment, for a receiving buffer with a length of 2M+K=730, sliding frames with a frame length of N=80 are sequentially intercepted from the time [K+1, K+2M-N], that is, from the time 11 to the time 650. After calculating the cross-correlation value C(x) and the frame energy E(x) of these sliding frames, the synchronization event judgment as shown in expression (4) is performed in sequence. The calculation results are as follows: Figure 6 As shown. In formula (4), E s represents the frame energy of the local training sequence frame A, T s represents a preset threshold coefficient, which is 0.875 in this embodiment. R(x) is the ratio of the cross-correlation value defined in equation (2) to the frame energy defined in equation (3).

[0044]

[0045] If R(x) corresponding to all sliding frames is less than This indicates that no synchronization event occurs in the receive buffer; if one or more sliding frames correspond to R(x) greater than or equal to This indicates that a synchronization event has occurred in the receiving buffer. The initial timing detection range is set to the precise synchronization service. At this time, the sliding frame corresponding to the maximum value of R(x) is taken and recorded as x. max, record the starting time of the frame, recorded as t0, and use it as the center of the search time to search the cross-correlation value forward and backward, and obtain a local maximum point before and after, which are recorded as R(x1) and R(x2) respectively, and the corresponding times are recorded as t1 and t2. Finally, the time range of [x1, x2] is set in the receiving buffer as the preliminary timing detection range.

[0046] 2. Next, the precise synchronization in step 2 is explained.

[0047] The training sequence frame used for precise synchronization is an OFDM modulated signal. The specific generation method in this embodiment is: after testing the best carrier frequency band of the real channel, a group of data agreed in advance by the transceiver is used to place it on the subcarrier corresponding to the best test frequency band through OFDM modulation, and finally a training sequence frame B is generated, with a frame length of M=360.

[0048] The receiving end receives the training sequence frame B in the receiving buffer, takes the time range of [t1, t2] as the starting point of the received signal, and intercepts the sliding frame with M = 360 as the frame length. Then, all sliding frames are demodulated in turn to obtain the demodulated data, and then the difference between the demodulated data and the OFDM pre-modulation data of the pre-agreed training sequence frame B is calculated, and the difference is measured according to the residual square sum of formula (5). Among them, SSE is the residual square sum, m represents the length of the data sequence before demodulation of the training sequence frame B, y i Indicates the data sequence value before demodulation of training sequence frame B. Indicates the data sequence value from sliding frame demodulation:

[0049]

[0050] Among all the sliding frames, find the sliding frame corresponding to the minimum residual square sum value, and determine the precise synchronization point from the starting point of the frame to complete the precise synchronization process. Adding one from this point is the starting point of the decryption FFT window, after which the decryption operation of the encrypted signal can be performed.

[0051] It should be pointed out that the described embodiments are only some embodiments of the present invention. For those skilled in the art, they can change and improve them without departing from the concept of the present invention. These situations are all within the protection scope of the present invention.

Claims

1. A timing synchronization method for end-to-end voice encryption in a mobile communication network, characterized in that: include: Step 1: Use training sequence frame A for initial synchronization to determine the initial timing detection range; Step 2: Within the timing detection range of the initial synchronization, use the training sequence frame B for accurate synchronization to obtain the precise starting position of the encrypted signal.

2. The timing synchronization method for end-to-end voice encryption in a mobile communication network according to claim 1, characterized in that: include: Training sequence frame setting preparation: the training sequence frame A agreed upon by both ends of the transmitter and receiver is a fixed data sequence; the training sequence frame B agreed upon by both ends of the transmitter and receiver is an orthogonal frequency division multiplexing (OFDM) frame generated by modulation corresponding to a fixed data sequence; the frame lengths of the training sequence frames A and B are set to N and M respectively, that is, A contains N data, B contains M data, and N<M; In the step 1: in the initial synchronization, the transmitting end sends a training sequence frame A, and a first-in-first-out receiving buffer is set at the receiving end to store the received signal, and the received signal is slidably intercepted with a frame length of the training sequence frame A as a unit, as a signal sliding frame, and the energy of each sliding frame and the cross-correlation value with the training sequence frame A are calculated, and finally the preliminary timing detection range is determined by threshold judgment; In the step 2: after the transmitting end sends the training sequence frame A, it continues to send the training sequence frame B, and the receiving end receives the signal to the buffer; taking the timing detection range of the initial synchronization as the starting time, taking the frame length of the training sequence frame B as the unit, slidingly intercepting the received signal as the signal sliding frame, performing OFDM demodulation on each sliding frame, calculating the residual sum of squares of the demodulated data and the original data, and finding the minimum error point to complete precise synchronization.

3. According to the timing synchronization method for end-to-end voice encryption in a mobile communication network according to claim 2, step 1 comprises three sub-steps, including: Step 1-1: Assume that the length of the receiving buffer is 2M+K, where K is a preset data length, to ensure that the buffer can receive data of 2 frames of signal length and has K redundant data bits to ensure robust reception; Step 1-2: Use the local training sequence A of the receiving end and the received signal in the buffer to perform a cross-correlation operation as shown in expression (1), and calculate the energy of the received signal according to expression (2). Then, based on the cross-correlation value, energy and preset threshold, as shown in decision expression (3), a synchronization event decision is made. The expressions of formulas (1) to (3) are as follows: In formula (1), C(x) represents the cross-correlation value between the sliding frame x with a starting time of time t and a data frame length of N and the local training sequence A, represents the signal value of the sliding frame x at time t+k-1, and A(k) represents the signal value of the local training sequence A at the kth time; In formula (2), E(x) represents the frame energy of sliding frame x; In formula (3), E s represents the frame energy of the local training sequence frame A, T s represents a preset threshold coefficient; R(x) is the ratio of the cross-correlation value defined in equation (1) to the frame energy defined in equation (2). If the R(x) value is less than It indicates that no synchronization event occurs at time t, and synchronization detection is stopped; Step 1-3: For the receiving buffer with a length of 2M+K, sequentially intercept sliding frames with a frame length of N from the K+1th time to the K+2M-Nth time, that is, the starting time range of the sliding frame is [K+1, K+2M-N]; Step 1--4: For the sliding frames taken in step 1-3, if R(x) corresponding to all sliding frames is less than This indicates that no synchronization event occurs in the receive buffer; if one or more sliding frames correspond to R(x) greater than or equal to This indicates that a synchronization event has occurred in the receiving buffer. At this time, the sliding frame corresponding to the maximum value of R(x) is taken and recorded as x. max , record the start time of the frame, recorded as t0, and use it as the search time center to search the cross-correlation value forward and backward to obtain a local maximum point before and after, recorded as R(x1) and R(x2) respectively, and the corresponding time is recorded as t1 and t2. Finally, the time range of [x1, x2] is set in the receiving buffer as the preliminary timing detection range.

4. According to the timing synchronization method for end-to-end voice encryption in a mobile communication network as claimed in claim 2, said step 2 comprises two sub-steps, including: Step 2-1: The receiving end receiving buffer receives the training sequence frame B, takes the time range of [t1, t2] as the starting point of the received signal, and intercepts the sliding frame according to the frame length M; Step 2-2: Perform OFDM demodulation on all sliding frames in turn, then calculate the difference between the demodulated data and the data before OFDM modulation of the training sequence frame B, and measure the difference according to the residual square sum of formula (4): In formula (4), SSE is the residual sum of squares, m represents the length of the data sequence before modulation of the training sequence frame B, and y i Indicates the data sequence value before the training sequence frame B modulation, Indicates the data sequence value obtained by sliding frame demodulation; Step 2-3: Among all sliding frames, find the sliding frame with the smallest residual square sum, and determine the precise synchronization point as the starting point of the frame to complete the precise synchronization process.