An Adaptive Adjustment Interference Method Based on Microphone Nonlinear Anti-Eavesdropping

Abstract

The present invention belongs to the field of information security technology, and specifically relates to an adaptive interference adjustment method based on microphone nonlinear anti-eavesdropping. The method includes: step one, construction of the interference signal: separating the interference signal into four 4kHz bandwidth bands through a bandpass filter; adaptively controlling the separated signals of the four different frequency bands; the four bandwidth bands respectively correspond to three different ultrasonic sensor center frequencies and the sampling frequency of the eavesdropping device; step two, transmission of the interference signal: enhancing the center frequency signal; adaptively adjusting the overall power of the interference signal through a power amplifier; and then transmitting the interference signal through an ultrasonic sensor array; the ultrasonic sensor array is composed of three ultrasonic sensors with different center frequencies. The method of the present invention is different from the common microphone anti-eavesdropping device broadcasting strong noise interference masking method, and has the advantages of low cost and high security.

Description

Technical Field

[0001] The present invention belongs to the field of information security technology, and in particular relates to a method for adaptively adjusting interference of a microphone to prevent eavesdropping. Background Art

[0002] Microphone anti-eavesdropping technology has a wide range of applications. With the advancement of science and technology, eavesdropping technology is becoming increasingly sophisticated. Eavesdropping has become a serious problem, posing a privacy and security risk. The leakage of eavesdropped voice information not only violates personal privacy, harms corporate interests, but can even threaten national security. This is why anti-eavesdropping technology has emerged. Today, anti-eavesdropping technology has become an essential part of life, with a wide range of applications, such as anti-eavesdropping devices installed on mobile phones, PN junction detectors, and various eavesdropping alarms.

[0003] Anti-eavesdropping technologies can be categorized into three types: electromagnetic interference, audible sound cancellation, and ultrasonic interference. Because humans can only hear a limited range of sounds, ultrasonic interference can disrupt eavesdropping devices without humans noticing, making it a common application.

[0004] Existing ultrasonic interference mainly relies on the nonlinear effect of the amplifier inside the microphone. Ultrasonic waves can leak energy into the sound spectrum, broadcasting high-frequency, inaudible strong noise. Due to the nonlinear effect of the microphone, the high-frequency noise is demodulated into the low-frequency human vocal band. The noise power is greater than the sound intensity of human speech, thus covering the sound.

[0005] There are also some defects in the implementation of this existing ultrasonic jamming technology. On the one hand, it ignores the opponent's ability to extract meaningful information from sounds corrupted by noise, that is, security. On the other hand, the ultrasonic jamming system does not take into account the impact of factors such as sound attenuation, frequency response, and aliasing effects on the jamming effect, so as to further improve the jamming effect and reduce the jamming cost. Summary of the Invention

[0006] The present invention aims to overcome the shortcomings and deficiencies of existing ultrasonic jamming technologies, such as the common jamming method, which uses a jamming signal composed of Gaussian white noise modulated at 4kHz, 8kHz, or 12kHz bandwidths. This method, which aims to increase coverage by increasing the noise coverage bandwidth, lacks a more targeted theoretical basis; the security of the jammed audio, which prevents the use of audio recovery methods to extract information content; the high cost; and the interference distance. The present invention proposes an anti-eavesdropping jamming technology that leverages factors such as microphone frequency response, sound attenuation characteristics, and aliasing effects to further improve the utilization of jamming signals, reduce costs, and enable adaptive adjustment.

[0007] To solve the above problems, the present invention adopts a technical solution: an adaptive interference adjustment method based on microphone nonlinear anti-eavesdropping, comprising:

[0008] Step 1: Construction of interference signal

[0009] Separate the interference signal into four 4kHz bandwidth bands; adaptively adjust the signals of the four separated frequency bands; the center frequencies of the four bandwidth bands correspond to the center frequencies of three different ultrasonic transducers and the sampling frequency of the eavesdropping device;

[0010] Step 2: Send interference signal

[0011] The center frequency signal is enhanced; the overall power of the interference signal is adaptively adjusted through a power amplifier; and the interference signal is then sent through an ultrasonic transducer array; the ultrasonic transducer array is composed of three ultrasonic transducers with different center frequencies.

[0012] Furthermore, the center frequencies of the four bandwidth bands are 25 kHz, 32 kHz, 40 kHz and the sampling frequency of the eavesdropping device.

[0013] Furthermore, in step 1, the adaptive adjustment of the separated signals of the four different frequency bands is specifically as follows: with respect to the distance, the frequency response of the eavesdropping device, and the sampling frequency, a higher gain coefficient is allocated to the frequency band with poor interference effect based on the sound attenuation coefficient and the frequency response coefficient.

[0014] Furthermore, the interference signal is Gaussian white noise.

[0015] As a microphone nonlinear anti-eavesdropping technology, this invention is different from the common microphone anti-eavesdropping equipment's strong noise interference masking method. Specifically,

[0016] (1) Lower overhead: This invention utilizes the principle of sound cancellation to ensure that the canceling signal strength is equal to the audible sound strength. Compared with common microphone anti-eavesdropping devices that use strong noise interference to mask the audible sound, the overhead is lower. At the same power as the canceling interference, using noise cannot effectively interfere.

[0017] (2) Strong security: After testing and recognizing the audio after interference cancellation, the word error rate (WER) reached over 99%. Two audio recovery methods were used to recover the voice content of the audio after interference. Humans were unable to distinguish the voice content of the recovered audio.

[0018] (3) Selectivity: The present invention can effectively implement separate interference against unauthorized eavesdropping without affecting the recording of voice content by authorized recording equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 Flowchart of a method for adaptively adjusting interference based on microphone nonlinear anti-eavesdropping in an embodiment of the present invention;

[0020] Figure 2 is the attenuation characteristics of different frequencies based on distance;

[0021] Figure 3 The microphone's ability to respond to different frequencies;

[0022] Figure 4 The test results of different factors affecting the effectiveness of the method of the present invention are as follows: (a) the effect of different languages; (b) the effect of different distances; (c) the effect of different angles;

[0023] Figure 5 This is the audio anti-restoration effect processed by the method of the present invention. DETAILED DESCRIPTION

[0024] The following will be combined with the embodiments of the present invention and the accompanying drawings to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. The following description of at least one exemplary embodiment is actually only illustrative and in no way serves as any limitation on the present invention and its application or use. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0025] The present invention provides an adaptive interference adjustment method based on microphone nonlinear anti-eavesdropping, and the basic ideas and technical principles involved are mainly based on the nonlinear characteristics of the microphone and the principle of sound cancellation.

[0026] 1. Microphone nonlinearity

[0027] Microphone nonlinear characteristics But in practice, microphone amplifiers usually exhibit nonlinearity, and the frequency f of the input signal S s Higher than the maximum receiving frequency f of the microphone c When , the input signal will be affected by nonlinearity and the output signal will be in a nonlinear state. The relationship is expressed as:

[0028]

[0029] Although the nonlinear output is theoretically an infinite power series, since the energy of the third-order harmonics and above is very weak, we only consider the second-order terms. At this point, we can express the microphone nonlinearity as:

[0030] S out =A1S+A2S2

[0031] Therefore, when an ultrasonic wave composed of frequencies f1 and f2 is played, the signal can be expressed as:

[0032] S=cos(2πf1t)+cos(2πf2t)

[0033] When the ultrasonic microphone diaphragm amplifier outputs S out We can express it as:

[0034] S out =A1(f1+f2)+A2(f1+f2) 2

[0035] =A1{cos(2πf1t)+cos(2πf2t)}+A2{cos 2 (2πf1t)+cos 2 (2πf2t)+2cos(2πf2t)cos(2πf2t)}

[0036] Now, the frequencies generated by the first-order terms are 2πf1 and 2πf2, which are outside the cutoff point of the microphone. The second-order terms can be expressed as:

[0037]

[0038] Most mobile phone microphones pass through a low-pass filter before being sampled by the ADC. The cutoff frequency of this filter is generally 24kHz. Both f1 and f2 are ultrasonic frequencies, and frequencies above 24kHz are filtered out by the low-pass filter, with the exception of Cos(2π(f1-f2)t). When ((f1-f2)) < 24kHz, a non-ultrasonic frequency is obtained, which is retained by the low-pass filter. This means that the high-frequency sound emitted by ultrasound, which is inaudible, is then received by the microphone as audio that is audible to humans.

[0039] 2. Sound attenuation

[0040] Sound attenuation refers to the decrease in the intensity of a sound wave as it propagates through a medium. To improve the interference effect at long distances, the present invention analyzes sound attenuation, taking into account the three main sources of acoustic attenuation: 1) the inverse square law; 2) absorption attenuation; and 3) diffraction.

[0041] (1) The intensity of a sound wave is inversely proportional to the square of the distance it propagates. As the distance increases, the amplitude of the sound wave per unit area decreases and the degree of attenuation increases.

[0042] (2) During the propagation of sound, the energy consumed by the internal friction and viscosity between particles in the air medium and the coherent molecular motion of the sound waves convert the sound energy into heat energy, which is related to the acoustic frequency. Therefore, the high-frequency signals received by microphones placed at different locations are attenuated more than the low-frequency signals.

[0043] (3) Sound waves diffract when they encounter obstacles or corners. For example, when an obstacle is in the path of a spreading wave, the wave will diffuse along the edge of the obstacle and eventually reach the shadowed area behind it. The amount of diffraction is inversely proportional to the acoustic frequency, meaning that the attenuation of diffraction depends on the frequency of the sound, with higher frequencies causing faster attenuation.

[0044] By measuring and analyzing the attenuation characteristics of smart device microphones at different frequencies, we can use the sound attenuation differences to adaptively allocate different frequency bands, achieving a longer interference distance while reducing losses.

[0045] 3. Aliasing effect

[0046] According to the Nyquist theorem, the sampling frequency must be greater than twice the highest frequency of the signal in order to retain the complete information of the signal without distortion. That is, when the signal is greater than the Nyquist frequency, the sampled signal will be aliased, that is, the high-frequency signal above the Nyquist frequency will be mapped to the observable area below the Nyquist frequency, and the high-frequency signal and the low-frequency signal will be aliased together. The present invention found that although there is a low-pass filter in the microphone circuit to prevent the aliasing effect, aliasing will still occur. This effect is used to increase the complexity of the interference signal by sending a frequency band near the sampling frequency.

[0047] 4. Frequency response

[0048] The reception of sound depends on the frequency response of the microphone. Due to factors such as the characteristics and structure of the hardware material itself, the microphone's response ability to signals of different frequencies varies. Therefore, interference signals in different frequency bands have poor interference effects due to their poor frequency response capabilities. Therefore, it is necessary to analyze the interference effects of different frequencies, determine the frequency band with the best interference effect, and improve the interference effect.

[0049] This paper optimizes the components of the interference signal using the above theory, constructing four segments of 4kHz bandwidth Gaussian white noise, with the center frequencies of each segment at 25kHz, 32kHz, 40kHz, and the eavesdropping sampling frequency. The intensity and power of the four signals are adaptively adjusted to ensure the long-range interference effect and signal complexity.

[0050] Example 1 The present invention provides a method for adaptively adjusting interference based on microphone nonlinear anti-eavesdropping, the process is as follows: Figure 1 The specific steps are as follows:

[0051] Step 1: Construction of interference signal

[0052] (1) Gaussian white noise is used as the main interference signal.

[0053] Gaussian white noise has infinite bandwidth. Based on the sound attenuation characteristics, frequency response characteristics, and the sampling frequency of the eavesdropping device, the present invention separates the Gaussian white noise into four 4kHz bandwidth bands using a bandpass filter at the center frequency of the ultrasonic transducer and the sampling frequency of the eavesdropping device. The four 4kHz bandwidth bands are 23-27kHz, 30-34kHz, 38-42kHz, and (the sampling frequency of the eavesdropping device ± 2)kHz. The center frequencies of the four bandwidth bands are 25, 32, 40kHz, and the sampling frequency of the eavesdropping device, respectively.

[0054] (2) Adaptive adjustment of different frequency bands based on the frequency response of the microphone device.

[0055] like Figure 3 As shown in the figure, the microphone has different response capabilities to different frequencies. The microphone has poor response capabilities to the 25 and 32kHz frequency bands. In order to ensure the effectiveness of the interference signal strength on voice interference, adaptive adjustment is performed for the three frequency bands of 23-27kHz, 30-34kHz, and 38-42kHz. The adaptive adjustment is specifically: roughly, the entire link from the signal sent by the playback device to the microphone receiving it can be regarded as a linear relationship. Therefore, for the input signal strength x, the sending frequency is 25, 32, and 40kHz respectively. At this time, the output signal F_x strength depends on the microphone's response to different frequencies, which can be expressed as: x =k 25 x,F x =k 32 x,F x =k 40 x, where k 25 、k 32 、k 40 They are respectively expressed as the frequency response coefficients of the microphone for the three signals, which can be measured experimentally. Ultimately, the output of signals of different frequency bands is made to have the same intensity. The frequency intensity of the input signal after adaptive adjustment of signal x based on the microphone frequency response is:

[0056] x 25 =k 32 k 40 x,x 32 =k 25 k 40 x,x 40 =k 25 k 32 x, at this time, the microphone output signal can be expressed as F x =k25 k 32 k 40 x,, to achieve adaptive adjustment for different frequency bands. The three frequency bands after adaptive adjustment are superimposed as interference signals.

[0057] Step 2: Signal sending phase

[0058] (1) The ultrasonic transducer array is used for transmission. The ultrasonic transducer array consists of 9 ultrasonic transducers of 3 types. The center frequencies of the three ultrasonic transducers are 25, 32, and 40 kHz respectively, ensuring that interference signals of different frequencies can be transmitted.

[0059] (2) Adding 25, 32, and 40 kHz single-frequency signals to the interference signal can reduce the nonlinear effect and make the noise superimposed in the human voice frequency band. Figure 2 As shown in the figure, experimental testing of the changes in sound intensity at different frequencies at different distances revealed that the change in frequency intensity is primarily related to distance: the farther the distance, the quieter the sound. Therefore, based on this characteristic, the power amplifier dynamically adjusts the overall output power of the jamming signal based on the interference distance, ensuring effective interference against eavesdropping devices at increasing distances, making it applicable to more scenarios and increasing its practicality.

[0060] Example 2 tests the effectiveness of the method of the present invention from the following aspects: the influence of different languages, the influence of distance, the influence of angle, the influence of equipment, and the resistance to speech restoration effect.

[0061] Experimental environment:

[0062] All experiments were conducted in an office with an ambient noise level of approximately 35dB, and the loudness of the audible speech audio obtained by the receiving device was guaranteed to be no less than 75dB.

[0063] The following metrics are used to evaluate the audio quality after accurate cancellation, i.e., intelligibility:

[0064] Perceptual Evaluation of Sound Quality (PESQ), Short-Term Objective Intelligibility (STOI) and Word Error Rate (WER).

[0065] PESQ is standardized in P.862 and is widely used for objective speech quality evaluation. Technically, the input consists of a clear speech signal as a reference and a signal to be measured. The output is a mean opinion score (MOS)

[28] ranging from -0.5 to 4.5. A high PESQ score means that the corresponding speech has a high listening quality, and vice versa. Generally, a PESQ value from 1.00 to 1.99 means "there is no feasible way to understand it".

[0066] WER is the percentage of incorrect words divided by the total number of words in a standard word sequence, and is evaluated using both human and machine recognition. The present invention asked 10 volunteers to listen to the uninterrupted and interfered audio recordings and count the words heard.

[0067] Short-term objective intelligibility (STOI) is an important metric for measuring speech intelligibility. A word in a speech signal can be either intelligible or incomprehensible. In this paper, an STOI value of 0.4 is used as the baseline. Values below 0.4 are considered unintelligible.

[0068] (1) The influence of different languages

[0069] The present invention adopts four languages, namely Chinese, English, Italian and Japanese, and each language is divided into two voices, male and female, for testing.

[0070] The results are as follows Figure 4 As shown in (a), the PESQ and STOI of different languages and genders are not much different, indicating that different languages and genders have little impact on the interference effect.

[0071] (2) Influence of distance

[0072] The eavesdropping devices were placed at distances of 50 cm, 100 cm, 150 cm, 200 cm, 250 cm, and 300 cm from the ultrasonic array, respectively.

[0073] The results are as follows Figure 4 As shown in (b), PESQ and STOI remain consistent within the distance range of 50cm-300cm, indicating that the interference effect is stable within this distance range.

[0074] (3) Influence of angle

[0075] In order to evaluate the effect of the interference angle θ, the horizontal angle θ of the ultrasonic transducer array facing the top microphone of the mobile phone is set to 0°, and θ is changed from -90° to 90° (when the top microphone of the mobile phone is rotated horizontally counterclockwise to perpendicular to the ultrasonic transducer array, θ = 90°, when it is rotated vertically clockwise, θ is equal to -90°), with a change step of 30°. The results are shown in Figure 2. Figure 4 As shown in (c), as the angle θ increases from -90° to 90°, the PESQ and STOI values do not change significantly, with PESQ ranging from 1.03 to 1.06 and STOI ranging from 0.06 to 0.08. This means that interference is effective at all angles, and changes in angle have little effect on the interference effect.

[0076] (4) Impact of different devices

[0077] The present invention uses four smartphones, namely iPhone 13ProMax, Oppo Reno5 Pro, Redmi K30 Ultra, iPadAir3Huawei nova3, and a tablet as eavesdropping devices for testing, and uses two recovery methods: 1) DPCRN convolutional recursive network voice enhancement; 2) using the SOX tool to extract noise features for noise elimination, simulating the situation where the opponent attempts to use these methods to recover information from unauthorized recordings, to prove the security of the method of the present invention.

[0078] Table 1 Interference effect of the method of the present invention on different eavesdropping devices

[0079]

[0080] The table above shows that the PESQ of the five devices after interference ranges from 1.02 to 1.05. The STOI fluctuates greatly due to factors such as different device hardware, but is still below the established baseline. The error rates of human and machine recognition indicate that the audio after interference is still incomprehensible.

[0081] Example 3 Comparison of the present invention's method with different bandwidth interference ranges

[0082] like Figure 5 As shown, the interference signal provided by the method of the present invention is compared with three different broadband interference signals. Except for the frequency range of 50Hz to 1kHz and 1kHz to 4kHz, the interference signal strength of the interference signal provided by the present invention is slightly weaker than the 4kHz bandwidth and 8kHz bandwidth respectively. However, in terms of overall performance, compared with other broadband interference signals, the interference frequency coverage range of the present invention is wider and the interference effect is the best.

Claims

1. An adaptive interference adjustment method based on microphone nonlinear anti-eavesdropping, characterized in that: include: Step 1: Construction of interference signal Separate the interference signal into four 4kHz bandwidth bands; Adaptively adjust the signals of the four separated frequency bands; the center frequencies of the four bandwidth bands correspond to the center frequencies of three different ultrasonic transducers and the sampling frequency of the eavesdropping device; Step 2: Send interference signal The center frequency signal is enhanced; the overall power of the interference signal is adaptively adjusted through a power amplifier; and the interference signal is then sent through an ultrasonic transducer array; the ultrasonic transducer array is composed of three ultrasonic transducers with different center frequencies.

2. The adaptive interference adjustment method based on microphone nonlinear anti-eavesdropping according to claim 1 is characterized in that: The center frequencies of the four bandwidth bands are 25kHz, 32kHz, 40kHz and the sampling frequency of the eavesdropping device.

3. The adaptive interference adjustment method based on microphone nonlinear anti-eavesdropping according to claim 1 is characterized in that: In the step 1, adaptive control is performed on the separated signals of the four different frequency bands, specifically: a higher gain coefficient is allocated to the frequency band with poor interference effect based on the sound attenuation coefficient and the frequency response coefficient in view of the distance, the frequency response of the eavesdropping device, and the sampling frequency.

4. The method for adaptively adjusting interference based on microphone nonlinear anti-eavesdropping according to any one of claims 1 to 3, characterized in that: The interference signal is Gaussian white noise.

Images