Anti-interference sound signal data extraction method based on energy data perception

Abstract

The invention provides an anti-interference sound signal data extraction method based on energy data perception, and belongs to the technical field of underwater sound signal processing, and the method comprises the steps: carrying out the A / D sampling and digital orthogonal demodulation of a received sound signal, and obtaining a baseband signal; performing low-pass filtering and downsampling on the baseband signal to obtain a preprocessed signal; performing matched filtering on the pre-processed signal by using a reference signal; the key improvement lies in that a local energy estimation function of a preprocessed signal is calculated, and normalization processing is carried out on a matched filtering result and the energy estimation function to obtain a normalized detection signal. The method effectively suppresses the shielding effect of a strong interference signal on a weak target signal under a non-orthogonal reference signal and a strong near-end and weak far-end scene, remarkably reduces the misrecognition probability, and improves the accuracy and reliability of the extraction of a long-distance and low-signal-to-noise-ratio target beacon signal in a complex underwater acoustic environment. The method is suitable for underwater navigation, positioning and communication systems.

Description

Technical Field

[0001] This invention relates to the field of underwater acoustic signal data extraction technology, and in particular to an anti-interference acoustic signal data extraction method based on energy data sensing. Background Technology

[0002] Current underwater beacon signal extraction mainly relies on three technical approaches: signal processing algorithm optimization, co-design of coding and modulation, and intelligent noise reduction methods. However, due to the complex environment of multipath effects, dynamic noise, and time-varying channel characteristics, existing technologies still face significant bottlenecks in practical applications. Current mainstream systems are based on signal processing algorithms as their core architecture. For example, real-time noise suppression schemes based on adaptive filtering technology can effectively cope with steady-state interference, but they are highly sensitive to sudden impulse interference, and time-frequency analysis algorithms such as wavelet transform suffer from incompatibility issues between computational complexity and real-time performance. Regarding anti-interception enhancement, coding and modulation techniques can reduce the bit error rate caused by multipath effects through redundant coding mechanisms such as LDPC-GMSK, but the inherent limitations of pre-set static channel models make them difficult to adapt to dynamic propagation environments such as thermohaline layers. While intelligent noise reduction models, which have emerged in recent years, have shown advantages in feature extraction in low signal-to-noise ratio scenarios, their training relies on large-scale labeled datasets, and the deployment cost due to the large number of parameters limits their application in embedded devices.

[0003] Because the reference signals between different beacons are not perfectly orthogonal, even after matched filtering, the beacon signal closer to the receiving device still occupies a larger amplitude in the matching result, potentially masking the true detection result of the target beacon signal. This leads to misjudgment of long-distance beacon signals. To address this problem, this invention proposes an anti-interference acoustic signal data extraction method based on energy data sensing. Summary of the Invention

[0004] Therefore, the purpose of this invention is to provide an anti-interference acoustic signal data extraction method based on energy data perception, so as to at least solve the above problems.

[0005] The technical solution adopted in this invention is as follows: An anti-interference acoustic signal data extraction method based on energy data sensing, the method comprising the following steps: S1. Perform A / D sampling on the received acoustic signal for each cycle to obtain the sampled sequence signal. ; for the sampled sequence signal Digital quadrature baseband demodulation is performed to obtain the baseband signal. ; S2, the baseband signal Represented in frequency domain form, wherein the frequency domain form is based on the demodulation carrier frequency. Sampling period Phase correction factor Index of the received signal sampling point

[0006] S3, regarding the baseband signal Low-pass filtering and downsampling are performed to obtain the preprocessed signal. ; S4. Use reference signal For the preprocessed signal Perform matched filtering to obtain the filtered signal. ; S5. Based on the preprocessed signal Calculate the energy estimation function ; S6. Based on the filtered signal and the energy estimation function Normalization is performed to obtain the normalized detection signal. It is used to identify target signals.

[0007] Furthermore, in step S1, the sampling rate of the A / D sampling is set to 80 kS / s; the digital quadrature baseband demodulation is achieved through the following formula:

[0008] Among them, carrier frequency 7.5 kHz, phase correction factor Used to compensate for the initial phase offset of the carrier; each period of the received acoustic signal is 10 seconds.

[0009] Furthermore, in step S2, the baseband signal Represented in frequency domain form, specifically through mathematical equivalent transformations to express it in frequency domain spectral form:

[0010] Frequency domain representation is used to guide the parameter design of low-pass filters, so as to achieve the refraction of the original signal spectrum around ± The filter is shifted to the vicinity of DC, thereby using a low-order low-pass filter to achieve sidelobe suppression and bandwidth limitation.

[0011] Furthermore, in step S3, the low-pass filtering is implemented using a finite impulse response (FIR) filter or an infinite impulse response (IIR) filter, with a passband frequency range of 0 to 5 kHz, used to retain effective baseband information and suppress high-frequency noise components; the downsampling process is an integer multiple downsampling with a downsampling factor of 4, performed after the low-pass filtering, to obtain the downsampled signal. As a preprocessed signal, it is represented as:

[0012] in, The impulse response of the low-pass filter is given.

[0013] Furthermore, in step S4, the reference signal The baseband template for the transmitted signal is generated by applying the same signal processing procedure as S1 and S3 to the known transmitted signal; the matched filtering process calculates the preprocessed signal. With the reference signal The cross-correlation is achieved, and the specific formula is as follows:

[0014] in, The length of the matched filter is equal to the reference signal. Length, This is the index of the matched filter.

[0015] Furthermore, the energy estimation function By calculating the preprocessed signal In length of The local energy within the sliding window is obtained; the specific calculation formula is:

[0016] Introducing energy operators The energy estimation function is obtained. :

[0017] Wherein, the length of the sliding window The length is consistent with that of the matched filter described in step S4.

[0018] Furthermore, in step S6, the normalization process is performed by normalizing the filtered signal. Divided by the energy estimation function To achieve this, we obtain the normalized detection signal:

[0019] The normalized detection signal Used to suppress amplitude masking effects caused by nearby strong interfering beacon signals.

[0020] Furthermore, after step S6, a target signal identification step is also included: based on the normalized detection signal Set an adaptive detection threshold; when When the peak value exceeds the adaptive detection threshold, it is determined to be a valid peak value of the target signal; the adaptive detection threshold is dynamically set based on the noise statistical characteristics of the normalized detection signal in the segment without target signal, specifically the noise mean plus K times the standard deviation, where K is a preset constant.

[0021] Furthermore, the target signal identification step also includes: if the normalized detection signal If a peak value exceeding the adaptive detection threshold is identified, the largest peak value is selected as the main target signal peak value; and other peak values ​​are subjected to consistency verification, which includes determining whether the peak interval is greater than the minimum time interval threshold or whether the peak position across cycles is stable.

[0022] Compared with the prior art, the beneficial effects of the present invention are: This invention introduces a normalization step based on local signal energy. In typical scenarios where multiple beacons operate simultaneously, especially using non-orthogonal or similar reference signals at varying distances from the receiver, the absolute amplitude of the traditional matched filter output is significantly amplified by the strong signal from the near-end beacon, causing the peak value of the weak signal at the far end to be submerged. This invention addresses this issue by introducing a normalization step based on local signal energy. Divided by local energy estimation The detection quantity is converted from "absolute correlation amplitude" to "relative signal-to-noise ratio intensity", which makes the peak value of weak targets at the far end stand out under the background of strong near-end interference, fundamentally solving the technical problem of "near end occlusion of far end". Attached Figure Description

[0023] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only preferred embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1 This is a flowchart illustrating an embodiment of the present invention.

[0025] Figure 2 This is a schematic diagram of the original sequence of an embodiment of the present invention.

[0026] Figure 3 This is a schematic diagram of the matched filter output result according to an embodiment of the present invention. Figure 4 This is a schematic diagram of the final signal result after normalization according to an embodiment of the present invention. Detailed Implementation

[0027] The principles and features of the present invention are described below with reference to the accompanying drawings. The listed embodiments are only used to explain the present invention and are not intended to limit the scope of the present invention.

[0028] Reference Figure 1 This invention provides a method for extracting anti-interference acoustic signal data based on energy data sensing, the method comprising the following steps: S1. Perform A / D sampling on the received acoustic signal for each cycle to obtain the sampled sequence signal. ; for the sampled sequence signal Digital quadrature baseband demodulation is performed to obtain the baseband signal. ; S2, the baseband signal Represented in frequency domain form, wherein the frequency domain form is based on the demodulation carrier frequency. Sampling period Phase correction factor Index of the received signal sampling point

[0029] S3, regarding the baseband signal Low-pass filtering and downsampling are performed to obtain the preprocessed signal. ; S4. Use reference signal For the preprocessed signal Perform matched filtering to obtain the filtered signal. ; S5. Based on the preprocessed signal Calculate the energy estimation function ; S6. Based on the filtered signal and the energy estimation function Normalization is performed to obtain the normalized detection signal. It is used to identify target signals.

[0030] In step S1, the sampling rate of the A / D sampling is set to 80 kS / s; the digital quadrature baseband demodulation is achieved through the following formula:

[0031] Among them, carrier frequency 7.5 kHz, phase correction factor Used to compensate for the initial phase offset of the carrier; each period of the received acoustic signal is 10 seconds.

[0032] In step S2, the baseband signal Represented in frequency domain form, specifically through mathematical equivalent transformations to express it in frequency domain spectral form:

[0033] Frequency domain representation is used to guide the parameter design of low-pass filters, so as to achieve the refraction of the original signal spectrum around ± The filter is shifted to the vicinity of DC, thereby using a low-order low-pass filter to achieve sidelobe suppression and bandwidth limitation.

[0034] In step S3, the low-pass filtering is implemented using a finite impulse response (FIR) filter or an infinite impulse response (IR) filter, with a passband frequency range of 0 to 5 kHz, used to retain effective baseband information and suppress high-frequency noise components; the downsampling process is an integer multiple downsampling with a downsampling factor of 4, performed after the low-pass filtering, to obtain the downsampled signal. As a preprocessed signal, it is represented as:

[0035] in, The impulse response of the low-pass filter is given.

[0036] In step S4, the reference signal The baseband template for the transmitted signal is generated by applying the same signal processing procedure as S1 and S3 to the known transmitted signal; the matched filtering process calculates the preprocessed signal. With the reference signal The cross-correlation is achieved, and the specific formula is as follows:

[0037] in, The length of the matched filter is equal to the reference signal. Length, This is the index of the matched filter.

[0038] The energy estimation function By calculating the preprocessed signal In length of The local energy within the sliding window is obtained; the specific calculation formula is:

[0039] Introducing energy operators The energy estimation function is obtained. :

[0040] Wherein, the length of the sliding window The length is consistent with that of the matched filter described in step S4.

[0041] In step S6, the normalization process is performed by changing the filtered signal. Divided by the energy estimation function To achieve this, we obtain the normalized detection signal:

[0042] The normalized detection signal Used to suppress amplitude masking effects caused by nearby strong interfering beacon signals.

[0043] Step S6 is followed by a target signal recognition step: based on the normalized detection signal Set an adaptive detection threshold; when When the peak value exceeds the adaptive detection threshold, it is determined to be a valid peak value of the target signal; the adaptive detection threshold is dynamically set based on the noise statistical characteristics of the normalized detection signal in the segment without target signal, specifically the noise mean plus K times the standard deviation, where K is a preset constant.

[0044] The target signal identification step further includes: if the normalized detection signal If a peak value exceeding the adaptive detection threshold is identified, the largest peak value is selected as the main target signal peak value; and other peak values ​​are subjected to consistency verification, which includes determining whether the peak interval is greater than the minimum time interval threshold or whether the peak position across cycles is stable.

[0045] For example, this embodiment was applied to an underwater beacon positioning test in the northern part of the South China Sea, aiming to verify the effectiveness of the present invention in extracting weak beacon signals from distant locations in the presence of strong near-end interference.

[0046] Regarding the test environment and deployment, in terms of the marine environment: the test area has a water depth of over 3500 meters, and the environmental noise is mainly low-frequency ship noise and marine environmental noise. The channel exhibits significant multipath effects and time-varying characteristics. In terms of equipment and deployment: three underwater acoustic beacons (denoted as beacon1, beacon2, and beacon3) were deployed on the seabed. A survey vessel towed a hydrophone array for mobile reception. All beacons and the shipborne receiving system were synchronized using a high-precision atomic clock. In terms of scenario setup: the three beacons used the same spread spectrum modulation scheme, but used a non-completely orthogonal spread spectrum code sequence. Among them, beacon1 and beacon2 were closer to the receiving vessel (about 2-5 kilometers), while beacon3 was farther away (about 10-12 kilometers), forming a typical "strong near-end, weak far-end" test scenario.

[0047] Regarding signal parameters and front-end settings, in terms of signal frame structure: each beacon transmits an acoustic signal within a fixed time slot with a 10-second period; in terms of receiver parameters, the A / D sampling rate is 80 kS / s (thousand samples / second); and the receiving carrier frequency... 7.5kHz; In the analog front-end passband: effectively covers the operating frequency band, provides suppression of out-of-band noise (especially 0-1 kHz and above 40 kHz) of not less than 20 dB, and the equivalent input noise of the receiving system is less than 1.7μV.

[0048] In the signal processing flow: Taking the extraction of signals from the remote beacon 3 as an example, the processing flow is as follows: For step S1, the analog acoustic signal received by signal acquisition and quadrature demodulation is converted into a discrete sequence by A / D conversion. Then, digital quadrature demodulation is performed using the carrier frequency. =7.5kHz and the phase correction factor obtained through initial synchronization estimation According to the formula Down-converted to baseband to obtain baseband signal This step shifts the signal spectrum to near zero frequency.

[0049] For step S2, low-pass filtering and downsampling are performed on the baseband signal. Through an FIR low-pass filter with a passband of 0-5 kHz ( To retain valid information and suppress high-frequency noise, the filtered signal is downsampled by a factor of 4, resulting in a preprocessed signal with a lower data rate that is more conducive to subsequent processing. This reduces the amount of data required for subsequent calculations.

[0050] For step S3, matched filtering, the baseband reference signal is generated in advance based on the known transmit chips of beacon3. Its length M is consistent with the code length, and the preprocessed signal and Perform cross-correlation (i.e., matched filtering) to obtain the correlation output. Matched filtering can compress signal energy and provide processing gain.

[0051] For step S4, local energy estimation, a sliding window with the same length M as the matched filter can be used to calculate the preprocessed signal. The local energy is obtained, and the energy estimation function is derived. This function reflects the local energy background of the signal at each time point, including the sum of the energy of the target signal, noise, and potential strong interference signals.

[0052] For step S5, energy normalization detection, the matched filter output can be... Divided by local energy estimation The normalized detection signal is obtained: .

[0053] For step S6, target identification and decision, the normalized detection signal can be used. Perform peak search. Set an adaptive detection threshold based on... The noise mean and standard deviation are calculated during periods when there is no target signal (e.g., threshold = noise mean + 3 times standard deviation). When the peak value exceeds this threshold, it is determined that the target signal has been detected. If there are multiple peak values ​​exceeding the threshold, the highest peak value is selected as the arrival time of beacon3, and time interval consistency verification (such as checking whether the peak interval conforms to the beacon's transmission timing) is used to eliminate false peaks.

[0054] For the comparison of effects Figure 2 The display shows the baseband signal within one reception cycle, where the peak values ​​of the near-end beacon 1 and beacon 2 are significant, while the peak value of the far-end beacon 3 is weaker, and the background is mixed with environmental noise.

[0055] Figure 3 , 4 This displays the baseband signal within a reception cycle. or It can be observed that the strong impulse responses generated by beacon1 and beacon2 are much larger than the weak response of beacon3. The signal of beacon3 is almost drowned out by background noise and near-end interference.

[0056] Figure 3 This demonstrates that the traditional method only proceeds up to the output of the poorly matched S3 filter. It is evident that the correlation peaks of beacon1 and beacon2 are very sharp and have extremely high amplitudes, while the peak of beacon3 is almost invisible and is easily misjudged as noise or directly obscured by strong peaks.

[0057] Figure 4 The normalized detection signal obtained by applying the method of the present invention is shown. After energy normalization, the strong peaks of beacon1 and beacon2 were significantly suppressed (relative amplitudes were reduced), while the peak of beacon3 was clearly highlighted and could be accurately identified and extracted.

[0058] Therefore, this embodiment fully demonstrates the effectiveness of the method of the present invention. In harsh environments with strong near-end interference, the present invention successfully solves the problem of signal obscuring from weak targets at long distances through energy sensing normalization technology, significantly improving the accuracy and reliability of underwater acoustic signal detection.

[0059] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for extracting anti-interference acoustic signal data based on energy data sensing, characterized in that, The method includes the following steps: S1. Perform A / D sampling on the received acoustic signal for each cycle to obtain the sampled sequence signal. ; for the sampled sequence signal Digital quadrature baseband demodulation is performed to obtain the baseband signal. ; S2, the baseband signal Represented in frequency domain form, wherein the frequency domain form is based on the demodulation carrier frequency. Sampling period Phase correction factor Index of the received signal sampling point ; S3, regarding the baseband signal Low-pass filtering and downsampling are performed to obtain the preprocessed signal. ; S4. Use reference signal For the preprocessed signal Perform matched filtering to obtain the filtered signal. ; S5. Based on the preprocessed signal Calculate the energy estimation function ; S6. Based on the filtered signal and the energy estimation function Normalization is performed to obtain the normalized detection signal. It is used to identify target signals.

2. The method for extracting anti-interference acoustic signal data based on energy data sensing according to claim 1, characterized in that, In step S1, the sampling rate of the A / D sampling is set to 80 kS / s; the digital quadrature baseband demodulation is achieved through the following formula: Among them, carrier frequency 7.5 kHz, phase correction factor Used to compensate for the initial phase offset of the carrier; each period of the received acoustic signal is 10 seconds.

3. The method for extracting anti-interference acoustic signal data based on energy data perception according to claim 1 or 2, characterized in that, In step S2, the baseband signal Represented in frequency domain form, specifically through mathematical equivalent transformations to express it in frequency domain spectral form: Frequency domain representation is used to guide the parameter design of low-pass filters, so as to achieve the refraction of the original signal spectrum around ± The filter is shifted to the vicinity of DC, thereby using a low-order low-pass filter to achieve sidelobe suppression and bandwidth limitation.

4. The method for extracting anti-interference acoustic signal data based on energy data sensing according to claim 1, characterized in that, In step S3, the low-pass filtering is implemented using a finite impulse response (FIR) filter or an infinite impulse response (IR) filter, with a passband frequency range of 0 to 5 kHz, to preserve effective baseband information and suppress high-frequency noise components; the downsampling process is an integer multiple downsampling with a downsampling factor of 4, performed after the low-pass filtering, to obtain the downsampled signal. As a preprocessed signal, it is represented as: in, The impulse response of the low-pass filter is given.

5. The method for extracting anti-interference acoustic signal data based on energy data sensing according to claim 1, characterized in that, In step S4, the reference signal It is generated by using the same signal processing procedure as S1 and S3 on the known transmitted signal, and it serves as the baseband template for the transmitted signal; The matched filtering process calculates the preprocessed signal. With the reference signal The cross-correlation is achieved, and the specific formula is as follows: in, The length of the matched filter is equal to the reference signal. Length, This is the index of the matched filter.

6. The method for extracting anti-interference acoustic signal data based on energy data perception according to claim 5, characterized in that, The energy estimation function By calculating the preprocessed signal In length The local energy within the sliding window is obtained; the specific calculation formula is: Introducing energy operators The energy estimation function is obtained. : Wherein, the length of the sliding window The length is consistent with that of the matched filter described in step S4.

7. The method for extracting anti-interference acoustic signal data based on energy data perception according to claim 1, characterized in that, In step S6, the normalization process is performed by changing the filtered signal. Divided by the energy estimation function To achieve this, we obtain the normalized detection signal: The normalized detection signal Used to suppress amplitude masking effects caused by nearby strong interfering beacon signals.

8. A method for extracting anti-interference acoustic signal data based on energy data sensing according to claim 1 or 7, characterized in that, Step S6 is followed by a target signal recognition step: based on the normalized detection signal Set an adaptive detection threshold; when When the peak value exceeds the adaptive detection threshold, it is determined to be a valid peak value of the target signal; the adaptive detection threshold is dynamically set based on the noise statistical characteristics of the normalized detection signal in the segment without target signal, specifically the noise mean plus K times the standard deviation, where K is a preset constant.

9. The method for extracting anti-interference acoustic signal data based on energy data perception according to claim 8, characterized in that, The target signal identification step further includes: if the normalized detection signal If a peak value exceeding the adaptive detection threshold is identified, the largest peak value is selected as the main target signal peak value; and other peak values ​​are subjected to consistency verification, which includes determining whether the peak interval is greater than the minimum time interval threshold or whether the peak position across cycles is stable.

Images