A multi-base buoy fusion detection method and system based on complex interference

Abstract

The present invention discloses a multi-base buoy fusion detection method and system under complex interference. The method includes: preprocessing the received signal to obtain the threshold value for target detection; based on the threshold value, performing fusion detection through soft decision-making weighted by signal-to-noise ratio. The preprocessing includes: MMSE spectral estimation for noise reduction, frequency-domain matched filtering based on Doppler template matching, time reversal deconvolution processing, and cell averaging constant false alarm rate detection for the target. The present invention improves the detection performance, as well as the generalization and robustness under the background of complex ocean noise.

Description

Technical Field

[0001] The present invention relates to the target technical field, in particular to a multi-base buoy fusion detection method and system under complex interference. Background Art

[0002] In the current complex naval warfare environment, due to the improvement of the concealment performance of underwater targets such as submarines, the enhancement of the maneuverability, the complexity of the hydrological environment, and the increasing perfection of sonar weapon countermeasure performance, the underwater detection and tracking performance based on a single platform and a single base is insufficient. Moreover, when a single-base sonar is an active sonar with co-located transceiver, it is easy to expose its own position. This has promoted multi-base detection to become a new anti-submarine detection system. In a multi-base sonar detection system, the transceivers of each detection node are separated, and the concealment is better than that of co-located nodes. Moreover, compared with single-node detection, multi-node collaborative detection can expand the detection range through reasonable base station configuration. By fusing the target state information obtained by multiple nodes, the detection probability and positioning accuracy of underwater targets can be improved. Therefore, using a multi-base sonar system for target detection and tracking is the development trend of current underwater target detection and tracking technology.

[0003] In a conventional distributed multi-base sonar system, the matched filtering technology is used for detection at each single base station. However, since the separation of the transceiver in a multi-base sonar easily leads to direct wave interference, and combined with multi-path interference in a complex ocean environment, etc., the matched filtering detection is inaccurate. When performing multi-base fusion detection, it is easier to generate missed detections or false alarms, resulting in a decline in the detection performance of the system. Moreover, the traditional fusion does not consider the contribution of the local decisions of each base station to the decision of the fusion center. Summary of the Invention

[0004] In view of this, the present invention provides a multi-base buoy fusion detection method and system under complex interference to solve the above technical problems.

[0005] The present invention discloses a multi-base buoy fusion detection method under complex interference, which includes:

[0006] Preprocessing the received signal to obtain the threshold value for target detection;

[0007] Combining the threshold value, performing fusion detection based on soft decision with signal-to-noise ratio weighting;

[0008] The preprocessing includes: MMSE spectral estimation noise reduction, frequency-domain matched filtering based on Doppler template matching, time reversal deconvolution processing, and target cell average constant false alarm rate detection.

[0009] Further, performing MMSE spectral estimation noise reduction on the received signal includes:

[0010] Subjecting the received signal to pre-emphasis, frame division, and windowing processing;

[0011] The time-domain framed signal is subjected to short-time Fourier transform to obtain the energy spectrum:

[0012] Estimate the background noise power spectrum, the posterior signal-to-noise ratio, and the prior signal-to-noise ratio;

[0013] Calculate the gain function and estimate the signal spectrum;

[0014] Phase recovery and inverse Fourier transform are performed to obtain the time-domain signal estimate;

[0015] The time-domain signals of each segment are combined to output the denoised signal and the equivalent signal-to-noise ratio.

[0016] Furthermore, the energy spectrum is:

[0017] P(k) = |X(k)| = |FFT(x(n))|

[0018] where x(n) = s(n) + d(n) is the received signal, which belongs to the time-domain framed signal, s(n) is the pure signal to be detected, and d(n) is the noise;

[0019] The estimation of the background noise power spectrum, the posterior signal-to-noise ratio, and the prior signal-to-noise ratio includes:

[0020] Select the first few frames without signal arrival as the background noise, and estimate its power spectrum λ d (k) = E{|D(k)|}, and estimate the posterior signal-to-noise ratio γ k and the prior signal-to-noise ratio ξ k :

[0021]

[0022] where λ d (k) is the mean value of the noise variance when there is no signal arrival in several segments, m is the signal segmentation sequence, and a is a coefficient.

[0023] Furthermore, the calculation of the gain function and the estimation of the signal spectrum include:

[0024] The expression of the gain function is:

[0025]

[0026] where G k is the gain function, Γ(·) is the Gamma function, and I0(·) and I1(·) represent the zero-order and first-order Bessel functions respectively;

[0027] The spectrum estimation of each segment of the signal is obtained by the following formula:

[0028]

[0029] where The spectrum estimation for each segment of the signal;

[0030] The time-domain signal is estimated as:

[0031]

[0032] where θ k is obtained from the noisy signal;

[0033] The equivalent signal-to-noise ratio of the signal after noise reduction processing is:

[0034]

[0035] where M is the number of signal segments.

[0036] Further, perform frequency-domain matched filtering based on Doppler template matching on the signal after noise reduction processing, including:

[0037] Perform matched filtering processing on the signal after noise reduction processing in the frequency domain respectively, and the output is:

[0038]

[0039] where ξ i corresponds to different Doppler frequency offsets.

[0040] Further, perform time reversal deconvolution processing on the output of the matched filter, including:

[0041] Perform time reversal convolution processing on the output of the matched filter, that is

[0042] T Y (ω) = Y(-ω) * Y(ω)

[0043] Pass T Y (ω) through the interference suppression gate to eliminate the main noise interference:

[0044] Y(ω) = T Y (ω) · I(ω)

[0045] where,

[0046] Perform inverse Fourier transform on Y(ω), and the finally output time-domain signal is:

[0047] y(n) = IFFT(Y(ω))

[0048] Perform target cell-averaging constant false alarm rate detection on y(n), including:

[0049] Set M reference cells, estimate the interference noise power level within the reference window, and multiply by the threshold coefficient λ = P fa-1 / M -1 Obtain the threshold value T.

[0050] Furthermore, combining the threshold value, the fusion detection based on the soft decision weighted by the signal-to-noise ratio includes:

[0051] Each single base station adopts a three-level decision threshold, T2 < T < T1, that is, the observation space and are respectively further divided into two parts, where satisfies satisfies The soft decision rule is:

[0052]

[0053] After the fusion center receives the decision data of each base station, the final decision is achieved by estimating the detection statistic and the detection threshold.

[0054] Furthermore, combining the threshold value, the fusion detection based on the soft decision weighted by the signal-to-noise ratio specifically includes:

[0055] According to the false alarm probability, the received signal processing signal-to-noise ratio, and the reference unit of each base station, obtain the three-level threshold coefficients;

[0056] Each single base station makes a decision on the target according to the soft decision rule;

[0057] Calculate the detection statistic and the threshold value of the fusion center;

[0058] The fusion center makes the final decision using the calculated detection statistic and the detection threshold.

[0059] Furthermore, the three-level threshold coefficients are respectively:

[0060] λ = P fa -1 / M -1

[0061] λ1 = 2 -1 / M (η + P fa -1 / M ) - (1 + η), η = 10 0.1·SNR

[0062] λ2 = ((1 + P fa ) / 2) -1 / M -1

[0063] Each base multiplies the interference noise power level estimated by the unit average constant false alarm rate by the coefficients λ, λ1, and λ2 respectively to obtain the three-level threshold values;

[0064] Each single base station makes a decision on the target according to the soft decision rule, that is

[0065]

[0066] And transmit the corresponding judgment data and the signal-to-noise ratio information of the received signal processing to the fusion center.

[0067] Further, the calculation of the detection statistic and the threshold value of the fusion center includes:

[0068] The detection statistic of the fusion center is the weighted value of the judgment values of each base station. By introducing the threshold coefficients λ3 and λ4, the data corresponding to the judgment values of each base station are estimated, that is, λ3 = αλ1, λ4 = αλ2, where α is a given parameter. Then the data corresponding to the judgment value can be estimated as:

[0069]

[0070] The detection statistic of the fusion center can be estimated as where N is the number of base stations, and β i (k) is the weight of the k-th judgment of the i-th base station;

[0071] At each fusion judgment, the fusion center updates the weights of each base station, that is:

[0072]

[0073] where η i (k) represents the signal-to-noise ratio of the k-th judgment of the i-th base station, and β i (0) = 1, and

[0074] The judgment threshold of the fusion center can be estimated as: where T i is the threshold obtained by CFAR detection of each base station.

[0075] The present invention also discloses a system applicable to the above-mentioned multi-base buoy fusion detection method under complex interference. The system includes:

[0076] A preprocessing module for preprocessing the received signal to obtain the threshold value for target detection; the preprocessing includes: MMSE spectral estimation noise reduction, frequency-domain matched filtering based on Doppler template matching, time reversal deconvolution processing, and target cell average constant false alarm rate detection;

[0077] A fusion detection module for performing fusion detection based on soft decision-making weighted by the signal-to-noise ratio in combination with the threshold value.

[0078] Due to the above technical solution, the present invention has the following advantages: The present invention is based on a distributed multi-base buoy, and performs active detection with separated transceiver for the target. Through preprocessing such as MMSE signal enhancement technology, Doppler template matching technology, time reversal deconvolution interference suppression technology, etc. at each single base station, local soft decision detection of the target under complex interference is achieved at the single base station. Then, the detection performance, as well as the generalization and robustness under the complex ocean noise background, are greatly improved through signal-to-noise ratio weighted fusion detection. It can be widely applied to distributed multi-base systems of base station platforms such as buoys, UUVs, ships, etc., providing accurate target detection information for combat and civilian applications. BRIEF DESCRIPTION OF THE DRAWINGS

[0079] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the drawings described below are only some embodiments recorded in the embodiments of the present invention. For those of ordinary skill in the art, other drawings can also be obtained based on these drawings.

[0080] Figure 1 Schematic diagram of the multi-base buoy layout in the embodiment of the present invention;

[0081] Figure 2 Schematic diagram of the multi-base detection range distribution in the embodiment of the present invention;

[0082] Figure 3 Schematic diagram of the multi-base buoy fusion detection process based on complex interference in the embodiment of the present invention;

[0083] Figure 4 Schematic diagram of the processing result of the MMSE noise reduction algorithm in the embodiment of the present invention;

[0084] Figure 5 Schematic diagram of the comparison before and after Doppler template matching in the embodiment of the present invention;

[0085] Figure 6 Schematic diagram of the processing result of the time reversal deconvolution interference suppression algorithm in the embodiment of the present invention;

[0086] Figure 7 Schematic diagram of the comparison between the fusion detection center and the single-station detection curve in the embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0087] The present invention will be further described in conjunction with the drawings and embodiments. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art shall fall within the scope of protection of the embodiments of the present invention.

[0088] See Figure 1, this invention provides an embodiment of a multi-base buoy fusion detection method under complex interference.

[0089] 1. Distributed multi-base buoy system

[0090] The multi-base buoy layout is a one-transmitter and three-receiver system as shown in Figure 1 . The multi-base sonar can be regarded as a combination of multiple bistatic sonars. The bistatic sonar equation under the condition of considering noise suppression is:

[0091] SL - TL1 - TL2 + TS - NL + DI = DT

[0092] The condition for detecting a target is that the left side of the equation is not less than the right side. If the propagation loss is spherical wave loss and the acoustic absorption of the medium is not considered, then:

[0093] r1·r2 ≤ 10 (SL+TS-NL+DI-DT) / 10

[0094] where r1 and r2 are the distances from the transmitter and the receiver to the target respectively. It can be seen that the action range of the bistatic sonar is a Cassini oval curve. Therefore, the detection range of the multi-base buoy system is as shown in Figure 2 . In addition, the multi-base system of this invention uses the Gold sequence coded signal as the transmitted signal. Because of its excellent orthogonal performance and signal processing gain, it is suitable for multi-base detection.

[0095] 2. Multi-base fusion detection method under complex interference

[0096] In the actual ocean environment, the underwater acoustic channel is complex and changeable, and the signal-to-noise ratio is low. For a multi-base sonar with separated transmitter and receiver, there are also serious direct wave interference and multi-path interference, etc. Therefore, this invention includes preprocessing detection of each single base station under complex interference and soft decision fusion detection based on signal-to-noise ratio weighting. The process is as shown in Figure 3 .

[0097] The specific steps are as follows:

[0098] Step 1: MMSE spectral estimation noise reduction

[0099] For the noise interference in ocean acoustic propagation, signal enhancement technology is adopted, which can effectively suppress the noise component while retaining the information of the signal and improve the detection ability. The underwater acoustic signal is a non-stationary random process that changes with time. When the signal is divided according to a small time interval, it can be approximately considered that the signal is short-time stationary within each time segment. For stationary signals, this invention adopts a signal enhancement technology based on the minimum mean square error spectral estimation method (MMSE). Its estimation criterion is based on the criterion of the minimum mean square error of the estimated amplitude and the actual amplitude. The specific process includes:

[0100] (1) Signal preprocessing

[0101] The received signal is processed through pre-emphasis, framing, and windowing.

[0102] (2) Calculate the short-time energy spectrum

[0103] The time-domain framed signal x(n) = s(n) + d(n) is transformed through the short-time Fourier transform of N points to obtain the energy spectrum:

[0104] P(k) = |X(k)| = |FFT(x(n))|

[0105] (3) Estimate the background noise power spectrum, posterior SNR, and prior SNR

[0106] Select the first few frames without signal arrival as the background noise, and estimate its power spectrum λ d (k) = E{|D(k)|}, and estimate the posterior SNR γ k and the prior SNR ξ k :

[0107]

[0108] where λ d (k) is the mean value of the noise variance when there is no signal arrival in several segments, m is the signal segmentation sequence, and the coefficient a can be taken as 0.98.

[0109] (4) Calculate the gain function G k , and estimate the signal spectrum

[0110] The expression of the gain function is:

[0111]

[0112] where, Γ(·) is the Gamma function, and I0(·) and I1(·) represent the zero-order and first-order Bessel functions, respectively.

[0113] The spectrum estimation of each segment of the signal is obtained by the following formula:

[0114]

[0115] (5) Phase recovery and inverse Fourier transform to obtain the estimated time-domain signal

[0116]

[0117] where, θ k is obtained from the noisy signal.

[0118] (6) Combine each segment of the time-domain signal and output the complete time-domain signal after noise reduction and the equivalent SNR

[0119] Figure 4 The signal after MMSE noise reduction is given, and it can be seen that the signal quality has been greatly improved.

[0120] Step 2: Frequency-domain matched filtering based on Doppler template matching

[0121] For the Gold transmitted signal, its Doppler sensitivity determines that when the direct wave interference passes through the matched filter of the target echo frequency offset, the output drops significantly, which is the core of direct wave interference suppression.

[0122] The specific implementation steps are as follows: When there is a certain prior knowledge of the target speed change range, estimate the target frequency offset change range, design a group of matched filters with different Doppler frequency offsets, and perform matched filtering on the signal after noise reduction in step 1 in the frequency domain respectively. The output is:

[0123]

[0124] The filter with the largest output peak is the one that best matches the target Doppler. Figure 5 The comparison before and after Doppler template matching is given, and it can be seen that the direct wave interference is suppressed to a lower-level sidelobe.

[0125] Step 3: Time reversal deconvolution to suppress multipath interference

[0126] In the shallow sea environment, there is a strong multipath effect, which makes the sound wave emitted by the sound source may reach the hydrophone along several different paths successively. The received signal will not only be attenuated but also severely distorted, which will directly affect the performance of underwater target detection. The time reversal deconvolution technology can reconstruct the multipath signal without any prior environmental information, adaptively match the channel, and has time compression and space focusing performance, making it have a certain anti-multipath interference performance.

[0127] The specific steps of time reversal deconvolution interference suppression are as follows:

[0128] (1) Time reversal deconvolution

[0129] Perform time reversal deconvolution on the signal processed in step 2, that is

[0130] T Y (ω) = Y(-ω) * Y(ω)

[0131] (2) Interference suppression

[0132] Pass T Y (ω) through the interference suppression gate to eliminate the main noise interference:

[0133] y(n) = T Y(ω)·I(ω)

[0134] Among them,

[0135] (3) Inverse Fourier transform to output the time-domain signal

[0136] After performing the inverse Fourier transform on the above signal, the finally output time-domain signal is:

[0137] y(n) = IFFT(Y(ω))

[0138] Figure 6 The detection results after time-reversed deconvolution interference suppression processing are given.

[0139] Step 4: Target CA-CFAR detection

[0140] Perform cell-averaging constant false alarm rate (CA-CFAR) detection on the output result in Step 3. Set M reference cells, estimate the interference noise power level within the reference window, and multiply by the threshold coefficient λ = P fa -1 / M -1 to obtain the threshold value T.

[0141] Step 5: Soft decision fusion detection based on signal-to-noise ratio weighting

[0142] Soft decision re-partitions the observation space through multiple threshold values, can make full use of the distance relationship between the detection statistic and the threshold, and improve the decision credibility. Each single base station of the present invention adopts a three-level decision threshold (T2 < T < T1), that is, the observation space and are respectively further divided into 2 parts (partitioned according to the false alarm probability and detection probability required by the system), where satisfies satisfies The soft decision rule is:

[0143]

[0144] After the fusion center receives the decision data of each base station, the final decision is realized by estimating the detection statistic and the detection threshold. Therefore, the fusion detection algorithm of the present invention mainly includes two parts: local decision of a single base station and fusion decision of the fusion center. The specific steps are as follows:

[0145] (1) Calculate the three-level threshold values

[0146] According to the false alarm probability, received signal processing signal-to-noise ratio, and reference cells of each base station, obtain the three-level threshold coefficients, that is:

[0147] λ = P fa -1 / M -1

[0148] λ1 = 2 -1 / M (η + P fa -1 / M ) - (1 + η), η = 10 0.1·SNR

[0149] λ2 = ((1 + P fa ) / 2) -1 / M - 1

[0150] Among them, the signal-to-noise ratio SNR is obtained by the MMSE algorithm in Step 1. Each base station multiplies the background noise power spectrum value estimated by CA-CFAR by the coefficients λ, λ1, and λ2 to obtain the three-level threshold value.

[0151] (2) Single base station local decision

[0152] Each single base station makes a decision on the target according to the soft decision rule, that is

[0153]

[0154] And transmits the corresponding decision data and the signal-to-noise ratio information of the received signal processing to the fusion center.

[0155] (3) Calculate the detection statistic and threshold value of the fusion center

[0156] The detection statistic of the fusion center is the weighting of the decision values of each base station. By introducing the threshold coefficients λ3 and λ4, the data corresponding to the decision values of each base station are estimated, that is, λ3 = αλ1, λ4 = αλ2, where α is a given parameter and can be taken as 0.1. Then the data corresponding to the decision value can be estimated as:

[0157]

[0158] The detection statistic of the fusion center can be estimated as where N is the number of base stations, and β i (k) is the weight of the kth decision of the ith base station.

[0159] At each fusion decision, the fusion center updates the weights of each base station, that is:

[0160]

[0161] where η i (k) represents the signal-to-noise ratio of the kth decision of the ith base station, β i (0) = 1, and This makes the contribution value of the base station user with a high signal-to-noise ratio in the final output value of the fusion center large and the weight heavy, while the decision-making role of the base station user with a low signal-to-noise ratio becomes weak and the role played in the fusion process is small.

[0162] The decision threshold of the fusion center can be estimated as: where T i is the threshold obtained by CFAR detection at each base station.

[0163] (4) Detection decision at the fusion center

[0164] The fusion center makes a final decision using the calculated detection statistic and the detection threshold.

[0165] Figure 7 The soft decision fusion detection curve is given. The false alarm probabilities required by each base station and the fusion center are all set to 0.001. It can be seen from the figure that compared with a single base station, the detection performance of the soft decision at the fusion center has been greatly improved.

[0166] The present invention also provides an embodiment of a system applicable to the above multi-base buoy fusion detection method based on complex interference. The system includes:

[0167] A preprocessing module for preprocessing the received signal to obtain the threshold value for target detection; the preprocessing includes: MMSE spectral estimation noise reduction, frequency-domain matched filtering based on Doppler template matching, time reversal deconvolution processing, and target cell average constant false alarm rate detection;

[0168] A fusion detection module for performing fusion detection based on soft decision weighted by signal-to-noise ratio in combination with the threshold value.

[0169] Each single base station of the present invention effectively suppresses the noise component while retaining the signal information through the MMSE spectral estimation algorithm, improving the input signal-to-noise ratio during detection;

[0170] Each single base station of the present invention adopts the Doppler template matching technology to effectively suppress the direct wave interference signal by designing a group of matched filters with different Doppler frequency offsets;

[0171] Each single base station of the present invention adopts the time reversal deconvolution interference suppression technology to compress and spatially focus the signal, effectively suppressing the multipath interference signal;

[0172] Each single base station of the present invention uses a three-level decision threshold to re-divide the observation space into four non-overlapping parts, and fully utilizes the distance relationship between the detection statistic and the threshold to achieve soft decision detection;

[0173] The fusion center of the present invention adopts the soft decision fusion detection technology based on signal-to-noise ratio weighting. By non-quantitatively estimating the data corresponding to the decisions of each base station and using the signal-to-noise ratio weighting method to represent the contribution values of each base station user in the fusion decision, the detection performance is improved.

[0174] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. Although the present invention has been described in detail with reference to the above embodiments, those of ordinary skill in the art should understand that: modifications or equivalent substitutions can still be made to the specific implementation manners of the present invention, and any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the protection scope of the claims of the present invention.

Claims

1. A multi-base buoy fusion detection method under complex interference, characterized in that, including: preprocessing the received signal to obtain the threshold value for target detection; combining the threshold value and performing fusion detection based on soft decision with SNR weighting; the preprocessing includes: MMSE spectral estimation for noise reduction, frequency-domain matched filtering based on Doppler template matching, time-reversal deconvolution processing, and target cell-averaged constant false alarm rate detection; the combining the threshold value and performing fusion detection based on soft decision with SNR weighting includes: Each single base station adopts a three-level decision threshold, T2 < T < T1, that is, the observation space and are respectively further divided into two parts, where satisfies satisfies The soft decision rule is: after the fusion center receives the decision data from each base station, the final decision is made by estimating the detection statistic and the detection threshold; the combining the threshold value and performing fusion detection based on soft decision with SNR weighting specifically includes: obtaining the three-level threshold coefficients according to the false alarm probability, received signal processing SNR, and reference cell of each base station; each single base station makes a decision on the target according to the soft decision rule; calculating the detection statistic and the threshold value of the fusion center; the fusion center makes the final decision using the calculated detection statistic and the detection threshold; the three-level threshold coefficients are respectively: λ = P fa -1 / M -1 λ1 = 2 -1 / M (η + P fa -1 / M ) - (1 + η), η = 10 0.1·SNR λ2 = ((1 + P fa ) / 2) -1 / M - 1 each base multiplies the interference noise power level estimated by the cell-averaged constant false alarm rate by the coefficients λ, λ1, and λ2 respectively to obtain the three-level threshold values; each single base station makes a decision on the target according to the soft decision rule, that is and transmits the corresponding decision data and the received signal processing SNR information to the fusion center; the calculating the detection statistic and the threshold value of the fusion center includes: the detection statistic of the fusion center is the weighting of the decision values of each base station. By introducing the threshold coefficients λ3 and λ4, the data corresponding to the decision values of each base station is estimated, that is, λ3 = αλ1, λ4 = αλ2, where α is a given parameter, and the data corresponding to the decision value can be estimated as: The detection statistic of the fusion center can be estimated as where N is the number of base stations, and β i (k) is the weight of the k-th decision of the i-th base station; in each fusion decision, the fusion center updates the weights of each base station, that is: where η i (k) represents the signal-to-noise ratio of the k-th decision of the i-th base station, β i (0) = 1, and The decision threshold of the fusion center can be estimated as: where T i is the threshold obtained by CFAR detection at each base station.

2. The method according to claim 1, wherein performing MMSE spectral estimation for noise reduction on the received signal includes: subjecting the received signal to pre-emphasis, frame division, and windowing processing; obtaining the energy spectrum after performing short-time Fourier transform on the time-domain framed signal: estimating the background noise power spectrum, posterior SNR, and prior SNR; calculating the gain function and estimating the signal spectrum; phase recovery and inverse Fourier transform to obtain the time-domain signal estimation; combining each segment of the time-domain signal and outputting the signal after noise reduction processing and the equivalent SNR.

3. The method according to claim 2, wherein the energy spectrum is: P(k) = |X(k)| = |FFT(x(n))| where x(n) = s(n) + d(n) is the received signal, which belongs to the time-domain framed signal, s(n) is the pure signal to be detected, and d(n) is the noise; the estimating the background noise power spectrum, posterior SNR, and prior SNR includes: Select the first few frames without signal arrival as background noise and estimate its power spectrum λ d (k) = E{|D(k)|}, and estimate the posterior signal-to-noise ratio γ k and the prior signal-to-noise ratio ξ k : Among them, λ d (k) is the mean value of the noise variance when no signal arrives in several segments, m is the signal segmentation sequence, and a is a coefficient.

4. The method according to claim 3, characterized in that, the calculating the gain function and estimating the signal spectrum includes: the expression of the gain function is: where G k is the gain function, Γ(·) is the Gamma function, and I0(·) and I1(·) denote the zero-order and first-order Bessel functions, respectively; the spectrum estimation of each segment of the signal is obtained by the following formula: Among them, is the spectral estimation of each segment of the signal; the time-domain signal estimation is: where θ k is obtained from the noisy signal; the equivalent SNR of the signal after noise reduction processing is: where M is the number of signal segments.

5. The method according to claim 4, wherein performing frequency-domain matched filtering based on Doppler template matching on the signal after noise reduction processing includes: performing matched filtering processing on the signal after noise reduction processing in the frequency domain respectively, and the output is: Among them, ξ i corresponds to different Doppler frequency offsets.

6. The method according to claim 5, characterized in that, performing time-reversal deconvolution processing on the output of the matched filter includes: performing time-reversal deconvolution processing on the output of the matched filter, that is T Y ($\omega$) = Y(-$\omega$) * Y($\omega$) Apply T Y (ω) Through the interference suppression gate to eliminate the main noise interference: Y(ω) = T Y (ω)·I(ω) Among them, After performing the inverse Fourier transform on Y(ω), the finally output time-domain signal is: y(n) = IFFT(Y(ω)) Perform target cell-averaged constant false alarm rate detection on y(n), including: Set M reference units, estimate the interference noise power level within the reference window, and multiply by the threshold coefficient λ = P fa -1 / M -1 to obtain the threshold value T.

7. A system applicable to the multi-base buoy fusion detection method based on complex interference according to any one of claims 1-6, characterized in that The system includes: A preprocessing module, which is used to preprocess the received signal to obtain the threshold value for target detection; the preprocessing includes: MMSE spectral estimation noise reduction, frequency-domain matched filtering based on Doppler template matching, time reversal deconvolution processing, and target cell-averaged constant false alarm rate detection; A fusion detection module, which is used to perform fusion detection based on soft decision-making weighted by signal-to-noise ratio in combination with the threshold value.

Images