Double-path fusion spectrum detection method, device, equipment and medium

Abstract

The invention relates to a double-path fusion spectrum detection method and device, equipment and a medium. The method comprises the following steps: respectively inputting a communication signal into a waveform reconstruction detection channel and a cyclostationary spectrum sensing channel, in the waveform reconstruction detection channel, carrying out signal reconstruction and reconstruction error calculation on the communication signal to obtain a judgment result so as to judge a spectrum occupation state, and in the cyclostationary spectrum sensing channel, carrying out signal reconstruction on the communication signal to obtain a spectrum occupation state; detection results are obtained through cyclic spectrum calculation and feature extraction of communication signals, so that signal features are detected, after the two detection results are fused, comparison is performed through an adaptive threshold, a spectrum detection result is output, and a more comprehensive signal feature analysis capability is provided.

Description

Technical Field

[0001] The present application relates to the field of wireless communication technology, and in particular to a spectrum detection method, apparatus, device, and medium for dual-path fusion. Background Art

[0002] With the rapid development of wireless communication technology, the scarcity of spectrum resources has become increasingly prominent. Traditional fixed spectrum allocation methods have resulted in a large amount of idle spectrum resources. The core of dynamic spectrum access technology lies in accurate and reliable spectrum sensing. Currently, mainstream spectrum sensing methods mainly include energy detection, matched filtering detection, and cyclostationary feature detection. Energy detection methods are simple to implement and have low computational complexity, but their performance degrades sharply in low signal-to-noise ratio environments and they are relatively sensitive to noise uncertainty. Although matched filtering detection is theoretically the optimal detection method, it requires prior knowledge of the characteristic information of the primary user signal, which has significant limitations in practical applications. Cyclic stationary feature detection utilizes the cyclostationary characteristics of communication signals for detection and has strong noise resistance, but it has high computational complexity and a long detection time.

[0003] In recent years, researchers have proposed a variety of improved solutions, such as detection methods based on compressed sensing and collaborative spectrum sensing. However, these methods still face the following challenges in practical applications: First, existing detection methods struggle to accurately extract and analyze signal features in complex electromagnetic environments, resulting in unstable detection performance. Second, detection thresholds are often set using fixed methods or empirical formulas, lacking the ability to adapt to environmental changes. Third, most methods fail to fully utilize the multidimensional characteristic information of the signal, limiting the potential for improvement in detection performance. Summary of the Invention

[0004] Based on this, it is necessary to provide a spectrum detection method, device, equipment and medium for dual-path fusion to address the above technical problems.

[0005] A dual-path fusion spectrum detection method, the method comprising:

[0006] receiving communication signals;

[0007] inputting the communication signal into a waveform reconstruction detection channel and a cyclostationary spectrum sensing channel respectively;

[0008] In the waveform reconstruction detection channel, a judgment result is obtained by performing signal reconstruction and reconstruction error calculation on the communication signal;

[0009] In the cyclostationary spectrum sensing channel, a detection result is obtained by performing cyclic spectrum calculation and feature extraction on the communication signal;

[0010] The decision result and the detection result are fused in a double-path manner, compared with an adaptive threshold, and a spectrum detection result is output.

[0011] In one of the embodiments, further comprising: constructing a Hankel matrix from the communication signal as:

[0012]

[0013] wherein L is the number of rows of the matrix, and r(t) represents the communication signal;

[0014] Performing singular value decomposition on the Hankel matrix as:

[0015] H = UΣV T

[0016] wherein U and V are orthogonal matrices, and Σ is a singular value matrix;

[0017] According to the size of the singular value after the decomposition, separating the noise signal in the communication signal to obtain a reconstruction matrix, and reconstructing the communication signal by minimizing the Frobenius norm between the Hankel matrix and the reconstruction matrix.

[0018] In one of the embodiments, further comprising: calculating a cyclic autocorrelation function of the communication signal as:

[0019] R x (t,τ) = E[r(t)r(t+τ)]

[0020] wherein τ is a delay parameter, and R x (t,τ) represents the cyclic autocorrelation function, and E[·] represents the mathematical expectation value;

[0021] Performing Fourier transform on the cyclic autocorrelation function to obtain a cyclic spectrum as:

[0022]

[0023] wherein S x (f,α) represents the cyclic spectrum, f represents the frequency, and α represents the cyclic frequency;

[0024] According to the cyclic spectrum, extracting the frequency component of the specified frequency to obtain a detection result.

[0025] In one of the embodiments, further comprising: fusing the decision result and the detection result in a double-path manner as:

[0026] S final = w css ·S css + w wrd ·S wrd

[0027] wherein S final represents the communication signal fusion result, S css represents the detection result of the cyclic spectrum sensing channel, S wrd represents the decision result of the waveform reconstruction detection channel, W css and W wrd respectively represent the path weight of the cyclic spectrum sensing channel and the waveform reconstruction detection channel; W css = γ css , W wrd = 1- γ css , γ css represents the applicability weight of the cyclic spectrum sensing channel; the applicability weight is realized according to the signal-to-noise ratio and the bit error rate.

[0028] In one of the embodiments, further comprising: constructing the adaptive threshold as:

[0029] Λ = σ 2 (1+ α· SNR)· Q 1 (Pfa)· H(y)

[0030] wherein Λ represents the adaptive threshold, σ 2 represents the signal power, α represents the environmental correction coefficient, Q(a, b) represents the Marcum Q function, Pfa represents the target false alarm probability, H(y) represents the channel noise compensation function, H(y) = 1+ β· exp(-y / y0), β represents the compensation intensity coefficient, y0 represents the reference fading coefficient, y represents the actual measured fading coefficient.

[0031] In one of the embodiments, the adjustment of the adaptive threshold comprises: a fast adjustment layer and a slow adjustment layer; further comprising: in the fast adjustment layer, the adaptive threshold is fast adjusted based on the short-time statistical characteristics of the local window, and is represented as:

[0032] Λ f = Λ f-1 + β f (t- Λ f-1 )· W(SNR)

[0033] wherein Λ f represents the adaptive threshold value after the fth update of the fast adjustment layer, β f represents the fast adjustment learning rate, T represents the real-time average value of the statistical quantity in the short-time window, and W(SNR) represents the weight function of the SNR correlation;

[0034] in the slow adjustment layer, the adaptive threshold is slowly adjusted based on the sparse optimization of the long-time statistical characteristics, and is represented as:

[0035] Λs s -1+β s (T s -Λ s -1)·F(PFa)

[0036] wherein, Λ s represents the adaptive threshold value after the s-th update of the slow adjustment layer, β s represents the slow adjustment learning rate, T s represents the arithmetic mean of the statistics in the long time window, and F(Pfa) represents an adjustment function related to the false alarm probability.

[0037] According to the fast adjustment of the adaptive threshold and the slow adjustment of the adaptive threshold, the final threshold is obtained as follows:

[0038] Λ final = ω·Λ f +(1-ω)·Λ s

[0039] wherein ω is a dynamic adjustment coefficient.

[0040] A dual-path fusion spectrum detection device, the device comprising:

[0041] a signal receiving module for receiving a communication signal;

[0042] a signal input module for inputting the communication signal into a waveform reconstruction detection channel and a cyclic stationary spectrum sensing channel, respectively;

[0043] a waveform reconstruction detection module for obtaining a decision result by signal reconstruction and reconstruction error calculation on the communication signal in the waveform reconstruction detection channel;

[0044] a cyclic stationary detection module for obtaining a detection result by cyclic spectrum calculation and feature extraction on the communication signal in the cyclic stationary spectrum sensing channel;

[0045] a spectrum detection module for comparing the decision result and the detection result after dual-path fusion with an adaptive threshold, and outputting a spectrum detection result.

[0046] In one embodiment, the waveform reconstruction detection module is further configured to construct a Hankel matrix from the communication signal as follows:

[0047]

[0048] wherein L is the number of rows of the matrix, and ·r(t) represents the communication signal.

[0049] singular value decomposition of the Hankel matrix is as follows: ​

[0050] H = UΣV T

[0051] wherein U and V are orthogonal matrices, and Σ is a singular value matrix;

[0052] According to the singular value size after the decomposition, a noise signal in the communication signal is separated to obtain a reconstruction matrix, and the communication signal is reconstructed by minimizing the Frobenius norm between the Hankel matrix and the reconstruction matrix.

[0053] A computer device comprises a memory and a processor, the memory stores a computer program, and the processor implements the following steps when executing the computer program:

[0054] receiving a communication signal;

[0055] inputting the communication signal into a waveform reconstruction detection channel and a cyclic stationary spectrum sensing channel, respectively;

[0056] in the waveform reconstruction detection channel, a decision result is obtained by signal reconstruction and reconstruction error calculation on the communication signal;

[0057] in the cyclic stationary spectrum sensing channel, a detection result is obtained by cyclic spectrum calculation and feature extraction on the communication signal;

[0058] after double-path fusion of the decision result and the detection result, the fusion result is compared with an adaptive threshold to output a spectrum detection result.

[0059] A computer readable storage medium stores a computer program, and the computer program is executed by a processor to implement the following steps:

[0060] receiving a communication signal;

[0061] inputting the communication signal into a waveform reconstruction detection channel and a cyclic stationary spectrum sensing channel, respectively;

[0062] in the waveform reconstruction detection channel, a decision result is obtained by signal reconstruction and reconstruction error calculation on the communication signal;

[0063] in the cyclic stationary spectrum sensing channel, a detection result is obtained by cyclic spectrum calculation and feature extraction on the communication signal;

[0064] after double-path fusion of the decision result and the detection result, the fusion result is compared with an adaptive threshold to output a spectrum detection result.

[0065] The double-path fusion spectrum detection method, device, equipment and medium described above input the communication signal into a waveform reconstruction detection channel and a cyclic spectrum sensing channel respectively, in the waveform reconstruction detection channel, the decision result is obtained by signal reconstruction and reconstruction error calculation on the communication signal to judge the spectrum occupation state, in the cyclic spectrum sensing channel, the detection result is obtained by cyclic spectrum calculation and feature extraction on the communication signal to detect the signal feature, after the fusion of the two detection results, the spectrum detection result is output by comparison through the adaptive threshold, and more comprehensive signal feature analysis capability is provided. BRIEF DESCRIPTION OF DRAWINGS

[0066] Figure 1 It is a flowchart of the double-path fusion spectrum detection method in one embodiment.

[0067] Figure 2 It is a structural block diagram of the double-path fusion spectrum detection device in one embodiment.

[0068] Figure 3 It is an internal structure diagram of the computer equipment in one embodiment. DETAILED DESCRIPTION

[0069] In order to make the purpose, technical scheme and advantages of the present application clearer, the present application is further described in detail below in combination with the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and do not limit the present application.

[0070] In one embodiment, as shown in Figure 1 a double-path fusion spectrum detection method is provided, including the following steps:

[0071] Step 102, receiving a communication signal.

[0072] The present embodiment can be applied to dynamic spectrum access, spectrum resource management and other scenarios in 5G / 6G communication system, therefore, according to the different application scenarios, the communication signal is received from the communication system for spectrum detection.

[0073] Step 104, inputting the communication signal into a waveform reconstruction detection channel and a cyclic spectrum sensing channel respectively.

[0074] In this step, the waveform reconstruction detection (WRD) channel judges the spectrum occupation state by reconstructing the received signal based on the signal subspace theory. The cyclic spectrum sensing (CSS) channel utilizes the cyclic spectrum characteristics of the signal to realize detection by analyzing the cyclic spectrum correlation function of the signal.

[0075] Step 106, in the waveform reconstruction detection channel, the decision result is obtained by signal reconstruction and reconstruction error calculation on the communication signal.

[0076] Step 108, in the cyclic stationary spectrum sensing channel, a detection result is obtained by performing a cyclic spectrum calculation and feature extraction on the communication signal.

[0077] Step 110, after the decision result and the detection result are fused in a double path, the fused result is compared with an adaptive threshold, and a spectrum detection result is output.

[0078] In this step, the detection result can update the environmental parameter, so that the adaptive threshold is updated using the environmental parameter, thereby realizing adaptive updating of the threshold and forming a logical closed loop.

[0079] In the spectrum detection method of the above double-path fusion, the communication signal is respectively input into a waveform reconstruction detection channel and a cyclic stationary spectrum sensing channel. In the waveform reconstruction detection channel, a decision result is obtained by performing signal reconstruction and reconstruction error calculation on the communication signal, so as to judge the spectrum occupation state. In the cyclic stationary spectrum sensing channel, a detection result is obtained by performing a cyclic spectrum calculation and feature extraction on the communication signal, so as to detect the signal feature. After the two detection results are fused, the fused result is compared with an adaptive threshold, and a spectrum detection result is output, thereby providing more comprehensive signal feature analysis capability.

[0080] In one of the embodiments, the initially received communication signal also needs to be preprocessed. A high-precision sampling strategy and a weighted filter are adopted to significantly reduce the noise in the communication signal and improve the quality of the communication signal, thereby providing an optimized input communication signal for the subsequent processing module. The communication signal is represented as:

[0081] r(t) = s(t) + n(t)

[0082] r(t) represents the communication signal, s(t) represents the ideal target signal, and n(t) represents the additive white Gaussian noise.

[0083] In one of the embodiments, in the waveform reconstruction detection channel, a Hankel matrix is constructed from the communication signal as follows:

[0084]

[0085] wherein L is the number of rows of the matrix, and r(t) represents the communication signal; the Hankel matrix is singular value decomposed as follows:

[0086] H = UΣV T

[0087] wherein U and V are orthogonal matrices, and Σ is a singular value matrix; according to the size of the singular value after the decomposition, the noise signal in the communication signal is separated to obtain a reconstruction matrix, and the communication signal is reconstructed by minimizing the Frobenius norm between the Hankel matrix and the reconstruction matrix.

[0088] In one embodiment, the cyclic autocorrelation function of the communication signal is calculated in the cyclic spectrum sensing channel as follows:

[0089] R x (t,τ)=E[r(t)r(t+τ)]

[0090] where τ is a delay parameter, R x (t,τ) represents the cyclic autocorrelation function, and E[·] represents the mathematical expectation value; the Fourier transform of the cyclic autocorrelation function is performed to obtain the cyclic spectrum as follows:

[0091]

[0092] where S x (f,α) represents the cyclic spectrum, f represents the frequency, and α represents the cyclic frequency; the frequency component of a specified frequency is extracted from the cyclic spectrum to obtain the detection result.

[0093] In one embodiment, the decision result and the detection result are fused in a double-path manner as follows:

[0094] S final =w css ·S css +w wrd ·S wrd

[0095] where S final represents the communication signal fusion result, S css represents the detection result of the cyclic spectrum sensing channel, S wrd represents the decision result of the waveform reconstruction detection channel, W css and W wrd represent the path weights of the cyclic spectrum sensing channel and the waveform reconstruction detection channel, respectively; W css =γ css , W wrd =1-γ css , and γ css represents the applicability weight of the cyclic spectrum sensing channel; the applicability weight is optimized and implemented according to the signal-to-noise ratio and the bit error rate.

[0096] In this embodiment, the applicability weight can be defined as follows:

[0097] γ css =f(SNR,BER,α)∈[0,1]

[0098] γ css →1: high SNR + low BER + significant cyclic frequency → complete trust in the channel; γ css→ 0: low SNR + high BER + no cyclic property → drop the channel.

[0099] In one embodiment, the step of constructing the adaptive threshold comprises:

[0100] The adaptive threshold is constructed as:

[0101] Λ = σ 2 (1 + a · SNR) · Q 1 (Pfa) · H(y)

[0102] Wherein, Λ represents the adaptive threshold, σ 2 is the signal power, a represents the environmental correction coefficient, Q(a, b) represents the Marcum Q function, Pfa represents the target false alarm probability, H(y) represents the channel noise compensation function, H(y) = 1 + β · exp(-y / y0), β represents the compensation intensity coefficient, y0 represents the reference fading coefficient, y represents the actual measured fading coefficient.

[0103] For the actual measured fading coefficient, it is specifically defined as:

[0104] y = 10log 10 (|h|2 / E[|h|2])

[0105] Wherein, h is the channel response, y <-10dB represents the deep fading state, -10dB ≤ y <-3dB represents the moderate fading, y ≥ -3dB represents the light fading or no fading, which directly reflects the fading degree of the current channel and is used to adjust the detection threshold to compensate for the performance loss caused by fading.

[0106] The channel noise compensation function: the introduction of H(y) enables the channel noise compensation degree to have the ability of adaptive adjustment threshold value, improving the detection performance of the system in deep fading environment.

[0107] Specifically, the Marcum Q function is defined as:

[0108]

[0109] In this embodiment, the WRD-CSS dual-path structure is used to realize precise adaptation in multiple scenes, which needs to pass through three core links of dynamic feature fusion, scene awareness weight distribution and lightweight deployment optimization. The specific implementation method is as follows: input image (WRD and CSS dual-path feature extraction) → scene analysis (classifier + feature statistic calculation) → dynamic fusion (according to a weighted or gate fusion) → output result (optimized features adapted to the current scene).

[0110] In one embodiment, the adjustment of the adaptive threshold comprises: a fast adjustment layer and a slow adjustment layer; the step of adjusting the adaptive threshold comprises:

[0111] In the fast adjustment layer, the adaptive threshold is fast adjusted based on the short-time statistics of the local window, denoted as:

[0112] Λ f =Λ f-1 +β f (T-Λ f-1 )·W(SNR)

[0113] wherein, Λ f represents the adaptive threshold value of the fast adjustment layer after the fth update, β f represents the fast adjustment learning rate, T represents the real-time average value of the statistics in the short-time window, and W(SNR) represents the weight function of the SNR correlation; in the slow adjustment layer, the adaptive threshold is slowly adjusted based on the long-time statistical characteristics for sparse optimization, denoted as:

[0114] Λ s =Λ s -1+β s (T s -Λ s -1)·F(Pfa)

[0115] wherein, Λ s represents the adaptive threshold value of the slow adjustment layer after the sth update, β s represents the slow adjustment learning rate, T s represents the arithmetic average value of the statistics in the long-time window, and F(Pfa) represents the adjustment function related to the false alarm probability; according to the adaptive threshold fast adjustment and the adaptive threshold slow adjustment, the final threshold is obtained as:

[0116] Λ final =ω·Λ f +(1-ω)·Λ s

[0117] wherein, ω is a dynamic adjustment coefficient.

[0118] In one specific embodiment, a differentiated optimization strategy is adopted according to different SNR intervals to realize fine adjustment of the threshold:

[0119] 1. High SNR interval (SNR≥y0): environment correction coefficient: α=0.8, fast adjustment learning rate: β f =0.2, slow adjustment learning rate: β s =0.2, compensation intensity coefficient: β=0.4

[0120] 2. Medium SNR interval (y1≤SNR<y0): environment correction coefficient: α=0.6, fast adjustment learning rate: β f =0.4, slow adjustment learning rate: β s= 0.4 • compensation strength coefficient: β = 0.4

[0121] 3. Low SNR interval (SNR < y1): environment correction coefficient: a = 0.2, fast adjustment learning rate: β f = 0.5 • slow adjustment learning rate: β s = 0.6.

[0122] In the high SNR interval, the system response is more agile, while in the low SNR interval, the system strengthens the compensation strategy when the signal-to-noise ratio is low, thereby improving the reliability and adaptability of detection.

[0123] It should be understood that, although Figure 1 the steps in the flowchart of the method are displayed in sequence according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order limitation for the execution of these steps, and these steps can be executed in other orders. Moreover, Figure 1 at least part of the steps in the method can include multiple sub-steps or multiple stages, which are not necessarily executed at the same time, but can be executed at different times, and the execution order of these sub-steps or stages is not necessarily sequential, but can be alternately executed with at least part of other steps or sub-steps or stages of other steps.

[0124] In one embodiment, as shown in Figure 2 a dual-path fusion spectrum detection device is provided, comprising: a signal receiving module 202, a signal input module 204, a waveform reconstruction detection module 206, a cyclostationary detection module 208, and a spectrum detection module 210, wherein:

[0125] The signal receiving module 202 is configured to receive a communication signal.

[0126] The signal input module 204 is configured to input the communication signal into a waveform reconstruction detection channel and a cyclostationary spectrum sensing channel, respectively.

[0127] The waveform reconstruction detection module 206 is configured to obtain a decision result by performing signal reconstruction and reconstruction error calculation on the communication signal in the waveform reconstruction detection channel.

[0128] The cyclostationary detection module 208 is configured to obtain a detection result by performing cyclic spectrum calculation and feature extraction on the communication signal in the cyclostationary spectrum sensing channel.

[0129] The spectrum detection module 210 is configured to compare the decision result and the detection result after dual-path fusion with an adaptive threshold, and output a spectrum detection result.

[0130] In one embodiment, the waveform reconstruction detection module 206 is further configured to construct a Hankel matrix from the communication signal as:

[0131]

[0132] Where L is the number of rows in the matrix, r(t) represents the communication signal;

[0133] The singular value decomposition of the Hankel matrix is:

[0134] H=UΣV T

[0135] Where U and V are orthogonal matrices, and Σ is a singular value matrix;

[0136] The noise signal in the communication signal is separated according to the size of the decomposed singular value to obtain a reconstruction matrix, and the communication signal is reconstructed by minimizing the Frobenius norm between the Hankel matrix and the reconstruction matrix.

[0137] In one embodiment, the cyclostationary detection module 208 is further configured to calculate the cyclic autocorrelation function of the communication signal as:

[0138] R x (t,τ)=E[r(t)r(t+τ)]

[0139] Where τ is the delay parameter, R x (t,τ) represents the cyclic autocorrelation function, E[·] represents the mathematical expectation;

[0140] Performing Fourier transform on the cyclic autocorrelation function, the cyclic spectrum is obtained as follows:

[0141]

[0142] Among them, S x (f,α) represents the cyclic spectrum, f represents the frequency, and α represents the cyclic frequency;

[0143] According to the cyclic spectrum, the frequency component of the specified frequency is extracted to obtain a detection result.

[0144] In one embodiment, the spectrum detection module 210 is further configured to perform dual-path fusion of the decision result and the detection result to form:

[0145] S final =w css ·S css +w wrd ·S wrd

[0146] Among them, S finalS represents the fusion result of the communication signal, S css S represents the detection result of the cyclic spectrum sensing channel, S wrd W represents the decision result of the waveform reconstruction detection channel, W css and W wrd represent the path weights of the cyclic spectrum sensing channel and the waveform reconstruction detection channel respectively; W css = γ css , W wrd = 1- γ css , γ css represents the applicability weight of the cyclic spectrum sensing channel; the applicability weight is realized according to the signal-to-noise ratio and the bit error rate.

[0147] In one embodiment, the spectrum detection module 210 is further configured to construct an adaptive threshold as follows:

[0148] Λ = σ 2 (1+ α·SNR)·Q 1 (Pfa)·H(y)

[0149] wherein Λ represents the adaptive threshold, σ 2 represents the signal power, α represents the environmental correction coefficient, Q(a, b) represents the Marcum Q function, Pfa represents the target false alarm probability, H(y) represents the channel noise compensation function, H(y) = 1+ β·exp(-y / y0), β represents the compensation intensity coefficient, y0 represents the reference fading coefficient, y represents the actual measured fading coefficient.

[0150] In one embodiment, the adjustment of the adaptive threshold comprises a fast adjustment layer and a slow adjustment layer; the spectrum detection module 210 is further configured to perform adaptive threshold fast adjustment in the fast adjustment layer based on the short-time statistical characteristics of the local window, which is represented as follows:

[0151] Λ f = Λ f-1 + β f (T- Λ f-1 )·W(SNR)

[0152] wherein Λ f represents the adaptive threshold value after the fth update in the fast adjustment layer, β f represents the fast adjustment learning rate, T represents the real-time average value of the statistical quantity in the short-time window, and W(SNR) represents the weight function of the SNR correlation;

[0153] In the slow adjustment layer, adaptive threshold slow adjustment is performed based on the long-time statistical characteristics and sparse optimization, which is represented as follows:

[0154] Λ s = Λ s-1+β s (T s -Λ s -1)·F(Pfa)

[0155] Among them, Λ s represents the adaptive threshold value after the slow adjustment layer is updated for the sth time, β s Indicates slow adjustment of learning rate, T s represents the arithmetic mean of the statistic in the long-term window, and F(Pfa) represents the adjustment function related to the false alarm probability;

[0156] According to the adaptive threshold fast adjustment and the adaptive threshold slow adjustment, the final threshold is obtained as follows:

[0157] Λ final =ω·Λ f +(1-ω)·Λ s

[0158] Among them, ω is the dynamic adjustment coefficient.

[0159] For the specific limitations of the dual-path fusion spectrum detection device, please refer to the limitations of the dual-path fusion spectrum detection method above and will not be repeated here. The various modules in the above-mentioned dual-path fusion spectrum detection device can be implemented in whole or in part through software, hardware, or a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in hardware form, or can be stored in the memory of the computer device in software form, so that the processor can call and execute the operations corresponding to the above modules.

[0160] In one embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as follows: Figure 3 As shown. The computer device includes a processor, memory, network interface, display screen and input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, a dual-path fusion spectrum detection method is implemented. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or a key, trackball or touchpad provided on the computer device housing, or an external keyboard, touchpad or mouse.

[0161] Those skilled in the art will understand that Figure 3The structure shown in the figure is only a block diagram of part of the structure related to the scheme of the present application, and does not constitute a limitation on the computer device to which the scheme of the present application is applied. The specific computer device can include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0162] In one embodiment, a computer device is provided, comprising a memory and a processor, the memory storing a computer program, and the processor implementing the steps of the method in the above embodiments when executing the computer program.

[0163] In one embodiment, a computer readable storage medium is provided, storing a computer program, and the computer program is executed by a processor to implement the steps of the method in the above embodiments.

[0164] A person of ordinary skill in the art can understand that all or part of the processes in the above embodiments can be completed by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer readable storage medium and can include the processes of the above embodiments when executed. Any reference to memory, storage, database or other medium used in the embodiments provided by the present application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration but not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0165] The technical features of the above embodiments can be combined in any manner. In order to make the description concise, not all possible combinations of the technical features in the above embodiments are described, but as long as the combinations of the technical features do not exist, they should be considered as the scope of the present application.

[0166] The above-described embodiments are merely illustrative of several embodiments of the present application, which are described in more detail and in a specific and detailed manner, but should not be construed as limiting the scope of the patent. It should be noted that for those skilled in the art, several modifications and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application should be subject to the appended claims.

Claims

1. A spectrum detection method with dual-path fusion, characterized in that: The method comprises: receiving communication signals; inputting the communication signal into a waveform reconstruction detection channel and a cyclostationary spectrum sensing channel respectively; In the waveform reconstruction detection channel, a judgment result is obtained by performing signal reconstruction and reconstruction error calculation on the communication signal; In the cyclostationary spectrum sensing channel, a detection result is obtained by performing cyclic spectrum calculation and feature extraction on the communication signal; After dual-path fusion of the decision result and the detection result, the result is compared with an adaptive threshold and a spectrum detection result is output.

2. The method according to claim 1, characterized in that Obtaining a judgment result by performing signal reconstruction and reconstruction error calculation on the communication signal, including: The Hankel matrix constructed from the communication signal is: Where L is the number of rows in the matrix, and r(t) represents the communication signal; The singular value decomposition of the Hankel matrix is: H=UΣV T Where U and V are orthogonal matrices, and Σ is a singular value matrix; The noise signal in the communication signal is separated according to the size of the decomposed singular value to obtain a reconstruction matrix, and the communication signal is reconstructed by minimizing the Frobenius norm between the Hankel matrix and the reconstruction matrix.

3. The method according to claim 1, characterized in that Obtaining a detection result by calculating a cyclic spectrum and extracting features of the communication signal, including: The cyclic autocorrelation function of the communication signal is calculated as: R x (t,τ)=E[r(t)r(t+τ)] Where τ is the delay parameter, R x (t,τ) represents the cyclic autocorrelation function, E[·] represents the mathematical expectation; Performing Fourier transform on the cyclic autocorrelation function, the cyclic spectrum is obtained as follows: Among them, S x (f,α) represents the cyclic spectrum, f represents the frequency, and α represents the cyclic frequency; According to the cyclic spectrum, the frequency component of the specified frequency is extracted to obtain a detection result.

4. The method according to claim 1, wherein Performing dual-path fusion on the judgment result and the detection result, including: The judgment result and the detection result are dual-path fused to form: S final =w css ·S css +w wrd ·S wrd Among them, S final Indicates the communication signal fusion result, S css represents the detection result of the cyclostationary spectrum sensing channel, S wrd Represents the decision result of the waveform reconstruction detection channel, W css and W wrd They represent the path weights of the cyclostationary spectrum sensing channel and the waveform reconstruction detection channel respectively; W css =γ css ,W wrd =1-γ css , γ css represents the suitability weight of the cyclostationary spectrum sensing channel; the suitability weight is optimized according to the signal-to-noise ratio and the bit error rate.

5. The method according to any one of claims 1 to 4, characterized in that The steps to construct an adaptive threshold include: Construct the adaptive threshold as: L=s 2 (1+α·SNR)·Q 1 (Pfa)·H(y) Where Λ represents the adaptive threshold, σ 2 is the signal power, α is the environmental correction coefficient, Q(a,b) is the Marcum Q function, Pfa is the target false alarm probability, H(y) is the channel noise compensation function, H(y) = 1 + β exp(-y / y0), β is the compensation intensity coefficient, y0 is the reference fading coefficient, and y is the actual measured fading coefficient.

6. The method according to claim 5, characterized in that The adjustment of the adaptive threshold includes: a fast adjustment layer and a slow adjustment layer; The step of adjusting the adaptive threshold includes: In the fast adjustment layer, the adaptive threshold is quickly adjusted based on the short-term statistical characteristics of the local window, which is expressed as: L f =L f-1 +b f (T-Λ f-1 )·W(SNR) Among them, Λ f represents the adaptive threshold value after the fast adjustment layer is updated for the fth time, β f Indicates rapid adjustment of the learning rate, T represents the real-time average of the statistics in a short window, and W(SNR) represents the weight function of SNR correlation; In the slow adjustment layer, sparse optimization is performed based on long-term statistical characteristics to perform adaptive threshold slow adjustment, which is expressed as: L s =L s -1+b s (T s -L s -1)·F(Pfa) Among them, Λ s represents the adaptive threshold value after the slow adjustment layer is updated for the sth time, β s Indicates slow adjustment of learning rate, T s represents the arithmetic mean of the statistic in the long-term window, and F(Pfa) represents the adjustment function related to the false alarm probability; According to the adaptive threshold fast adjustment and the adaptive threshold slow adjustment, the final threshold is obtained as follows: L final =ω·L f +(1-ω)·L s Among them, ω is the dynamic adjustment coefficient.

7. A dual-path fusion spectrum detection device, characterized in that: The device comprises: A signal receiving module, configured to receive communication signals; A signal input module, configured to input the communication signal into a waveform reconstruction detection channel and a cyclostationary spectrum sensing channel respectively; A waveform reconstruction detection module, configured to obtain a judgment result by performing signal reconstruction and reconstruction error calculation on the communication signal in the waveform reconstruction detection channel; a cyclostationary detection module, configured to obtain a detection result by performing cyclic spectrum calculation and feature extraction on the communication signal in the cyclostationary spectrum sensing channel; The spectrum detection module is used to perform dual-path fusion on the decision result and the detection result, compare them with the adaptive threshold, and output the spectrum detection result.

8. The device according to claim 7, characterized in that The waveform reconstruction detection module is further configured to construct a Hankel matrix from the communication signal: Where L is the number of rows in the matrix, r(t) represents the communication signal; The singular value decomposition of the Hankel matrix is: H=UΣV T Where U and V are orthogonal matrices, and Σ is a singular value matrix; The noise signal in the communication signal is separated according to the size of the decomposed singular value to obtain a reconstruction matrix, and the communication signal is reconstructed by minimizing the Frobenius norm between the Hankel matrix and the reconstruction matrix.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images