Blind signal time domain detection method based on autocorrelation three-window judgment

Abstract

The invention discloses a blind signal time domain detection method based on self-correlation three-window judgment, belongs to the technical field of communication, and aims to provide a blind signal time domain detection method based on self-correlation three-window judgment under the condition of low signal-to-noise ratio. The blind signal time domain detection method based on the self-correlation three-window judgment is suitable for communication signals of various modulation types and radar signals with low duty ratio. The technical key points of the invention are as follows: the method disclosed by the invention can realize self-adaption of the detection judgment threshold. For signals with different amplitudes and different durations, a self-adaptive threshold can be set by adopting the idea of combining a relative threshold and an absolute threshold according to the size of a correlation value of operation and according to the types and characteristics of the signals in a window; the problems of signal detection segmentation caused by correlation value fluctuation of amplitude modulation signals, signal missing detection caused by received data amplitude difference of aliasing signals, signal detection false alarm caused by high pulse correlation peak values of radar signals and the like are solved, and finally the detection accuracy of the algorithm on various types of signals is improved. A data set detection result shows that when the signal-to-noise ratio is larger than 8dB, the detection correct probability is 95% or above.

Description

Technical Field

[0001] The present invention belongs to the field of communication technologies, and particularly relates to a blind signal time-domain detection method based on autocorrelation three-window decision applicable to a blind signal reconnaissance receiver in this field. Background Art

[0002] With the rapid development of modern social communication technologies, various wireless communication methods and devices are constantly increasing. Each radio frequency device will occupy a certain spectrum, and the continuous increase of these devices makes many radio devices concentrated in some frequency bands. The blind signal detection technology refers to the detection of signals under the condition that the prior information of the signals is uncertain. The research on the methods of this technology is of great significance for the improvement of non-cooperative communication and cognitive radio systems. Blind signal detection methods can generally be divided into two categories: time-domain detection and frequency-domain detection. Based on the characteristics of simple calculation and high real-time performance of time-domain detection, the present invention focuses on the research of blind signal time-domain detection algorithms.

[0003] Traditional time-domain detections mainly include energy detection method, double-window sliding detection method, correlation detection method, etc. The energy detection method is the most basic and classic signal detection method, and its principle is to detect the existence of signals by relying on the energy or power of the signals. When performing blind signal detection, the energy detection method does not require any prior information, has a low algorithm complexity, and is relatively easy to implement. The decision result is obtained by comparing the energy calculated from the received signal with a threshold. The decision threshold value of the algorithm is closely related to the power of the noise. When using the energy detection algorithm when the noise variance is affected by the external environment, the algorithm performance will decline due to the difficulty in determining the decision threshold. The double-window sliding detection method is an improved algorithm based on the energy detection method. The energy ratio of two adjacent windows is calculated respectively, and the sum of the two is used as the decision function. When one of the two windows is noise and the other is a signal, the energy difference between the two windows is the largest, that is, the decision function appears as a peak, and the position of the signal is determined according to the peak position of the decision function. However, this algorithm depends on the energy difference between adjacent windows. Even if there is no noise, there will still be an energy difference between single signals and overlapping signals in adjacent windows, which may lead to misjudgment of the signal position. The correlation detection algorithm uses the essential difference that noise has no correlation while signals have correlation, and detects the position of the signal according to whether the correlation value of the received data exceeds a set threshold, and has good anti-noise performance and real-time performance. However, how to determine the optimal detection threshold is the key research content therein. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to provide a blind signal time-domain detection method based on autocorrelation three-window decision applicable to communication signals of various modulation types and low-duty-cycle radar signals under low signal-to-noise ratio conditions, which has the advantages of good real-time performance and strong adaptability.

[0005] The present invention adopts the following technical solutions:

[0006] A blind signal time-domain detection method based on autocorrelation three-window decision, the improvement of which includes the following steps:

[0007] Step 1: Perform a correlation operation on the received discrete data x(n) to obtain the autocorrelation function R n (n), where n represents the sampling point:

[0008] The expression of the discrete data x(n) is:

[0009]

[0010] In the above formula, is the actual signal; A is the signal amplitude; is the initial phase; T s is the sampling interval; f n is the carrier frequency, and w(n) is the zero-mean additive white Gaussian noise after sampling;

[0011] The calculation formula for the data autocorrelation function R n (n) is:

[0012]

[0013] In the above formula, N is the correlation window length, * represents taking the conjugate of the data, and i sam represents the discrete sampling point;

[0014] Step 2: According to the length L of x(n), set three windows at the beginning, middle, and end of the data for data type decision within the windows. The value of the decision window label m is 1, 2, 3, and the decision window length is defined as N Det , and Calculate the average power and the maximum autocorrelation value R m within the three windows respectively, and obtain the ratio coefficient k of the two:

[0015]

[0016] Step 3: Set the ratio coefficient decision index k ind = 1.8. If k ≥ k ind , there is amplitude-modulated data within the window, and calculate the absolute threshold value th abs_1 of the received data;

[0017] Step 4: If k < k ind , calculate the relative threshold th rel within the decision window m respectively. According to the specific situation where the autocorrelation value within the window exceeds the relative threshold, calculate the absolute threshold value th abs_2or th abs_3 ;

[0018] Step 5, compare the absolute threshold values calculated in Step 3 and Step 4 with the data autocorrelation function R n (n), and set both the sliding window length and the sliding interval to L win . If there is an autocorrelation value greater than the absolute threshold within the sliding window, it is determined that the entire sliding window contains signals. After the sliding window traverses all the received data, the signal positions are output.

[0019] Further, Step 3 includes:

[0020] Step 31, determine the window numbers j where amplitude-modulated data exists, j ∈ m, and calculate the maximum autocorrelation value R j max ;

[0021] Step 32, take the minimum value R j max and multiply it by the threshold coefficient α lower than the signal to calculate the absolute threshold value th j ' max of the received data: abs_1 :

[0022] th abs_1 = R j ' max *α.

[0023] Further, Step 4 includes:

[0024] Step 41, calculate the maximum autocorrelation value R m within the decision window m respectively, multiply R m by the relative threshold decision coefficient ε to calculate the relative threshold th rel of the received data:

[0025] th rel = R m *ε

[0026] Step 42, determine whether the data type within the window is noise or signal;

[0027] Step 43, if the data types determined by the three decision windows are all noise, calculate the maximum autocorrelation value R i max within the decision window m respectively, i ∈ m, take the maximum value R i max and multiply it by the threshold coefficient λ higher than the noise to calculate the absolute threshold value th i ' max of the received data: abs_3 :

[0028] th abs_3 = R i ' max * λ

[0029] Step 44, if the data types determined by the three decision windows are not all noise, calculate the absolute threshold value th according to different data types abs_2 :

[0030] If all three windows are signals, calculate the maximum autocorrelation value R k max , k ∈ m, and take the minimum value R k max of R k ' max and multiply it by the threshold coefficient α below the signal to calculate the absolute threshold value th of the received data abs_2 :

[0031] th abs_2 = R k ' max * α

[0032] If not all three windows are signals, select the minimum value from the maximum autocorrelation values of all windows with signals and define it as Rmini1. Then find the maximum autocorrelation value in the received data and define it as RmaxAll. If the ratio of RmaxAll to Rmini1 is greater than the decision coefficient μ, the absolute threshold value th abs_2 is:

[0033] th abs_2 = R p ' max * λ

[0034] In the above formula, p is the serial number of the window with noise, p ∈ m, and calculate the maximum autocorrelation value R p max , take the minimum value R p max of R p ' max , and λ is the threshold coefficient above the noise;

[0035] If the ratio of RmaxAll to Rmini1 is less than the decision coefficient μ, the absolute threshold value th abs_2 is:

[0036] th abs_2 = R q ' max * α

[0037] In the above formula, q is the serial number of the window with signals, q ∈ m, and calculate the maximum autocorrelation value R of the data in the q windowq max , take the minimum value R of R q max q ' max , where α is a threshold coefficient lower than the signal.

[0038] Furthermore, α is taken as 0.5, ε is taken as 0.2 - 0.5, λ is taken as 1.5, and μ is taken as 1.8.

[0039] A blind signal time - domain detection method system based on autocorrelation three - window decision. This system has program modules corresponding to the steps of the above - mentioned technical solution, and when running, it executes the steps in the above - mentioned blind signal time - domain detection method based on autocorrelation three - window decision.

[0040] A computer - readable storage medium stores a computer program, and the computer program is configured to implement the steps of the above - mentioned blind signal time - domain detection method based on autocorrelation three - window decision when called by a processor.

[0041] The beneficial effects of the present invention are:

[0042] The technical solution proposed by the present invention is a blind signal time - domain detection method based on autocorrelation three - window decision for communication signals of various modulation types and low - duty - cycle radar signals under low signal - to - noise ratio conditions. Key points of the present invention: In the method disclosed by the present invention, the detection decision threshold can be adaptive. For signals with different amplitudes and different durations, according to the magnitude of the calculated correlation value, the idea of combining relative threshold and absolute threshold can be adopted. According to the signal type and characteristics within the window, an adaptive threshold is set, which improves problems such as signal detection segmentation caused by the fluctuation of the correlation value of amplitude - modulated signals, signal missed detection caused by the amplitude difference of received data for mixed signals, and signal detection false alarm caused by high pulse correlation peaks for radar signals. Finally, the detection accuracy of the algorithm for multiple types of signals is improved.

[0043] The method disclosed by the present invention has strong anti - noise performance. Due to the essential difference that noise has no correlation and signals have correlation, by performing autocorrelation operation on the received data, the positions of signals and noise can be distinguished according to the correlation amplitude difference, and the signal position can be detected under low signal - to - noise ratio conditions such as 0 dB.

[0044] The method disclosed by the present invention has a small amount of computation and can efficiently process received data. When making a data type decision, three windows at the beginning, middle, and end of the received data are selected for decision detection. Too few windows may affect the accuracy of the overall type decision of the received data, and too many windows will increase the amount of computation. Therefore, in order to balance the relationship between the two, three - window decision is selected, which avoids performing threshold decision operations on all data, improves the operation efficiency, and has good real - time performance.​

[0045] The method disclosed by the present invention has no signal type restriction on the input data. It is applicable to communication signals of various modulation types and radar signals with different duty cycles. Since the algorithm sets different detection processes according to the judgment results of the data types, and the correlation value of the amplitude modulation signal changes with the signal amplitude, it is impossible to make a unified judgment with non-amplitude modulation signals such as frequency modulation and phase modulation. Moreover, since the amplitude difference between the pulse part and the rest period of the radar signal equates the pulse part to amplitude modulation data and the rest period part to noise data, corresponding process detection is performed according to the data type.

[0046] The detection decision threshold of the method disclosed by the present invention can be adaptive. For signals with different amplitudes and different durations, the combination idea of relative threshold and absolute threshold can be adopted according to the size of the calculated correlation value. According to the signal type and characteristics within the window, an adaptive threshold is set, which improves problems such as signal detection segmentation caused by the fluctuation of the correlation value of the amplitude modulation signal, signal missing detection caused by the amplitude difference of received data for overlapping signals, and signal detection false alarm caused by the high pulse correlation peak of radar signals. Finally, the detection accuracy of the algorithm for multiple types of signals is improved. It has been verified that the detection result of the data set shows that when the signal-to-noise ratio is greater than 8 dB, the detection correct probability is above 95%. Brief Description of the Drawings

[0047] Figure 1 is a schematic flow chart of the method of the present invention;

[0048] Figure 2 is a schematic flow chart of non-amplitude modulation data detection in the method of the present invention;

[0049] Figure 3 is a schematic diagram of GMSK acquisition signal detection when the signal-to-noise ratio is 0 dB;

[0050] Figure 4 is a waveform diagram of a radar signal when the signal-to-noise ratio is 2 dB;

[0051] Figure 5 is a schematic diagram of radar signal detection when the signal-to-noise ratio is 2 dB;

[0052] Figure 6 is a waveform diagram of a single signal of various modulation types in a simulation data set;

[0053] Figure 7 is a waveform diagram of an overlapping signal in a simulation data set;

[0054] Figure 8 is a schematic diagram of low duty cycle radar signal detection. Detailed Embodiment

[0055] To make the objectives, technical solutions and advantages of the present invention more clear and understandable, the following further elaborates on the present invention in conjunction with the appended Figure 1-8 drawings and embodiments. It should be understood that the specific embodiments described herein are merely for explaining the present invention and are not intended to limit the present invention.

[0056] Embodiment 1 discloses a blind signal time-domain detection method based on self-correlation three-window decision. By adopting the idea of setting thresholds for three detection windows, combining relative thresholds and absolute thresholds, and determining the detection threshold jointly according to the data types of the three detection windows, finally, the sliding detection of the self-correlation value of the received data is performed to determine the signal position. As Figure 1 shown, it includes the following steps:

[0057] Step 1: Perform a correlation operation on the received discrete data x(n) to obtain the autocorrelation function R n (n), where n represents the sampling point:

[0058] The expression of the discrete data x(n) is:

[0059]

[0060] In the above formula, is the actual signal; A is the signal amplitude; is the initial phase; T s is the sampling interval; f n is the carrier frequency, and w(n) is the zero-mean additive white Gaussian noise after sampling;

[0061] The calculation formula of the data autocorrelation function R n (n) is:

[0062]

[0063] In the above formula, N is the length of the correlation window, * represents taking the conjugate of the data, and i sam represents the discrete sampling point;

[0064] In this embodiment, since as the signal carrier frequency increases, the relative positions of the sampling points in different periods are uncertain, resulting in a significant deterioration of the correlation value. To optimize this problem, the absolute value of the signal can be taken during the autocorrelation operation, and the autocorrelation function of the signal can be updated by sliding recursion, which is a fast algorithm that is easy to implement in hardware. The expression is:

[0065] R n (n + 1) = R n (n) + |x(n + N)x*(n + N + 1)|

[0066] -|x(n)x*(n + 1)|

[0067] Step 2: According to the length L of x(n), set three windows at the beginning, middle, and end of the data for data type decision within the windows. The data type within each window can be divided into three categories: noise, constant modulus signal, and amplitude-varying signal. The average power P of noise and constant modulus signal m and the maximum autocorrelation value R m are approximately the same numerically, while there is a large difference between the two for amplitude-varying signals. Among them, there are obvious amplitude differences between the pulse part and the rest period of the radar signal, which can be considered as amplitude-varying signals during window decision. If the window is all the rest period of the radar signal, it is considered as noise. Then the possible data forms of amplitude-varying signals within the window are: noise-signal, single-signal - aliased signal, amplitude modulation signal, radar pulse signal. Since there must be signals within the window and the data amplitude changes, for simplicity, they are collectively referred to as amplitude modulation data. The value of the decision window label m is 1, 2, 3, and the decision window length is defined as N Det , and Calculate the average power and the maximum autocorrelation value R m within the three windows respectively, and obtain the proportionality coefficient k between the two:

[0068]

[0069] Step 3: Set the proportionality coefficient decision index k ind = 1.8 (obtained according to the characteristics of amplitude modulation data). If k ≥ k ind , then there are signals within the window and the amplitude changes, that is, there is amplitude modulation data, and the absolute threshold value th of the received data is calculated abs_1 ;

[0070] Step 3 includes:

[0071] Step 31: Determine the window number j where amplitude modulation data exists, j ∈ m, and calculate the maximum autocorrelation value R j max of the data within the window respectively;

[0072] Step 32: Take the minimum value R j max of R j ' max and multiply it by the threshold coefficient α lower than the signal to calculate the absolute threshold value th abs_1 of the received data:

[0073] th abs_1 = R j ' max *α.

[0074] Since there must be a signal, the threshold should be less than the autocorrelation value of the signal, that is, α < 1. In order to distinguish the correlation value between the signal and the noise with the threshold, in this embodiment, α is taken as 0.5. If it is too large, it may lead to signal missed detection. If it is too small, the noise may be determined as a signal, resulting in detection false alarm.

[0075] Step 4, if k < k ind , then there may not be a signal in the window, and it is necessary to determine the threshold for the non - amplitude - modulated data. Calculate the relative threshold th rel within the decision window m respectively. According to the specific situation where the autocorrelation value in the window exceeds the relative threshold, calculate the absolute threshold value th abs_2 or th abs_3 ;

[0076] As Figure 2 shown, step 4 includes:

[0077] Step 41, when calculating the absolute threshold for the non - amplitude - modulated data, calculate the maximum autocorrelation value R m within the decision window m respectively. Multiply R m by the relative threshold decision coefficient ε. ε takes values from 0.2 to 0.5, and calculate the relative threshold th rel of the received data:

[0078] th rel = R m * ε

[0079] Step 42, the correlation value of the noise within the window is random and fluctuates around the relative threshold th rel . The correlation value of the constant - modulus signal is stable. According to this characteristic difference between the noise and the signal, determine whether the data type within the window is noise or signal;

[0080] Step 43, if the data types determined by the three decision windows are all noise, calculate the maximum autocorrelation value R i max within the decision window m respectively, i ∈ m. Take the maximum value R i max as R i ' max and multiply it by the threshold coefficient λ higher than the noise to calculate the absolute threshold value th abs_3 of the received data:

[0081] th abs_3 = R i ' max * λ

[0082] Since the decision windows are all noise, the threshold should be higher than the autocorrelation value of the noise, i.e., λ > 1. To distinguish the correlation value of the signal from that of the noise by the threshold, in this embodiment, λ is taken as 1.5. If it is too large, it may lead to signal missed detection; if it is too small, it may misjudge the noise as a signal, resulting in false detection.

[0083] Step 44, if the decision data types of the three decision windows are not all noise, that is, there is non - AM data, then three absolute threshold values th are calculated according to different data types. abs_2 :

[0084] The first case is that if all three windows are signals, then the maximum autocorrelation value R within the decision window m is calculated respectively, where k ∈ m, and the minimum value R' of R is taken. Multiply it by the threshold coefficient α lower than the signal to calculate the absolute threshold value th of the received data. k max , k ∈ m, take the minimum value R' of R k max and multiply it by the threshold coefficient α lower than the signal to calculate the absolute threshold value th of the received data. k ' max : abs_2 :

[0085] th abs_2 = R k ' max * α

[0086] If not all three windows are signals, then select the minimum value from the maximum autocorrelation values of all windows with signals and define it as Rmini1. Then find the maximum autocorrelation value in the received data and define it as RmaxAll. The second case is that if the ratio of RmaxAll to Rmini1 is greater than the decision coefficient μ. In this embodiment, μ is taken as 1.8 according to the difference between the radar signal pulse correlation peak and the correlation value during the rest period. Then the absolute threshold value th abs_2 is:

[0087] th abs_2 = R p ' max * λ

[0088] In the above formula, p is the serial number of the window with noise, p ∈ m, calculate the maximum autocorrelation value R of the data within the p - window respectively p max , take the minimum value R' of R p max and λ is the threshold coefficient higher than the noise; p ' max

[0089] The third case is that if the ratio of RmaxAll to Rmini1 is less than the decision coefficient μ, then the absolute threshold value th abs_2 is:

[0090] th abs_2= R q ' max *α

[0091] In the above formula, q is the window serial number where the signal exists, q ∈ m, and the maximum autocorrelation value R of the data within the q window is calculated respectively q max , and take R q max The minimum value R q ' max , and α is the threshold coefficient lower than the signal.

[0092] Step 5, compare the absolute threshold values calculated in Step 3 and Step 4 with the autocorrelation function R n (n), and set both the sliding window length and the sliding interval to be L win . If there is an autocorrelation value greater than the absolute threshold within the sliding window, it is determined that the entire sliding window has a signal. After the sliding window traverses all the received data, the signal position is output.

[0093] To sum up, due to the signal amplitude fluctuation of the amplitude modulation data, the autocorrelation function fluctuates. Some positions where the signal exists may have a low autocorrelation function value. Judging that there is a signal only by the autocorrelation function being greater than the threshold is likely to cause signal omission and detection segmentation. The method of the present invention effectively improves the signal detection accuracy through sliding window sliding judgment.

[0094] The performance of the method of the present invention is illustrated through simulation experiments. Use a software-defined radio platform to collect a GMSK continuous signal with a signal-to-noise ratio of 0 dB, a signal symbol rate of 9600 sps, a signal carrier frequency of 70 MHz, a signal sampling rate of 245.76 MHz, and the data length for time-domain detection is 2,457,600 sampling points, that is, 10 milliseconds Figure 3 is the detection schematic diagram of the GMSK collected signal when the signal-to-noise ratio is 0 dB. From Figure 3 it can be seen that the vast majority of the autocorrelation functions of the data are above the decision threshold, and there are individual sampling points below the threshold. Since the sliding window makes a sliding judgment on the signal, the detected signal position is the entire data interval, that is, the detection is correct. The time-domain detection situation of a radar signal with a duty cycle of 1% when the signal-to-noise ratio is 2 dB is verified by simulation Figure 4 and Figure 5 are the radar signal waveform and the radar signal detection schematic diagram respectively. From Figure 4 it can be seen that due to the low signal-to-noise ratio and the short pulse duration of the radar, the position of the pulse cannot be determined from the waveform diagram. From Figure 5It can be seen that after autocorrelating the data, the amplitude of the pulse signal part is obvious and exceeds the decision threshold, and the signal can be correctly detected. The above detection results of communication and radar signals under low signal-to-noise ratio show that signals can be correctly detected at signal-to-noise ratios of 0 dB and 2 dB, and the method of the present invention has strong anti-noise performance.

[0095] In order to further verify the detection performance of the method of the present invention, a scenario where signals and noise alternate and appear randomly in a real situation is simulated, and a data set is generated using MATLAB simulation. The data set has different signal-to-noise ratios, data types, and modulation methods, and the specific content is shown in Table 1 below.

[0096] Table 1: MATLAB simulation data set

[0097]

[0098] Figure 6 are the single-signal waveforms of various modulation types in the simulation data set. The carrier frequency is set to 70 MHz, the sampling frequency is set to 160 MHz, and 8 single signals are simulated in the modulation order of BPSK, QPSK, 8PSK, 2FSK, 4FSK, 2ASK, 16QAM, and GMSK, and then a communication signal combination with random amplitude and random position is generated. The 8 single signals are combined in pairs for aliasing, and there are 64 combination results, as shown in Table 2 below.

[0099] Table 2: Simulation combination of aliased signals

[0100] Modulation method BPSK QPSK 8PSK 2FSK 4FSK 2ASK 16QAM GMSK BPSK √ √ √ √ √ √ √ √ QPSK √ √ √ √ √ √ √ √ 8PSK √ √ √ √ √ √ √ √ 2FSK √ √ √ √ √ √ √ √ 4FSK √ √ √ √ √ √ √ √ 2ASK √ √ √ √ √ √ √ √ 16QAM √ √ √ √ √ √ √ √ GMSK √ √ √ √ √ √ √ √

[0101] Select one aliased signal for illustration - BPSK signal aliased with BPSK signal. Set the carrier frequency of signal 1 to 70 MHz, the carrier frequency of signal 2 to 50 MHz, and the sampling frequency to 160 MHz. Figure 7 are the waveforms of the aliased signals in the simulation data set. Two natural numbers are randomly generated to change the amplitude of the baseband signal, and the natural numbers are in the range of [1 - 5] to achieve the effect of random signal amplitude. When fusing data, the signal position is uncertain, noise is added before and after the signal, presenting in the form of noise-signal-noise, and the length of the noise added before and after the signal is determined by a randomly generated positive integer. Each signal-to-noise ratio data set contains 740 random signals, and the specific detection results are shown in Table 3 below.

[0102] Table 3: Detection results of data sets under different signal-to-noise ratio conditions

[0103]

[0104] For the data set of software simulation, all generated signals (single signal, aliased signal, communication signal, radar signal) are integrated and divided into four packets of data according to the signal-to-noise ratio. The SNRs are 8, 10, 15, and 20 dB respectively. As can be seen from Table 3, the correct detection probability of each packet of data is above 95%. Although there are cases of signal undetected, specifically, the detected interval fails to completely contain the true signal interval, and there is no complete signal undetected. It proves that the method of the present invention has universality, does not need to distinguish between communication signals and radar signals, is applicable to various modulation methods of communication signals, and can correctly detect pulse signals even in the case of a low radar duty cycle such as 0.06%. Figure 8 is a schematic diagram of low duty cycle radar signal detection. As can be seen from Figure 8 it, for the position of the radar pulse signal, the amplitude of the autocorrelation function is higher than the adaptive decision threshold, that is, the detection is correct. To sum up, the method of the present invention has high efficiency, anti-noise performance, universality, and adaptability.

[0105] After verification, the method proposed by the present invention solves the technical problems proposed by the present invention. The method of the present invention has been verified by simulation experiments and practical applications, and the technical effects recorded in the specification of the present invention have been verified.

[0106] The algorithm (method) proposed by the present invention is the underlying technical core of the present invention, and various products can be derived based on the algorithm.

[0107] Based on the algorithm (method) proposed by the present invention, a blind signal time-domain detection system based on autocorrelation three-window decision is developed using a programming language. The system has program modules corresponding to the steps of the above technical solution, and executes the steps in the above-mentioned blind signal time-domain detection method based on autocorrelation three-window decision when running.

[0108] The computer program of the developed system (software) is stored on a computer-readable storage medium. The computer program is configured to implement the steps of the above-mentioned blind signal time-domain detection method when called by a processor. That is, the present invention is materialized on a carrier to become a computer program product.

[0109] A blind signal time-domain detection device based on autocorrelation three-window decision, the device includes at least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the above-mentioned blind signal time-domain detection method based on autocorrelation three-window decision to achieve blind signal time-domain detection.

[0110] The various embodiments of the systems and techniques described herein can be implemented in digital electronic circuitry, integrated circuit systems, application specific ASICs (application specific integrated circuits), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special or general purpose programmable processor that receives data and instructions from a storage system, at least one input device, and at least one output device, and transmits the data and instructions to the storage system, the at least one input device, and the at least one output device.

[0111] The computational programs (also referred to as programs, software, software applications, or code) in the present invention include machine instructions for a programmable processor and can implement these computational programs using high-level procedural and / or object-oriented programming languages, and / or assembly / machine languages. As used in the present invention, the terms "machine-readable medium" and "computer-readable medium" refer to any computer program product, device, and / or apparatus (e.g., magnetic disks, optical disks, memories, programmable logic device PLD) for providing machine instructions and / or data to a programmable processor, including a machine-readable medium that receives machine instructions as a machine-readable signal. The term "machine-readable signal" refers to any signal for providing machine instructions and / or data to a programmable processor.

[0112] The above content is a further detailed description of the present invention in combination with specific preferred embodiments, and it cannot be determined that the specific implementation of the present invention is only limited to these descriptions. For those of ordinary skill in the technical field to which the present invention pertains, without departing from the concept of the present invention, several simple deductions or substitutions can still be made, and all should be regarded as belonging to the protection scope of the present invention.

Claims

1. A blind signal time domain detection method based on autocorrelation three-window decision, characterized in that: The steps include: Step 1: Perform correlation operation on the received discrete data x(n) to obtain the autocorrelation function R of the data. n (n), n represents the sampling point: The expression of discrete data x(n) is: In the above formula, is the actual signal; A is the signal amplitude; is the initial phase; T s is the sampling interval; f n is the carrier frequency, w(n) is the sampled zero-mean additive white Gaussian noise; Data autocorrelation function R n The calculation formula for (n) is: In the above formula, N is the correlation window length, * means to conjugate the data, i sam represents discrete sampling points; Step 2: According to the length L of x(n), set the beginning, middle and end windows of the data to judge the data type in the window. The values ​​of the judgment window number m are 1, 2, and 3. The judgment window length is defined as N. Det ,and Calculate the average power in the three windows respectively and the maximum autocorrelation value R m , and obtain the proportional coefficient k between the two: Step 3: Set the proportionality factor judgment index k ind =1.8, if k≥k ind , then there is amplitude modulated data in the window, and the absolute threshold value th of the received data is calculated. abs_1 ; Step 4: If k < k ind , respectively calculate the relative threshold th within the decision window m rel , according to the specific situation that the autocorrelation value in the window exceeds the relative threshold, the absolute threshold value th of the received data is calculated abs_2 or abs_3 ; Step 5: Compare the absolute threshold values ​​calculated in steps 3 and 4 with the data autocorrelation function R n (n) Compare, set the sliding window length and sliding interval to L win If there is an autocorrelation value greater than the absolute threshold in the sliding window, it is determined that there is a signal in the entire sliding window, and the sliding window outputs the signal position after traversing all received data.

2. The blind signal time domain detection method based on autocorrelation three-window decision according to claim 1 is characterized in that: Step 3 includes: Step 31, determine the window number j where the amplitude modulated data exists, j∈m, and calculate the maximum autocorrelation value R of the data in the window respectively. j max ; Step 32, take R j max The minimum value R j ' max Multiply it by the threshold coefficient α below the signal to calculate the absolute threshold value th of the received data abs_1 : th abs_1 =R j ' max *a.

3. The blind signal time domain detection method based on autocorrelation three-window decision according to claim 2 is characterized in that: Step 4 includes: Step 41, respectively calculate the maximum autocorrelation value R in the decision window m m , R m Multiply it by the relative threshold decision coefficient ε to calculate the relative threshold th of the received data rel : the rel =R m *ε Step 42, determining whether the data type in the window is noise or signal; Step 43: If the data types of the three decision windows are all noise, the maximum autocorrelation value R in the decision window m is calculated respectively. i max , i∈m, take R i max The maximum value R i ' max Multiply it by the threshold coefficient λ higher than the noise to calculate the absolute threshold value th of the received data abs_3 : th abs_3 =R i ' max *l Step 44: If the data types determined by the three decision windows are not all noise, then the absolute threshold value th is calculated based on different data types. abs_2 : If all three windows are signals, the maximum autocorrelation value R in the decision window m is calculated respectively. k max , k∈m, take R k max The minimum value R k ' max Multiply it by the threshold coefficient α below the signal to calculate the absolute threshold value th of the received data abs_2 : th abs_2 =R k ' max *a If all three windows are not signals, the minimum value is selected from the maximum values ​​of the autocorrelation of all windows with signals and is defined as Rmini1. Then the maximum value of the autocorrelation in the received data is defined as RmaxAll. If the ratio of RmaxAll to Rmini1 is greater than the decision coefficient μ, the absolute threshold value th abs_2 for: th abs_2 =R p ' max *l In the above formula, p is the window number with noise, p∈m, and the maximum autocorrelation value R of the data in the p window is calculated respectively. p max , take R p max The minimum value R p ' max , λ is the threshold coefficient higher than the noise; If the ratio of RmaxAll to Rmini1 is less than the decision coefficient μ, the absolute threshold value th abs_2 for: th abs_2 =R q ' max *a In the above formula, q is the window number where the signal exists, q∈m, and the maximum autocorrelation value R of the data in the q window is calculated respectively. q max , take R q max The minimum value R q ' max , α is the threshold coefficient below the signal.

4. The blind signal time domain detection method based on autocorrelation three-window decision according to claim 3 is characterized in that: α is 0.5, ε is 0.2-0.5, λ is 1.5, and μ is 1.

8.

5. A blind signal time domain detection method system based on autocorrelation three-window judgment, characterized by: The system has a program module corresponding to the steps of any one of claims 1 to 4 above, and executes the steps in the blind signal time domain detection method based on autocorrelation three-window decision when running.

6. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and the computer program is configured to implement the steps of a blind signal time domain detection method based on autocorrelation three-window decision according to any one of claims 1 to 4 when called by a processor.

Images