A Fast Time Difference Prediction Method Based on Short-Time Fourier Transform Data

Abstract

This invention discloses a method for rapid time difference prediction based on short-time Fourier transform data, belonging to the field of radar target detection signal processing. This invention directly utilizes the spectral signal obtained from the short-time Fourier transform during echo signal preprocessing for rapid time difference prediction, eliminating the step of converting the echo time-domain signal to the frequency domain in the processing flow. When the radar system detects a valid target, the two spectral signal matrices obtained in the preprocessing process are subjected to constant false alarm rate (CFAR) detection and then multiplied in the time and frequency domains. The resulting time-frequency domain signal matrices are then superimposed along the time dimension. The superimposed column signals are then subjected to a fast inverse Fourier transform to achieve rapid prediction of the time difference of signals received by multiple passive radars. This invention fully utilizes the spectral signal obtained from the short-time Fourier transform signal preprocessing of the passive radar system, simplifying the signal processing flow and reducing computational load while maintaining the same accuracy, thus improving the efficiency of target localization.

Description

Technical Field

[0001] This invention relates to a method for rapid time difference prediction based on short-time Fourier transform data, belonging to the field of radar target detection signal processing. Background Technology

[0002] The detection and location of unknown aerial targets is of great significance in modern radar-based conventional air defense missions. After years of research, radar-based aerial target detection and location technology has made significant progress, and various military and civilian radars are now widely used. For active radar, the commonly used method for measuring the distance between the radar and the target is primarily based on the delay of the echo relative to the transmitted signal. Depending on the radar's transmitted signal, the delay time can typically be measured using the pulse method, frequency method, or phase method. For passive radar, since it only receives signals and does not transmit them, it is impossible to know the time it takes for the electromagnetic wave to leave the target, making it difficult to directly calculate the distance between the target and the receiving station. Therefore, passive radar ranging generally requires measurements from multiple observation stations. A common method is to simultaneously measure the time difference of arrival of the electromagnetic wave signals emitted by the aerial target from multiple receiving stations. Multi-station time difference positioning, also known as hyperbolic positioning, has become the most promising positioning method in modern passive positioning technology. Furthermore, compared to active radar, passive radar has superior four-fold resistance characteristics: detection capability against low-altitude / ultra-low-altitude intruding targets, comprehensive electronic jamming capability, detection capability against stealth targets, and resistance to anti-radiation missiles. Furthermore, since passive radar utilizes signals from radiation sources already present in the airspace, it does not require high-power transmitters; only corresponding receiving equipment is needed, resulting in a small equipment footprint and low cost. Therefore, using passive radar to detect and locate unknown aerial targets has become a current research hotspot and is widely used in airport and urban air surveillance, as well as airspace defense for important areas.

[0003] The time difference between the electromagnetic wave signal emitted by the target and the arrival of the two detection and receiving stations forms a hyperboloid (line). The position of the target can be obtained by the intersection of multiple hyperboloids (lines). A major problem in this positioning method is whether the time difference between the detected and received signals of multiple stations can be calculated quickly. The time difference calculation is an extremely important step in the target positioning process. In the process of passive radar detecting electromagnetic wave signals in the air, the signal preprocessing step of the radar system is to perform spectrum analysis on the detected and received electromagnetic wave signals through short-time Fourier transform. If characteristic signals (such as step frequency signals) are found in the spectrum, it can be determined that there is an air target activity. After a conventional radar system detects an unknown air target, it will take the following steps to calculate the time difference. Taking the solution of the time difference between the received signals of two passive radars as an example: (1) Perform fast Fourier transform (FFT) on the time domain signals of the two stations to convert the time domain signals to the frequency domain; (2) Multiply the two frequency domain signals in the frequency domain; (3) Perform fast inverse Fourier transform (IFFT) on the signal obtained by multiplying in the frequency domain to complete the fast time difference calculation. This process is quite complex, involves a huge amount of data processing, and consumes the limited computing resources of the radar system. Therefore, how to quickly calculate the time difference of the detected electromagnetic wave signals is a challenge that needs to be studied to improve the radar's performance in rapidly locating unknown aerial targets. Summary of the Invention

[0004] The purpose of this invention is to provide a method for rapid prediction of time difference based on short-time Fourier transform data, which is based on the short-time Fourier transform of multiple passive radar detection and reception signals. This method makes full use of the spectral signal obtained from the short-time Fourier transform signal preprocessing process of the passive radar system, simplifies the signal processing flow, reduces the amount of computation, and enables the passive radar to quickly locate unknown targets in the air while ensuring the same accuracy.

[0005] The objective of this invention is achieved through the following technical solution.

[0006] This invention discloses a fast time difference prediction method based on short-time Fourier transform data. It directly utilizes the spectral signal obtained from the short-time Fourier transform during echo signal preprocessing for rapid time difference prediction. When the radar system detects a valid target, the two spectral signal matrices obtained from the preprocessing process are subjected to constant false alarm rate (CFAR) detection and then multiplied in the time and frequency domains. The resulting time and frequency domain signal matrices are then superimposed along the time dimension. The superimposed column signal is then subjected to a fast inverse Fourier transform to complete the rapid time difference prediction. This invention fully utilizes the spectral signal obtained from the short-time Fourier transform signal preprocessing of the passive radar system. Compared with traditional time difference calculation methods, this method simplifies the signal processing flow, reduces computational load, and improves target localization efficiency while maintaining the same accuracy.

[0007] This invention discloses a method for rapid time difference prediction based on short-time Fourier transform data, comprising the following steps:

[0008] Step 1: Perform preprocessing and spectrum monitoring on the echo signals s1(m) and s2(m) detected by the two radars. Use the spectrum signal obtained from the short-time Fourier transform signal preprocessing process of the passive radar system for spectrum monitoring. Perform spectrum analysis on the two obtained spectrum signal matrices. If there are characteristic signals in the spectrum signal matrix, it is determined that there is an airborne target, and proceed to Step 2.

[0009] Step 1.1: Perform preprocessing operations on the detected signals, that is, transform the time-domain signals of the two radars into time-frequency domain signals based on short-time Fourier transform, and obtain the time-frequency spectrum matrices S1(n,f) and S2(n,f) respectively.

[0010] Step 1.2: If a characteristic signal is found in the signal spectrum, indicating the presence of an aerial target, proceed to Step 2.

[0011] Step 1.3: If no characteristic signal is found in the signal spectrum, continue spectrum monitoring.

[0012] The characteristic signals mentioned in step 1.2 include, but are not limited to, step frequency signals, frequency hopping signals, linear frequency modulation signals, etc.

[0013] Step 2: Using the two spectral signal matrices S1(n,f) and S2(n,f) obtained from the short-time Fourier transform during echo signal preprocessing in Step 1, perform constant false alarm rate (CFAR) detection. Retain spectral signal matrix data exceeding the threshold, and set spectral signal data below the threshold to zero. This significantly reduces the computational load. New spectral matrices S1(n,f) and S2(n,f) are obtained, and time-frequency domain complex multiplication is performed to obtain the time-frequency domain signal matrix Sn. 12 (n,f). Instead of converting the time-domain echo signals detected by the two radars into frequency-domain signals, the operation is performed directly on the spectrum signal matrix after the short-time Fourier transform.

[0014] Step 2.1: Perform constant false alarm rate (CFAR) detection. Based on the radar system and the influence of environmental noise, reasonably estimate the detection threshold. Use this threshold to perform CFAR detection on the two spectrum signal matrix data obtained in Step 1. Only retain the spectrum signal matrix data that exceeds the threshold, and set the spectrum signal matrix data that does not exceed the threshold to zero, thus obtaining new spectrum matrices S1(n,f) and S2(n,f).

[0015] Step 2.2: Perform complex conjugate multiplication. Perform complex conjugate multiplication on S1(n,f) to obtain S1. * (n,f), then S1 *Multiplying the corresponding elements of the spectrum matrix of (n,f) and S2(n,f) yields S 12 (n,f).

[0016]

[0017] Step 3: The time-frequency domain signal matrix obtained in Step 2 is superimposed along the time dimension. The superimposed column signal is then subjected to a fast inverse Fourier transform to calculate the time difference between the two radars. This enables rapid prediction of the time difference between the received signals from the two passive radars based on the short-time Fourier transform data.

[0018] Step 3.1: For S 12 The (n,f) spectrum matrix is ​​superimposed along the time dimension to obtain S. 12 (f) Sequence.

[0019] Step 3.2: For S 12 (f) Perform a fast inverse Fourier transform to obtain the sequence r. 12 (n). Sequence r 12 The time corresponding to the peak value of (n) is the time difference between radar receiving station 2 and radar receiving station 1, that is, the time difference between the detection and reception signals of the two passive radars is quickly predicted based on the short-time Fourier transform data.

[0020] Step 4: Based on Steps 1 to 3, and building upon the rapid prediction of the time difference between signals received by two passive radars, further achieve rapid prediction of the time difference between signals received by multiple passive radars.

[0021] Beneficial effects:

[0022] 1. Compared with conventional methods, the present invention discloses a method for rapid time difference prediction based on short-time Fourier transform data, which directly uses the radar detection echo signal spectrum based on short-time Fourier transform. In terms of processing flow, it directly eliminates the step of converting the echo time domain signal to the frequency domain, which can significantly simplify the radar signal processing flow and improve computational efficiency.

[0023] 2. This invention discloses a fast time difference prediction method based on short-time Fourier transform data. It utilizes the radar echo signal spectrum based on short-time Fourier transform, performs a conjugate complex multiplication operation on the obtained station 1 and station 2 spectra, then superimposes the spectrum matrices along the time dimension, and finally performs a fast inverse Fourier transform to complete the time difference calculation. Compared with conventional methods, this method reduces the data computation load of the Fourier transform process, significantly reduces the overall algorithm's computational load, and achieves fast time difference prediction.

[0024] 3. The present invention discloses a method for rapid prediction of time difference based on short-time Fourier transform data, which, based on the rapid prediction of the time difference of signals received by two passive radars, further realizes the rapid prediction of the time difference of signals received by multiple passive radars. Attached Figure Description

[0025] Figure 1 This is a schematic diagram of a radio detection radar system in an embodiment of the present invention, "A method for rapid prediction of time difference based on short-time Fourier transform data".

[0026] Figure 2 This is a flowchart of the algorithm of the present invention in an embodiment of the invention "A method for fast prediction of time difference based on short-time Fourier transform data".

[0027] Figure 3 This is a time spectrum diagram of the received signals from two radars in an embodiment of the present invention, "A method for fast prediction of time difference based on short-time Fourier transform data". Figure 3 (a) is the time-frequency diagram of the signal received at station 1. Figure 3 (b) is the time-frequency diagram of the signal received at station 2.

[0028] Figure 4 This is a time difference result diagram from an embodiment of the present invention, "A method for rapid prediction of time difference based on short-time Fourier transform data". Detailed Implementation

[0029] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings and specific examples. It should be noted that the described embodiments are only intended to facilitate the understanding of the present invention and do not serve any limiting purpose.

[0030] Example 1:

[0031] This embodiment illustrates how the "a fast time difference prediction method based on short-time Fourier transform data" of the present invention is used to perform fast time difference calculation through a simulation example.

[0032] In this embodiment, the simulation software is MATLAB. The relevant parameters of the emitted waveform from the airborne target radiation source are as follows: signal baseband bandwidth 8MHz, carrier frequency changes every 10ms, uniformly jumping from 30MHz to 70MHz (i.e., 30MHz, 40MHz, 50MHz, 60MHz, 70MHz), sampling rate 100MHz, and total signal duration 50ms. This embodiment employs a bistatic radar structure as follows: Figure 1As shown, a dual-station air detection radar system is formed, consisting of one airborne radiation source and two passive radar receiving systems, namely Station 1 radar and Station 2 radar. It is assumed that both Station 1 and Station 2 radar systems can completely receive the radio signals emitted by the airborne radiation source. The radiation source signal received by Station 2 is assumed to be delayed by 30 ns relative to that received by Station 1, for time difference simulation verification.

[0033] Step 1: Perform short-time Fourier transform (SFT) on the target signals s1(m) and s2(m) received by receiving stations 1 and 2. Set the window function w(m) of the SFT to a rectangular window with a window length W = 128. For the data within a time window, use the fast Fourier transform (FFT) method to calculate and obtain the time-frequency data S1(n,f) and S2(n,f), where n represents the discrete time and f represents the discrete frequency. The time-frequency diagrams of receiving stations 1 and 2 are shown below. Figure 3 As shown in Figures (a) and (b), it can be seen that both receiving station 1 and receiving station 2 have received the signal completely, and a clear step frequency spectrum can be seen.

[0034] Step 2: Due to signal frequency hopping, signal spectrum analysis reveals that the effective data duration for the same carrier frequency is 0–10ms, 10–20ms, 20–30ms, 30–40ms, and 40–50ms. Here, only the first 10ms of effective data is analyzed. Constant false alarm rate (CFAR) detection is performed on the signal spectra obtained from stations 1 and 2 respectively to obtain the effective data. The analysis process for other time periods of the same frequency signal is similar and will not be elaborated further. S1 is obtained by performing CFAR detection on the first 10ms of data from S1(n,f) and then performing a complex conjugate operation. * (n,f), S1 * Multiplying the corresponding elements of the effective data matrix of (n,f) and S2(n,f) in the first 10ms yields S 12 (n,f).

[0035] Step 3, for S 12 Adding data of the same frequency f (n, f) is equivalent to matrix column addition, resulting in S. 12 (f) and then S 12 (f) Perform an inverse fast Fourier transform to obtain the time-domain correlation sequence r. 12 (n). r 12 The peak value of (n) corresponds to a delay of 30ns, thus completing the verification. The verification results are as follows: Figure 4 As shown.

[0036] Comparison of computational complexity: The length of each received signal segment is N=2. pBoth conventional algorithms and the algorithm in this patent require performing a short-time Fourier transform (SFT) on the signal for time-frequency detection before calculating the time difference. Therefore, the SFT is not included in the computational complexity calculation. Conventional correlation sequence frequency domain calculation methods require two N-point FFTs, one N-point multiplication, and one N-point IFFT. Thus, the computational complexity of the conventional method is... Complex multiplication of the first order and complex addition of the third order (3·N·p). For the fast algorithm proposed in this patent, the window length W = 2. q This requires one complex multiplication of two N / W column W′ row data matrices, where W′ is the column signal length of the effective spectrum data matrix, one accumulation of N / W column W′ row data, and one W′ point IFFT. The computational cost is... Multiplication of complex numbers and Complex number addition.

[0037] Since the computation time for complex multiplication is much longer than that for complex addition in digital signal processing systems, the computational complexity of complex multiplication is the primary consideration for comparison. In this example, N = 1 × 10⁶ points, p ≈ 20, q = 7, W corresponds to 128, and W′ = 12. The conventional method requires 32.5M operations for complex multiplication, and this computational complexity increases with the number of points in N. The complex multiplication method of this patent requires only 96K operations. It can be seen that the computational complexity of this patented algorithm is far less than that of the conventional algorithm, reducing it by more than 300 times, thus proving the effectiveness of the algorithm.

[0038] Step 4: Based on Steps 1 to 3, and building upon the rapid prediction of the time difference between signals received by two passive radars, further achieve rapid prediction of the time difference between signals received by multiple passive radars.

[0039] Based on this, the present invention provides a rapid time difference prediction method for multiple passive radar detection and reception signals based on short-time Fourier transform. By directly utilizing the short-time Fourier transform-based signals from multiple passive radar detection and reception systems, a rapid time difference calculation process is completed, achieving the goal of rapid ranging of aerial targets. Compared with traditional time difference calculation methods, this invention fully utilizes the spectral signal obtained from the short-time Fourier transform during passive radar system signal preprocessing for rapid time difference calculation. The step of converting the echo time-domain signal to the frequency domain is directly eliminated in the processing flow. While maintaining the same accuracy, this greatly simplifies the radar signal processing flow and improves prediction efficiency.

[0040] The above detailed description further illustrates the purpose, technical solution, and beneficial effects of the invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for rapid time difference prediction based on short-time Fourier transform data, characterized in that: Includes the following steps, Step 1: Perform preprocessing and spectrum monitoring on the echo signals s1(m) and s2(m) detected by the two radars. Use the spectrum signal obtained from the short-time Fourier transform signal preprocessing process of the passive radar system for spectrum monitoring. Perform spectrum analysis on the two obtained spectrum signal matrices. If there are characteristic signals in the spectrum signal matrices, it is determined that there is an airborne target, and proceed to Step 2. Step 2: Using the two spectral signal matrices S1(n,f) and S2(n,f) obtained from the short-time Fourier transform during echo signal preprocessing in Step 1, perform constant false alarm rate (CFAR) detection. Retain spectral signal matrix data exceeding the threshold, and set spectral signal data below the threshold to zero. This significantly reduces the computational load. New spectral matrices S1(n,f) and S2(n,f) are obtained, and time-frequency domain complex multiplication is performed to obtain the time-frequency domain signal matrix Sn. 12 (n,f); Instead of converting the time-domain echo signals detected by the two radars into frequency-domain signals, the operation is performed directly on the spectrum signal matrix after the short-time Fourier transform. Step 3: The time-frequency domain signal matrix obtained in Step 2 is superimposed along the time dimension. The superimposed column signal is then subjected to a fast inverse Fourier transform to calculate the time difference between the two radars. This enables rapid prediction of the time difference between the received signals from the two passive radars based on the short-time Fourier transform data.

2. The method for fast time difference prediction based on short-time Fourier transform data as described in claim 1, characterized in that: It also includes step four, which, based on steps one to three, further enables rapid prediction of the time difference between the received signals from two passive radars.

3. A method for rapid time difference prediction based on short-time Fourier transform data as described in claim 1 or 2, characterized in that: The implementation method for step one is as follows: Step 1.1: Perform preprocessing operations on the detected signals, that is, transform the time-domain signals of the two radars into time-frequency domain signals based on short-time Fourier transform, and obtain the time-frequency spectrum matrices S1(n,f) and S2(n,f) respectively; Step 1.2: If a characteristic signal is detected in the signal spectrum, indicating the presence of an aerial target, proceed to Step 2. Step 1.3: If no characteristic signal is found in the signal spectrum, continue spectrum monitoring.

4. The method for rapid time difference prediction based on short-time Fourier transform data as described in claim 3, characterized in that: The second step is implemented as follows: Step 2.1: Perform constant false alarm rate (CFAR) detection; reasonably estimate the detection threshold based on the radar system and the influence of environmental noise, and use this threshold to perform CFAR detection on the two spectrum signal matrix data obtained in Step 1. Only retain the spectrum signal matrix data that exceeds the threshold, and set the spectrum signal matrix data that does not exceed the threshold to zero, thus obtaining the new spectrum matrices S1(n,f) and S2(n,f); Step 2.2: Perform complex conjugate multiplication; perform complex conjugate multiplication on S1(n,f) to obtain S1 * (n,f), then S1 * Multiplying the corresponding elements of the spectrum matrix of (n,f) and S2(n,f) yields S 12 (n,f); 5. The method for fast time difference prediction based on short-time Fourier transform data as described in claim 4, characterized in that: The method for implementing step three is as follows: Step 3.1: For S 12 The (n,f) spectrum matrix is ​​superimposed along the time dimension to obtain S. 12 (f) sequence; Step 3.2: For S 12 (f) Perform a fast inverse Fourier transform to obtain the sequence r. 12 (n); sequence r 12 The time corresponding to the peak value of (n) is the time difference between radar receiving station 2 and radar receiving station 1, that is, the time difference between the detection and reception signals of the two passive radars is quickly predicted based on the short-time Fourier transform data.

6. The method for fast time difference prediction based on short-time Fourier transform data as described in claim 5, characterized in that: The characteristic signals mentioned in step 1.2 include, but are not limited to, step frequency signals, frequency hopping signals, and linear frequency modulation signals.

Images