A data segmentation method for radio signal detection and identification

Abstract

The present application relates to a data segmentation method for radio signal detection and recognition, belonging to the field of radio signal analysis. The method comprises: (1) collecting radio signal samples. Collecting multiple representative radio signal sequence data to form a sample set. (2) initial segmentation. Selecting an initial segmentation length n, and dividing each sample in the sample set into multiple segments with the length of n. (3) calculating the autocorrelation of each segment. Calculating the correlation value of the first half and the second half of each segment. (4) determining the segmentation length. If the autocorrelation of more than half of the segments is less than a certain threshold, the segmentation length can be determined as n, otherwise, the value of n is increased, and step (2) is performed again until the segmentation length is determined. The present application can save the time for designing a radio signal detection and recognition method.

Description

Technical Field

[0001] This invention belongs to the field of radio signal analysis, and specifically relates to a data segmentation method for radio signal detection and identification. Background Technology

[0002] Detection and identification of various radio signals are fundamental to radio spectrum management and radio spectrum resources. Radio signals are typically sequences of data of varying lengths and durations. To detect and identify radio signals, these sequences need to be further divided into segments of fixed length based on detection and identification algorithms.

[0003] The commonly used data segmentation method involves determining a segment length *n* based on experience, dividing a sequence into multiple segments of fixed length *n*, and then inputting these segments into a radio signal detection and recognition algorithm. The size of the data segments *n* is adjusted based on the detection and recognition results. This data segmentation method requires extensive experimentation, is time-consuming and labor-intensive, and cannot guarantee satisfactory results.

[0004] Therefore, there is a need for a data segmentation method that can overcome this defect, and this invention is created in response to this real need. Summary of the Invention

[0005] (a) Technical problems to be solved

[0006] The technical problem to be solved by this invention is how to provide a data segmentation method for radio signal detection and identification, so as to solve the problem that existing data segmentation methods require a lot of experiments, are time-consuming and labor-intensive, and cannot guarantee satisfactory results.

[0007] (II) Technical Solution

[0008] To address the aforementioned technical problems, this invention proposes a data segmentation method for radio signal detection and identification, which includes the following steps:

[0009] S1. Collect a set of radio signal samples: Collect multiple representative radio signal sequence data to form a sample set;

[0010] S2. Initial segmentation: Select an initial segment length n, and divide each sample in the sample set into multiple segments of length n;

[0011] S3. Calculate the autocorrelation value of each segment: Calculate the autocorrelation value between the first half and the second half of each segment;

[0012] S4. Determine the segment length: If the autocorrelation values ​​of more than half of the segments are less than a certain threshold, then determine the segment length as n; otherwise, increment the value of n and repeat step S2 until the segment length can be determined.

[0013] Furthermore, in step S1, N valid radio signal sequence data are collected according to the needs of radio signal detection and identification to form a sample set.

[0014] Furthermore, N is not less than the length of the longest sequence in the sample set.

[0015] Furthermore, in step S2, the initial segment length n is not less than 1 / 2 of the length of the shortest sequence in the sample set. A sequence of length L is divided into L-n+1 segments, and the data overlap between two segments is at most n-1, resulting in a final number of segments M.

[0016] Furthermore, in step S3, if the data of a segment is a(1), a(2), ..., a(n), then when n is odd, i.e. n = 2m + 1, the first half of the data is a(1), a(2), ..., a(m + 1), and the second half of the data is a(m + 1), a(m + 2), ..., a(n).

[0017] Furthermore, in step S3, if a segment of data is a(1), a(2), ..., a(n), when n is an even number, i.e. n = 2m, then the first half of the data is a(1), a(2), ..., a(m), and the second half of the data is a(m+1), a(m+2), ..., a(n).

[0018] Furthermore, the correlation value C between the first half and the second half of the data is calculated as follows:

[0019]

[0020] In the formula, s(i) and t(i) are data segments of length k, s(i) is the first half and t(i) is the second half; when n is odd, k = m + 1, and when n is even, k = m.

[0021] Furthermore, in step S4, if the correlation values ​​of (M / 2)+1 segments out of the M segments are all less than a certain threshold, then the segment length is determined to be n; otherwise, the value of n is incremented to n+1, and step S2 is repeated until the segment length can be determined.

[0022] Furthermore, the threshold is 0.5.

[0023] Furthermore, segments of length n are input into the radio signal detection and recognition algorithm.

[0024] (III) Beneficial Effects

[0025] This invention proposes a data segmentation method for radio signal detection and identification. This invention utilizes the autocorrelation of radio signal segments to determine an appropriate segment length, avoiding the need for multiple trials to manually select segment lengths. This can save time in designing radio signal detection and identification methods and can play an important role in the field of radio signal processing. Attached Figure Description

[0026] Figure 1 This is a flowchart illustrating the data segmentation process for radio signal detection and identification according to the present invention. Detailed Implementation

[0027] To make the objectives, contents, and advantages of the present invention clearer, the specific embodiments of the present invention will be described in further detail below with reference to the accompanying drawings and examples.

[0028] This invention belongs to the field of radio signal analysis and relates to a method for segmenting radio signal data, providing data of appropriate length for radio signal detection and identification algorithms.

[0029] This invention discloses a data segmentation method for radio signal detection and identification. The method includes: (1) collecting radio signal samples. Collecting multiple representative radio signal sequence data to form a sample set. (2) initial segmentation. Selecting an initial segmentation length n, and dividing each sample in the sample set into multiple segments of length n. (3) calculating the autocorrelation of each segment. Calculating the correlation value between the first half and the second half of each segment. (4) determining the segment length. If the autocorrelation of more than half of the segments is less than a certain threshold, the segment length can be determined as n; otherwise, the value of n is incremented, and step (2) is repeated until the segment length can be determined.

[0030] The purpose of this invention is to provide a method for automatically segmenting radio signals to meet the needs of radio signal detection and identification.

[0031] To achieve the above objectives, this invention proposes a data segmentation method for radio signal detection and identification, the method comprising:

[0032] S1. Collect a set of radio signal samples. Collect multiple representative radio signal sequence data to form a sample set.

[0033] S2. Initial Segmentation. Choose an initial segment length n, and divide each sample in the sample set into multiple segments of length n.

[0034] S3. Calculate the autocorrelation value of each segment. Calculate the autocorrelation value between the first half and the second half of each segment.

[0035] S4. Determine the segment length. If the autocorrelation values ​​of more than half of the segments are less than a certain threshold, then determine the segment length as n; otherwise, increment the value of n and repeat step S2 until the segment length can be determined.

[0036] Example 1:

[0037] Figure 1 This is a flowchart of a data segmentation method for radio signal detection and identification. (Example:) Figure 1 As shown, the method includes:

[0038] (1) Collect a set of radio signal samples. Collect representative radio signal sequence data to form a sample set.

[0039] In practice, N valid radio signal sequence data can be collected according to the needs of radio signal detection and identification to form a sample set. N should not be less than the length of the longest sequence in the sample set.

[0040] (2) Initial segmentation. Select an initial segment length n, and divide each sample in the sample set into multiple segments of length n.

[0041] In practice, the initial segment length n should not be less than half the length of the shortest sequence in the sample set. Segments can overlap. A sequence of length L can be divided into (L-n+1) segments. The maximum number of overlapping data between two segments is (n-1), and the final number of segments is M.

[0042] (3) Calculate the autocorrelation of each segment. Calculate the autocorrelation value between the first half and the second half of each segment.

[0043] In practice, if the data of a segment is a(1), a(2), ..., a(n), then when n is odd (i.e., n = 2m + 1), the first half of the data is a(1), a(2), ..., a(m + 1), and the second half of the data is a(m + 1), a(m + 2), ..., a(n); when n is even (i.e., n = 2m), the first half of the data is a(1), a(2), ..., a(m), and the second half of the data is a(m + 1), a(m + 2), ..., a(n).

[0044] The correlation value C between the first half and the second half of the data is calculated as follows:

[0045]

[0046] In the formula, s(i) and t(i) are data segments of length k, s(i) is the first half and t(i) is the second half; when n is odd, k = m + 1, and when n is even, k = m.

[0047] (4) Determine the segment length. If the autocorrelation of more than half of the segments is less than a certain threshold, then the segment length is determined to be n. Otherwise, increment the value of n and repeat step (2) until the segment length can be determined.

[0048] In practice, if (M / 2)+1 of the M segments have a correlation value that is less than a certain threshold to avoid periodic repetition within a segment, such as 0.5, then the segment length is determined to be n. Otherwise, the value of n is incremented to n+1, and step (2) is repeated until the segment length can be determined.

[0049] Example 2:

[0050] A data segmentation method for radio signal detection and identification includes:

[0051] (1) Collect a set of radio signal samples. Collect multiple representative radio signal sequence data to form a sample set.

[0052] (2) Initial segmentation. Select an initial segment length n, and divide each sample in the sample set into multiple segments of length n.

[0053] (3) Calculate the autocorrelation of each segment. Calculate the autocorrelation value between the first half and the second half of each segment.

[0054] (4) Determine the segment length. If the autocorrelation of more than half of the segments is less than a certain threshold, then the segment length is determined to be n. Otherwise, increment the value of n and repeat step (2) until the segment length can be determined.

[0055] Furthermore, in step (1), the number of samples N should not be less than the length of the longest sequence in the sample set.

[0056] Furthermore, in step (2), the initial segment length n should not be less than half the length of the shortest sequence in the sample set. A sequence of length L can be divided into (L-n+1) segments, and there can be overlap between segments. The maximum number of overlapping data between two segments is (n-1).

[0057] Furthermore, in step (3), if the data of a segment is a(1), a(2), ..., a(n), then when n is odd, i.e. n = 2m + 1, the first half of the data is a(1), a(2), ..., a(m + 1), and the second half of the data is a(m + 1), a(m + 2), ..., a(n); when n is even, i.e. n = 2m, the first half of the data is a(1), a(2), ..., a(m), and the second half of the data is a(m + 1), a(m + 2), ..., a(n).

[0058] The correlation value C between the first half and the second half of the data is calculated as follows:

[0059]

[0060] Furthermore, in step (4), if among the M segments obtained in step (2), the correlation values ​​of (M / 2)+1 segments are all less than a certain threshold, then the segment length is determined to be n.

[0061] This invention utilizes the autocorrelation of radio signal segments to determine appropriate segment lengths, avoiding the need for multiple trials to manually select segment lengths. This saves time in designing radio signal detection and identification methods and can play an important role in the field of radio signal processing.

[0062] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the technical principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A data segmentation method for radio signal detection and identification, characterized in that, The method includes the following steps: S1. Collect a set of radio signal samples: Collect multiple representative radio signal sequence data to form a sample set; S2. Initial segmentation: Select an initial segment length n, and divide each sample in the sample set into multiple segments of length n; S3. Calculate the autocorrelation value of each segment: Calculate the autocorrelation value between the first half and the second half of each segment; S4. Determine the segment length: If the autocorrelation values ​​of more than half of the segments are less than a certain threshold, then determine the segment length as n; otherwise, increment the value of n and repeat step S2 until the segment length can be determined. in, In step S3, if the data of a segment is a(1), a(2), …, a(n), then when n is odd, i.e. n=2m+1, the data of the first half is a(1), a(2), …, a(m+1), and the data of the second half is a(m+1), a(m+2), …, a(n). In step S3, if the data of a segment is a(1), a(2), …, a(n), when n is an even number, i.e. n=2m, then the data of the first half segment is a(1), a(2), …, a(m), and the data of the second half segment is a(m+1), a(m+2), …, a(n); The correlation value C between the first half and the second half of the data is calculated as follows: In the formula, s(i) and t(i) are data segments of length k, s(i) is the first half and t(i) is the second half; when n is odd, k=m+1, and when n is even, k=m.

2. The data segmentation method for radio signal detection and identification as described in claim 1, characterized in that, In step S1, N valid radio signal sequence data are collected according to the needs of radio signal detection and identification to form a sample set.

3. The data segmentation method for radio signal detection and identification as described in claim 2, characterized in that, N is not less than the length of the longest sequence in the sample set.

4. The data segmentation method for radio signal detection and identification as described in any one of claims 1-3, characterized in that, In step S2, the initial segment length n is not less than 1 / 2 of the length of the shortest sequence in the sample set. A sequence of length L is divided into L-n+1 segments. The maximum number of overlapping data between two segments is n-1, and the final number of segments is M.

5. The data segmentation method for radio signal detection and identification as described in claim 4, characterized in that, In step S4, if (M / 2)+1 of the M segments have correlation values ​​that are all less than a certain threshold, then the segment length is determined to be n; otherwise, n is incremented to n+1, and step S2 is repeated until the segment length can be determined.

6. The data segmentation method for radio signal detection and identification as described in claim 5, characterized in that, The threshold is 0.

5.

7. The data segmentation method for radio signal detection and identification as described in claim 5, characterized in that, The segments of length n are input into the radio signal detection and recognition algorithm.

Images