A search algorithm for fast synchronization of chirp signals

Abstract

The application relates to the technical field of wireless communication, in particular to a search algorithm for quickly synchronizing a Chirp signal, which comprises the following steps: a maximum peak-to-average power ratio corresponding sample point interval is pre-judged by changing a sliding step; in order to not miss any sample point, a point-by-point sliding method is still adopted in the pre-judged interval, so that the accuracy of the search is ensured; in the process of determining the pre-judged interval, the increase of the sliding step not only reduces the sliding times, effectively reduces the calculation amount of searching the Chirp signal, but also can more quickly find the pre-judged interval, improves the search speed, effectively reduces the consumption of system resources, enables a general DSP chip to undertake the search of the Chirp signal, and reduces the system cost.

Description

Technical Field

[0001] This invention relates to the field of wireless communication technology, and in particular to a search algorithm for fast synchronization of Chirp signals. Background Technology

[0002] The chirp signal is a crucial signal for initial synchronization between the satellite and the mobile terminal in a satellite mobile communication system. It consists of two parts: an upscan (frequency increasing linearly with time) and a downscan (frequency decreasing linearly with time). The mobile terminal determines the start position of the received chirp signal to obtain the downlink signal's time synchronization information.

[0003] Typically, mobile terminals store sample information of local chirp signals. They perform sliding correlation calculations between the received airborne signal and the local chirp samples, and then use a Fast Fourier Transform (FFT) to obtain the sample values ​​for each sliding step in the frequency domain. The ratio of peak power to the average sample power is called the peak-to-average power ratio (PAPR). Whether the PAPR is greater than a threshold is used as the criterion for the presence or absence of a chirp signal. The starting position and frequency offset information of the chirp signal are calculated based on the sample position corresponding to the maximum PAPR.

[0004] To accurately find the maximum peak-to-average power ratio (PAPR), a sliding step of one sample point is typically used to avoid missing samples corresponding to the maximum PAPR and causing misjudgments. However, the number of received air signal samples is large each time, for example, 30720 air signal samples. Using 1024 as the typical Fast Fourier Transform (FFT) points, each sliding step requires complex multiplication of 1024 points, FFT, sample power calculation, and searching for the maximum PAPR, consuming significant system resources. This process needs to be repeated 30720 times. Even with the help of external hardware accelerators, commonly used general-purpose DSP chips struggle to meet real-time requirements.

[0005] Therefore, the urgent problem to be solved is how to improve the search algorithm, reduce the number of operations, and enable general-purpose DSP chips to perform search tasks.

[0006] Therefore, we propose a fast search algorithm for synchronizing Chirp signals to solve the existing problems. Summary of the Invention

[0007] The purpose of this invention is to provide a fast search algorithm for synchronizing Chirp signals to solve the problems mentioned in the background art.

[0008] To achieve the above objectives, the present invention provides the following technical solution: a search algorithm for fast synchronization of Chirp signals, comprising: S1: Multiply the input sequence with the local sequence to obtain the correlation sequence, and then obtain the frequency domain signal through fast Fourier transform; S2: Set the first sliding step size, find the maximum power and the second largest power of the frequency domain signal within the frequency domain range using the first sliding step size, and calculate the time offset; S3: After repeating the operation steps of S2 several times, calculate the mean of the time offset based on the time offset obtained in S2; S4: The second sliding step size is obtained based on the mean time offset obtained in S3. The input sequence is then directly slid into the prediction interval using the second sliding step size. S5: Set the third sliding step size, and slide the input sequence point by point through the third sliding step size until the maximum peak-to-average power ratio is found; S6: Calculate the starting position and frequency offset information of the Chirp signal by using the sample point position corresponding to the maximum peak-to-average power ratio.

[0009] Furthermore, in step S1, the input sequence is the Chirp signal received by the mobile terminal from the satellite.

[0010] Further, in step S1, the local sequence is a Chirp signal pre-stored locally on the mobile terminal.

[0011] Furthermore, in step S1, several zeros are added before the input sequence.

[0012] Furthermore, in step S2, the first sliding step size is greater than or equal to 10.

[0013] Furthermore, in step S2, the time offset is the interval between the points corresponding to the maximum power and the second-highest power.

[0014] Furthermore, in step S3, when repeating the operation steps of S2 several times, if the peak-to-average power ratio corresponding to the maximum power and the second-highest power is less than or equal to a set threshold, the time offset obtained this time is discarded; if the peak-to-average power ratio corresponding to the maximum power and the second-highest power is greater than the set threshold, the time offset obtained this time is retained; the average time offset is obtained based on the retained time offset.

[0015] Furthermore, in step S4, the second sliding step size is the difference between the product of the time-biased mean and a specific constant and the number of points in the predicted interval.

[0016] Furthermore, the specific constant is proportional to the number of points in the Fast Fourier Transform.

[0017] Furthermore, in step S5, the third sliding step size is 1.

[0018] Compared with the prior art, the beneficial effects of the present invention are: This invention predicts the interval containing the sample point corresponding to the maximum peak-to-average power ratio by changing the sliding step size. To ensure no sample points are missed, a point-by-point sliding method is still used within the predicted interval, guaranteeing search accuracy. In determining the predicted interval, increasing the sliding step size not only reduces the number of sliding operations, effectively reducing the computational load of searching for the chirp signal, but also allows for faster finding of the predicted interval, improving search speed. This effectively reduces system resource consumption, enabling general-purpose DSP chips to handle chirp signal searches and lowering system costs. Attached Figure Description

[0019] Figure 1 This is a flowchart illustrating the overall steps of the present invention; Figure 2 This is a flowchart of the calculation steps of the present invention. Detailed Implementation

[0020] The technical solution of the present invention will be further described below with reference to the accompanying drawings and specific embodiments. Example 1

[0021] like Figure 1 and Figure 2 As shown, the overall steps of a fast Chirp signal synchronization search algorithm include: S1: Multiply the input sequence with the local sequence to obtain the correlation sequence, and then obtain the frequency domain signal through fast Fourier transform. The input sequence is the Chirp signal received by the mobile terminal from the satellite, and the local sequence is the Chirp signal pre-stored locally by the mobile terminal. Several zeros are added before the input sequence to avoid the omission of target samples. S2: Set the first sliding step size, find the maximum power and the second largest power of the frequency domain signal within the frequency domain range using the first sliding step size, and calculate the time offset. The first sliding step size is greater than or equal to 10, thereby reducing the number of sliding steps. The time offset is the interval between the corresponding points of the maximum power and the second largest power. S3: After repeating the operation steps of S2 several times, calculate the mean time offset based on the time offset obtained in S2. When repeating the operation steps of S2 several times, if the peak-to-average power ratio corresponding to the maximum power and the second-highest power is less than or equal to a set threshold, discard the time offset obtained in this instance; if the peak-to-average power ratio corresponding to the maximum power and the second-highest power is greater than the set threshold, retain the time offset obtained in this instance; calculate the mean time offset based on the retained time offset to improve the accuracy of the calculation. S4: The second sliding step size is obtained based on the time offset obtained in S3. The input sequence is directly slid to the prediction interval through the second sliding step size. The second sliding step size is the difference between the product of the time offset and a specific constant and the number of points in the prediction interval. The specific constant has a proportional relationship with the number of points in the fast Fourier transform. S5: Set the third sliding step size, and slide the input sequence point by point through the third sliding step size until the maximum peak-to-average power ratio is found. The third sliding step size is 1 to avoid the problem of missing target samples. S6: Calculate the starting position and frequency offset information of the Chirp signal by using the sample point position corresponding to the maximum peak-to-average power ratio.

[0022] The working principle of the fast synchronization Chirp signal search algorithm based on Example 1 is as follows: Let the input sequence be r(p), the local sequence be h(m), the correlation sequence be s(t), and the frequency domain signal be v(f), then s(t)=r(p)*h(m) v(f) = FFT[s(t)] Where FFT[ ] represents Fast Fourier Transform.

[0023] Let the first sliding step size be , ≥10, find the maximum power pow(v(f) in the frequency domain). ) and second-highest power POW ( And calculate the corresponding peak-to-average power ratio (PAPR). ) and PAPR ( ).

[0024] The formula for calculating time offset td is as follows: td=|pos( )-pos( )| Where, pos( ) represents the point corresponding to the maximum power, pos( () represents the location corresponding to the second highest power.

[0025] The threshold value is set to threshold. Correlation calculations, fast Fourier transforms, and finding the pow(value) of v(f) are performed repeatedly on r(p) and h(m). ) and pow( When PAPR( )≤threshold or PAPR( If PAPR( )≤threshold, then discard the td obtained this time. ) > threshold and PAPR( If td > threshold, then retain the td obtained in this calculation. Calculate the partial mean tdmean based on the retained td.

[0026] Since this application is aimed at high-orbit narrowband language communication, high-orbit satellites are less affected by Doppler frequency shift in narrowband communication compared to geostationary satellites, and the time offset td is calculated as an average of multiple times, so it basically does not require preprocessing. It has been verified to be usable through actual communication.

[0027] Furthermore, the PAPR threshold setting needs to be based on actual conditions, considering a balance between sensitivity and accuracy. The peak PAPR of a bimodal structure decreases as the distance between the peaks increases. Chirp signals under various SNR conditions can be generated using a comprehensive measuring instrument to determine the balance between the bimodal distance and PAPR threshold.

[0028] Let the second sliding step size be ,but =tdmean*un Where u is a specific constant and n is the number of points in the predicted interval.

[0029] Furthermore, the value of n cannot be too large, and the maximum value cannot exceed tdmean*u. The center point of the actual pre-sliding range is tdmean*u. The maximum value is found by sliding n points forward and backward. n can be 1 / 2 / 3 as needed, but it cannot be greater than tdmean*u.

[0030] Let the third sliding step size be =1, through Slide r(p) directly to the predicted interval, and then... Slide r(p) point by point until the maximum peak-to-average power ratio PAPR is found. ), through PAPR ( The corresponding sample point position pos( Calculate the starting position and frequency offset information of the Chirp signal.

[0031] Wherein, if the target sample point is at the starting position of the input sequence r(p), the first sliding step size is... If the value is too large, it will cause the omission of target samples. Therefore, zeros need to be added before r(p) to prevent the omission of target samples and to recalculate.

[0032] There is a possibility of omissions when sample points are at the end of the sequence. The high-orbit chirp signal has a total of 480 sample points. Within the 30720 sample points captured in a single capture, the complete chirp signal is guaranteed, with the most extreme case being the last 480 chirp sample points. Because the local sequence's 1024 chirp signal points are in the first 480 points, overlap may occur. This can be addressed by adding 1024-480 points at the end of the sequence, or by placing the local sequence's chirp sample points at the end of the sequence. In practical applications, the impact is minimal due to the use of multiple captures of the chirp starting point and averaging, thus improving positioning accuracy.

[0033] If the target sample point is at the end of the input sequence r(p), and the length of the input sequence is L, then the number of slides is... Then you can find the maximum power POW ( ) and second-highest power POW ( The number of slides is This represents the maximum computational load.

[0034] For the zero-padding case, p zeros are added to the beginning of the sequence, and 544 zeros are added to the end. The maximum number of computations is... The zero-padding length *p* is determined based on the first sliding step size, ensuring that the chirp data is after two to three sliding points. For example, if the first sliding step size is 16, the padding length is 64 zeros. In the formula for calculating the number of slides, the number of zeros at the beginning and end can be adjusted to ensure the calculated number of slides is an integer. The 544 zeros at the end are the minimum number; to ensure the number of slides is an integer, this number can be increased appropriately.

[0035] The above specific embodiments are merely several preferred embodiments of the present invention. Based on the technical solutions of the present invention and the relevant teachings of the above embodiments, those skilled in the art can make various alternative improvements and combinations to the above specific embodiments.

Claims

1. A search algorithm for fast synchronization of Chirp signals, characterized in that, include: S1: Multiply the input sequence with the local sequence to obtain the correlation sequence, and then obtain the frequency domain signal through fast Fourier transform; where the input sequence is the Chirp signal received by the mobile terminal from the satellite, the local sequence is the Chirp signal pre-stored locally by the mobile terminal, and several zeros are added before the input sequence. S2: Set a first sliding step size, find the maximum power and the second largest power of the frequency domain signal within the frequency domain range using the first sliding step size, and calculate the time offset; wherein, the first sliding step size is greater than or equal to 10, and the time offset is the interval between the corresponding points of the maximum power and the second largest power; S3: After repeating the operation steps of S2 several times, calculate the average time offset based on the time offset obtained in S2; wherein, when repeating the operation steps of S2 several times, if the peak-to-average power ratio corresponding to the maximum power and the second largest power is less than or equal to a set threshold, then discard the time offset obtained in this instance; if the peak-to-average power ratio corresponding to the maximum power and the second largest power is greater than the set threshold, then retain the time offset obtained in this instance; calculate the average time offset based on the retained time offset. S4: The second sliding step size is obtained based on the mean time offset obtained in S3. The input sequence is then directly slid into the prediction interval using the second sliding step size. S5: Set the third sliding step size, and slide the input sequence point by point through the third sliding step size until the maximum peak-to-average power ratio is found; S6: Calculate the starting position and frequency offset information of the Chirp signal by the sampling point position corresponding to the maximum peak-to-average power ratio; Let the input sequence be r(p), the local sequence be h(m), the correlation sequence be s(t), and the frequency domain signal be v(f), then s(t) = r(p)·h(m) and v(f) = FFT[s(t)], where FFT[ ] represents Fast Fourier Transform; Let the first sliding step size be step1, step1 ≥ 10. Find the maximum power pow(f1) and the second largest power pow(f2) of v(f) in the frequency domain and calculate the corresponding peak-to-average power ratio PAPR(f1) and PAPR(f2). The formula for calculating the time offset td is: td = |pos(f1)-pos(f2)|, where pos(f1) is the point corresponding to the maximum power and pos(f2) is the point corresponding to the second maximum power. The threshold is set to threshold. When performing related calculations, fast Fourier transforms, and searching for pow(f1) and pow(f2) of v(f) repeatedly, if PAPR(f1) ≤ threshold or PAPR(f2) ≤ threshold, the td obtained in this case is discarded. If PAPR(f1) > threshold and PAPR(f2) > threshold, the td obtained in this case is retained. The time-partial mean tdmean is then calculated based on the retained td.

2. The search algorithm for fast synchronization Chirp signals according to claim 1, characterized in that: In step S4, the second sliding step size is the difference between the product of the time-biased mean and a specific constant and the number of points in the predicted interval.

3. The search algorithm for fast synchronization Chirp signals according to claim 2, characterized in that: The specific constant is proportional to the number of points in the Fast Fourier Transform.

4. The search algorithm for fast synchronization Chirp signals according to claim 1, characterized in that: In step S5, the third sliding step size is 1.

Images