Frequency offset estimation method of FSK (Frequency Shift Keying) signal

Abstract

The invention discloses a frequency offset estimation method of an FSK (Frequency Shift Keying) signal, and belongs to the technical field of frequency estimation. The method comprises the following steps: firstly, converting a signal on a time domain to a frequency domain through sliding discrete Fourier transform for analysis and processing, then calculating power near a frequency point position, then searching a maximum power value, judging whether the maximum power value is greater than a threshold, and finally comparing a frequency position corresponding to the power value greater than the threshold with an original position of the frequency point, and calculating the frequency offset. The method is suitable for FSK signals with different modulation orders, can be used in an FDMA system for processing continuous signals, and has the advantages of wide application range and low complexity.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of frequency estimation technology in wireless communication, in particular to a frequency offset estimation method of FSK signal, which can be used in frequency division multiple access (FDMA) wireless communication system using FSK modulation mode. BACKGROUND

[0002] Frequency shift keying (FSK) modulation signal has the characteristics of simple generation mode, strong anti-noise ability and anti-frequency offset ability, low engineering implementation complexity, and can rely on signal design to improve anti-frequency offset ability, but simply increasing the interval of each frequency point without processing and correcting frequency offset will affect the demodulation performance. In wireless communication system, frequency deviation inevitably exists in digital receiver. In order to improve the demodulation performance of the receiving end and make it work effectively and reliably, frequency estimation technology plays an important role in communication system. In frequency division multiple access communication system, one of the key steps for the receiving end to correctly demodulate is frequency estimation.

[0003] At present, the frequency estimation algorithms realized by FPGA include cycle accumulation estimation method, AR model square power spectrum method, etc. The cycle accumulation estimation method estimates frequency according to the periodic characteristics of signal change with time, but this method is not suitable for carrier frequency estimation of continuous phase FSK. The AR model square power spectrum method is suitable for continuous phase FSK and non-continuous phase FSK signal, but the calculation is complex and not conducive to engineering implementation. Therefore, it is necessary to study the frequency estimation technology of FSK modulation mode in FDMA wireless communication system. SUMMARY

[0004] The purpose of the present application is to overcome the shortcomings of the prior art, and to provide a frequency offset estimation method of FSK signal, which can use sliding FFT to perform power judgment on M useful signals with the same subcarrier interval, extract frequency difference value, and estimate frequency offset.

[0005] The technical scheme adopted by the present application is:

[0006] A frequency offset estimation method of FSK signal is used to estimate the frequency offset of FSK signal at the signal receiving end, the FSK signal has M subcarriers, and the signal receiving end knows the original position of each subcarrier corresponding frequency point in advance; the method comprises the following steps:

[0007] Step 1: decimating the FSK signal collected by the digital-analog converter;

[0008] Step 2: performing sliding FFT transformation on the decimated signal, converting the signal in time domain to frequency domain, and obtaining the signal spectrum;

[0009] Step 3: Find the original position of the frequency point corresponding to a subcarrier on the signal spectrum, select a neighborhood range of the original position, and calculate the power of each frequency in the neighborhood range;

[0010] Step 4: Compare the maximum power within the neighborhood with the threshold. If it exceeds the threshold, the frequency corresponding to the maximum power is considered to be the frequency offset position of the corresponding subcarrier frequency point. If it does not exceed the threshold, continue to receive new FSK signals and repeat steps 1 to 4.

[0011] Step 5: Compare the frequency offset position with the original position, calculate the frequency difference, and obtain a frequency offset estimate.

[0012] Furthermore, when performing the sliding FFT transformation in step 2, the interval between frequencies is greater than the frequency resolution F, F=fs / N, where fs is the sampling clock and N is the FFT length.

[0013] Furthermore, the neighborhood range is (f c -f d ,f c +f d ), where f c is the original position, f d It is half the interval between frequencies during sliding FFT transformation.

[0014] Furthermore, the threshold is four to eight times the non-peak mean value of the signal power within the neighborhood after FFT transformation.

[0015] The present invention has the following beneficial effects:

[0016] (1) The present invention is applicable to frequency deviation estimation of binary and multi-ary FSK signals;

[0017] (2) The present invention is not limited to frequency hopping anti-interference communication systems, but can also be used in FDMA systems that process continuous signals, and has a wide range of applications;

[0018] (3) The present invention adopts sliding FFT, which has low computational complexity and is easy to implement in engineering. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 The diagram is a schematic diagram showing the principle of a frequency deviation estimation method for an FSK signal.

[0020] Figure 2 This is a comparison chart of the FFT of the 4FSK original signal and the frequency shifted signal.

[0021] Figure 3 This is a schematic diagram of the spectrum lines of each frequency point of 4FSK.

[0022] Figure 4This is a schematic diagram of the frequency difference of the 4FSK signal before and after the frequency deviation. DETAILED DESCRIPTION

[0023] The present invention is described in further detail below.

[0024] A frequency offset estimation method for FSK signals is proposed. The method first converts the signal in the time domain to the frequency domain through sliding discrete Fourier transform for analysis and processing. Then, the power of the corresponding spectrum position is calculated respectively. Then, the maximum power of the spectrum line is searched within a frequency point range. The maximum power value is compared with a threshold. Finally, the position with a power value greater than the threshold is compared with the original frequency point position, and the frequency difference is calculated to achieve frequency offset estimation.

[0025] like Figure 1 As shown, the method specifically includes the following steps:

[0026] (1) Extract the FSK signal sampled by the digital-to-analog converter: the extraction coefficient can be flexibly changed according to the sampling clock and symbol rate, and the interval between frequencies is 2f d It only needs to be greater than the frequency resolution F, where F = fs / N, fs is the sampling clock, and N is the FFT length.

[0027] The following is a detailed explanation using the sampling rate of 80M, the symbol rate of 100K, the frequency component interval of 200K, and the 4FSK modulation method as an example (not limited to this sampling rate, symbol rate, frequency component interval, and modulation method). The frequency components are ±100K and ±300K respectively. Considering the implementation complexity and the accuracy of the sliding FFT, the decimation factor is set to 80. The sampling rate after decimation is 1M. In matlab, the data volume is simplified, the sampling frequency is set to 1000Hz, the symbol rate is 100sps, the frequency offset reception is manually performed, the modulation order is set to 4, and the result of the 1000-point FFT is as follows: Figure 2 shown.

[0028] (2) Perform sliding FFT operation on the extracted signal:

[0029] The FFT (Fast Fourier Transform) is used to determine the frequency domain. The sliding FFT operation is performed on each sample point, and the FFT is calculated once for each sliding sample point. The FFT converts the signal in the time domain to the frequency domain for analysis and processing.

[0030] (3) Obtain the maximum power within the neighborhood of the first frequency point:

[0031] Since the spectrum of the FSK signal on M subcarrier frequencies has discrete peaks and the frequency intervals are equal, the frequency can be estimated by searching for the position of the peaks. Taking the 4FSK signal as an example, Figure 3 As shown, take fc The frequency point of the position is the target frequency estimation signal, at f c -f d and f c +f d Search for the frequency peak position between c Indicates the original position of the first frequency point.

[0032] (4) Threshold decision for peak power:

[0033] Determine whether the highest peak value is greater than the threshold. If it is greater than the threshold, the frequency point is considered to be the original f c The true position of the spectrum line is f. Otherwise, continue to receive new signals and repeat the above process.

[0034] (5) Compare the power peak position with the original frequency point position and calculate the frequency deviation:

[0035] Calculate Δf=ff c , Δf is the frequency deviation.

[0036] Taking the FFT length of 256 points as an example, after the extracted signal is subjected to sliding FFT operation, the theoretical frequency deviation value is 20*256 / 1000=5.12. Figure 4 This is a comparison chart of the frequency difference of 4FSK before and after the frequency deviation. After processing by the method of the present invention, the calculated frequency deviation is 5, and the frequency error is:

[0037] (5.12-5) / 5.12=2.3%,

[0038] It can be considered that the frequency offset calculated by the present invention is correct.

[0039] The present invention is applicable to FSK signals of different modulation orders and can be used in an FDMA system for processing continuous signals. It has a wide application range. At the same time, the present invention is based on sliding FFT processing and has the advantage of low complexity.

Claims

1. A method for estimating the frequency offset of an FSK signal, for estimating the frequency offset of an FSK signal at a signal receiving end, wherein the FSK signal has M subcarriers and the signal receiving end knows in advance the original position of the frequency point corresponding to each subcarrier; characterized in that: The following steps are involved: Step 1, extracting the FSK signal collected by the digital-to-analog converter; Step 2: Perform a sliding FFT transform on the extracted signal to convert the signal in the time domain to the frequency domain to obtain the signal spectrum; Step 3: Find the original position of the frequency point corresponding to a subcarrier on the signal spectrum, select a neighborhood range of the original position, and calculate the power of each frequency in the neighborhood range; Step 4: Compare the maximum power within the neighborhood with the threshold. If it exceeds the threshold, the frequency corresponding to the maximum power is considered to be the frequency offset position of the corresponding subcarrier frequency point. If it does not exceed the threshold, continue to receive new FSK signals and repeat steps 1 to 4. Step 5: Compare the frequency offset position with the original position, calculate the frequency difference, and obtain a frequency offset estimate.

2. The method for estimating frequency deviation of an FSK signal according to claim 1, wherein: When performing the sliding FFT transformation in step 2, the interval between frequencies is greater than the frequency resolution F, F=fs / N, where fs is the sampling clock and N is the FFT length.

3. The method for estimating frequency deviation of an FSK signal according to claim 1, wherein: The neighborhood range is (f c -f d ,f c +f d ), where f c is the original position, f d It is half the interval between frequencies during sliding FFT transformation.

4. The method for estimating frequency deviation of an FSK signal according to claim 1, wherein: The threshold is four to eight times the non-peak value of the signal power within the neighborhood after FFT transformation.

Images