An amplitude modulation digital baseband signal identification method and device

Abstract

The present application belongs to the technical field of integrated circuit design, and also relates to the field of digital signal processing, in particular to an amplitude modulation baseband signal processing and identification method and device, which are used to solve the problem of digital signal identification of amplitude modulation transmission. The processing process includes: quadrature signal equalization, quadrature signal merging, normalized correlation detection, signal peak value search, signal strength screening, and signal decision. The method and device can process the baseband signal converted from analog-digital after being demodulated by the radio frequency analog circuit in the radio equipment, and finally identify the digital coding content of the amplitude modulation signal. It is mainly applied to amplitude modulation radio equipment and near field communication field, and is responsible for the decoding of digital signal. Through this method and device, the coding content can be efficiently and accurately identified; it has strong anti-interference ability and anti-signal distortion ability; it has sensitive identification ability for small signals and certain signal-to-noise ratio gain for low-speed signals.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of integrated circuit design, and also relates to the field of digital signal processing, and particularly relates to an amplitude modulation digital baseband signal identification method and device, and more particularly relates to a digital signal detection method and a digital signal processing method. BACKGROUND

[0002] In the field of amplitude modulation radio (AM) and near field communication (NFC), the design of a receiver is a key link of the whole system. For an on-off keying (OOK) signal with non-100% amplitude modulation, the performance of coherent demodulation is better than that of envelope detection and other methods. In a typical coherent demodulation architecture, an analog radio frequency front-end circuit is responsible for sending a higher sampling rate baseband signal with I-Q two-way orthogonal multi-bit sampling depth to a digital circuit after mixing, amplifying, filtering, and analog-digital conversion of an antenna signal; the digital circuit restores the original baseband signal according to the two-way signal, and further decodes the communication content according to the communication protocol.

[0003] A patent application “Self-adaptive threshold adjusting method in non-contact smart card demodulation system” (application number: CN201310586002.9, publication number: CN103606001A) applied by Beijing University of Technology and a patent application “OOK demodulation circuit” (application number: CN202110980205.0, publication number: CN113691475A) applied by Hunan Maikesenwei Electronic Technology Co., Ltd. respectively disclose a method for identifying a radio frequency communication signal. Both the above two methods use a threshold comparison method for hard decision, and the method has the disadvantages of high requirement for the conversion accuracy of an analog circuit, being easily affected by a direct current bias, communication noise, and impedance mismatch of a receiving antenna, limited performance, and poor stability and adaptability.

[0004] A patent application “System and method for peak adaptive sampling demodulation in radio frequency transceiver” (application number: CN201980001676.3, publication number: CN110521127B) applied by Shenzhen Hi-Pot Technology Co., Ltd. discloses a demodulation identification method based on a carrier instantaneous peak value. The method has the disadvantages of requiring a multi-phase analog signal processing of an input signal, high requirement for analog design, and being easily affected by an integrated circuit manufacturing process.

[0005] At present, there are few solutions to demodulate a radio frequency amplitude modulation baseband signal by using a correlation detection-based decision technology, and there is a lack of a complete system solution. SUMMARY

[0006] The application provides an amplitude modulation digital baseband signal identification method and device to solve the problem of identifying the amplitude modulation transmission digital signal in a receiver system.

[0007] The method and device provided by the application have an internal structure comprising: a quadrature equalizer (101), a quadrature combiner (102), a correlation detector (104), a signal strength filter (103), a peak value searcher (105), and a signal decision device (106).

[0008] The quadrature equalizer (101) is used to correct the quadrature of the I-Q two-way signal. When the analog radio frequency circuit of the radio receiver demodulates the antenna signal, the sampling time of the I-Q two-way signal is usually not exactly 90 degrees relative to the carrier. If no equalization is performed, the correctness of the output result will be reduced. The quadrature equalizer (101) uses first-order linear correction. The input values of the two-way signal are I and Q, and the equalization parameters are a, b, c, and d. The output after equalization is I2=I*a+Q*b; Q2=I*c+Q*d; wherein the equalization parameters a, b, c, and d can be configured by a register.

[0009] The quadrature combiner (102) is used to combine the two-way quadrature signal into one-way for subsequent unified processing. It comprises multiple combiners and a selector, and the output of one combiner can be selected according to the application environment; the number and combination of the combiners are not limited. In order to adapt to the general application scenario, the quadrature combiner (102) comprises three combiners and a selector, which can be switched as needed, and the three combiners are respectively:

[0010] The two-way combiner (201) periodically selects the I-way or Q-way signal for direct output. The selection principle is to calculate the alternating current energy of the I-way and Q-way input signals respectively in a period of time, and select the one with larger alternating current energy for output.

[0011] The linear combiner (202) outputs the amplitude F=(I2*m+Q2*n) / (m+n) by multiplying the I-way and Q-way input signals by the coefficients m and n respectively, and then dividing the sum of the two coefficients. The two coefficients m and n are periodically and dynamically updated. The updating method is that the new coefficient is equal to the ratio of the alternating current energy of the input signal to the noise of the input signal.

[0012] Square root combiner (203) outputs F=(I2 2 ×m 2 +Q2 2 ×n 2 ) 1 / 2 / (m+n);m,n periodically dynamically updated. The updating method is that the new coefficient is equal to the ratio of the AC energy of the input signal to the noise of the input signal.

[0013] The one-or-the-other combiner (201) is the simplest and has the smallest area and power consumption in implementation; the linear combiner (202) has the ability to suppress single-path spur and can obtain more reliable results by combining two signals; the square root combiner (203) has larger area and power consumption and can adapt to the special phase in which the two phases of the quadrature demodulation are located at ±45 degrees of the carrier, at which time the amplitudes of the I-Q two signals are equal in size and opposite in direction, and the linear combiner (202) will fail.

[0014] In the correlation detector (104), the correlation coefficients of the current input signals R1~R n and the template waveforms M1~M n are obtained in real time through correlation operation, indicating the similarity of the input waveforms to the template model. The correlation detector (104) uses normalized correlation calculation and is not affected by the amplitude of the input signal, but only detects whether the shape of the input waveform is similar to the template, so it can detect very small signals and has high detection sensitivity. The correlation coefficient C=Σ(R i ×M i ) / ((Σ(R i -ΣR i / n) 2 ) 1 / 2 ×(ΣM i 2 ) 1 / 2 ); wherein the operation scale is related to n, and the value of n of the correlation detector can be selected according to the template length and the input signal rate to balance power consumption and performance. The larger the value of n, the more information of a communication symbol is interpreted, and the stronger the anti-interference ability is. In theory, the processing gain of 1.5 dB can be obtained for each doubling of the symbol length n, so the processing process has signal-to-noise ratio gain for low-speed signals with high sampling rate.

[0015] The peak searcher (105) searches for local maximum values by saving the correlation coefficient values of the last several times and outputs an indication signal. For continuous input signals, the local maximum values of the correlation coefficients usually represent the most accurate time of sending the corresponding symbol, so the peak searcher (105) can improve the phase accuracy of the communication symbol reception and facilitate further processing of the subsequent communication protocol processing module.

[0016] The signal strength filter (103) records the sum of the AC amplitudes of the input signal over a period of time. By comparing with a preset threshold, the signal strength filter (103) can avoid identifying the noise and interference of small amplitude as valid signals when there is no signal.

[0017] The signal decision maker (106) gives an identification mark when the peak searcher (105) searches for the correlation coefficient peak, the strength filter (103) confirms that the signal strength exceeds the threshold, and the correlation detector (104) calculates the correlation coefficient exceeding the threshold, indicating that the template waveform is identified.

[0018] Compared with the prior art, the present application has the following advantages:

[0019] Firstly, since the present application uses orthogonal signal input, compared with single-ended input, it is not affected by the mixing phase of the front-end analog circuit, and sufficient signal amplitude can always be obtained. Since the circuit size of the correlation detection is large and the cost is high, the present application uses two-way merging and then performs a correlation detection, which can significantly reduce the circuit cost.

[0020] Secondly, since the present application uses correlation detection for the identification of amplitude modulation signals, compared with the traditional method of simply comparing the signal amplitude with the threshold, the present application has better identification accuracy and adaptability. In particular, the correlation detection used by the present application is normalized, that is, it focuses on comparing the similarity of the shape of the input signal and the shape of the target waveform, and can not be affected by the amplitude and DC bias of the input waveform, so the identification sensitivity of the small signal is similar to that of the large signal.

[0021] Thirdly, since the final decision condition of the present application is multi-dimensional, that is, the signal shape is similar, the signal strength reaches the preset value, and the signal waveform similarity reaches the peak, compared with the method of simply calculating the correlation value and making a decision, it can effectively filter out noise and interference and give a decision at the right time, providing a better signal source for subsequent protocol processing. BRIEF DESCRIPTION OF DRAWINGS

[0022] Figure 1 The figure is the structure diagram of the device and the flow chart of the method.

[0023] Figure 2 The figure is the structure diagram of the orthogonal merger (102) in the device.

[0024] Figure 3 The figure is the input / output simulation result diagram of the device and the method, the vertical axis is the amplitude, and the horizontal axis is the time. DETAILED DESCRIPTION

[0025] Figure 1The method and device provided by the application comprise the following modules: a quadrature equalizer (101), a quadrature combiner (102), a correlation detector (104), a signal strength filter (103), a peak searcher (105), and a signal decision device (106). The input two-way quadrature signals first enter the quadrature equalizer (101) for equalization, and then are sent to the quadrature combiner (102) to be combined into one-way signals. The combined signals are sent to the correlation detector (104) and the signal strength filter (103) respectively. The correlation detector (104) calculates the correlation coefficient, and the signal strength filter (103) calculates the amplitude strength of the signals. The calculated correlation coefficient is sent to the peak searcher (105) to calculate the peak time of the correlation coefficient and give a flag signal. The signal decision device (106) judges whether to identify the target symbol in combination with the correlation coefficient, the amplitude strength, and the peak flag.

[0026] For the quadrature combiner (102), an embodiment using three combining modes is provided as shown in Figure 2 . The three combining modes are two-choice mode, linear mode, and square root mode. However, in the specific implementation, the three modes are not limited, and only one or two modes can be implemented.

[0027] The implementation method of the quadrature equalizer (101) is that the input values of the two-way signals are I and Q, the equalization parameters are a, b, c, and d, the output after equalization is I2=I×a+Q×b; Q2=I×c+Q×d; and the equalization parameters a, b, c, and d can be configured by a register.

[0028] The two-choice combiner (201) in the quadrature combiner (102) periodically switches to the signal with larger alternating current amplitude in the I-Q two-way input. One calculation method of the alternating current amplitude is the sum of the input values minus the direct current bias, that is, the alternating current amplitude W=(∑(X i -X M ) 2 ) 1 / 2 . For example, the I-way input sequence is X1-X4, the average value X M =(X1+X2+X3+X4) / 4, and the alternating current amplitude W=((X1-X M ) 2 +(X2-X M ) 2 +(X3-X M ) 2 +(X4-X M ) 2 ) 1 / 2 .

[0029] The linear combiner (202) in the quadrature combiner (102) periodically calculates the ratio (e.g., m, n) of the AC amplitude and noise in the I and Q inputs, i.e., the output F = (I² × m + Q² × n) / (m + n). The calculation method for the AC amplitude can be the same as that for the two-to-one combiner (201); one method for calculating the noise is the difference between the sums of the odd and even input values. For example, if the I input sequence is X1 to X4, then the noise N = X2 + X4 - X1 - X3.

[0030] The square root combiner (203) in the quadrature combiner (102) periodically calculates the ratio (e.g., m, n) of the AC amplitude to noise in the I and Q inputs, i.e., the output F = (I2) / (Q2). 2 ×m 2 +Q2 2 ×n 2 ) 1 / 2 / (m+n). The calculation of noise and AC amplitude is the same as that of the linear combiner (202), and the circuit can be reused.

[0031] The intensity filter (103) is implemented by calculating the mean square error of the signal over a period of time, i.e., P = ∑(A i -A M ) 2 For example, a calculation is performed every four cycles, and the average value of the four signals A1 to A4 is A. M Then the signal amplitude intensity P = (A1 - A M ) 2 +(A2-A M ) 2 +(A3-A M ) 2 +(A4-A M ) 2 .

[0032] The implementation method of the correlation detector (104) is to calculate the correlation coefficient C = Σ(R i ×M i ) / ((Σ(R i -ΣR i / n) 2 ) 1 / 2 ×(ΣM i 2 ) 1 / 2 ), where R is the sliding history value of the input waveform, and M is the template waveform, with R and M having the same length. Specifically, if the template waveform is the rising edge of a square wave, then M1~M n / 2 =-1, M n / 2+1 ~M n / 2 If the value is 1, the above formula can be simplified to the correlation coefficient C = (Σ nn / 2+1 R i -Σ1 n / 2 R i ) / ((Σ(R i -ΣR i / n) 2 ) 1 / 2 ×(n) 1 / 2 This eliminates a large number of multiplication operations, reducing the cost and power consumption of digital circuits. Specifically, if the template waveform is a square wave with a falling edge, then M1~M n / 2 M is 1. n / 2+1 ~M n / 2 Since the value is -1, the above formula can be simplified to the correlation coefficient C = (Σ1) / (Σ1) n / 2 R i -Σ n n / 2+1 R i ) / ((Σ(R i -ΣR i / n) 2 ) 1 / 2 ×(n) 1 / 2 ).

[0033] For example, the calculation method for detecting the rising edge of a standard square wave using a correlation detector (104) with a template length of 4 is as follows: Let the input signal be R1~R4, and the template waveform be the rising edge, i.e., M1~M4 is {-1, -1, 1, 1}, then the average value of the input value R M = (R1+R2+R3+R4) / 4, correlation coefficient C = (R3+R4-R1-R2) / ((R1-R4) / 4) M ) 2 +(R2-R M ) 2 +(R3-R M ) 2 +(R4-R M ) 2 ) 1 / 2 / 2.

[0034] The implementation of the correlation detector (104) can include multiple sets of template waveforms for simultaneous comparison. Two preferred general templates are the rising edge and falling edge of a square wave. These two templates can effectively restore the signal to a single-bit level value, thereby achieving waveform shaping.

[0035] The peak searcher (105) is implemented as follows: for the input values ​​of the sliding record, such as S1, S2, and S3, if S2 ≥ S1 and S2 ≥ S3, the maximum value is determined to have been found; if S2 ≤ S1 and S2 ≤ S3, the minimum value is determined to have been found. It should be noted that the peak search length includes, but is not limited to, the comparison of three historical values, and can also include the comparison of more historical values.

[0036] The implementation method of the signal decision device (106) is that the peak searcher (105) searches for the correlation peak, the strength filter (103) calculates the signal strength exceeding the threshold value, and the correlation detector (104) calculates the correlation coefficient exceeding the threshold value, and then the identification mark is given to indicate that the template waveform is identified. The above-mentioned threshold values can be set through registers respectively.

[0037] For further illustration, an overall embodiment for decoding the Manchester coded signal is given, in which the above-mentioned three methods are implemented in the quadrature combiner (102), the template length and the history record length in the correlation detector (104) are both 64 points, and the template waveform is 32 zeros and 32 ones (i.e. 0000000000000000000000000000000011111111111111111111111111111111). The input of the TYF protocol in the NFC protocol at the speed of 424K is decoded and simulated by using the embodiment, and the result is as shown in Figure 3 It can be seen from Figure 3 that the decoding is correct, the waveform is regular and stable, and the input burr is not sensitive.

[0038] The above-mentioned embodiments are only examples for clearly illustrating the present application, and are not the limitation of the embodiments. Other different forms of changes or variations can be made on the basis of the above-mentioned description by those skilled in the art. All the embodiments do not need to be exhausted here, and the obvious changes or variations derived therefrom are still within the protection scope of the present application.

Claims

1. A signal processing and recognition apparatus, characterized by, The application relates to a quadrature equalizer and a quadrature combiner, and belongs to the field of signal processing. The quadrature equalizer (101) is used for compensating quadrature imbalance of I and Q input signals to generate two quadrature signals. The quadrature combiner (102) is used for combining the two quadrature signals equalized by the quadrature equalizer (101) into a digital signal. The signal strength filter (103) is used for calculating the swing strength of the digital signal generated by the quadrature combiner (102). The correlation detector (104) is used for calculating the correlation coefficient of the digital signal generated by the quadrature combiner (102) and a target template. The peak value searcher (105) is used for detecting the local maximum or minimum value of the correlation coefficient output by the correlation detector (104). The signal decision device (106) is used for giving a level decision or symbol classification of a current signal according to the results of the signal strength filter (103), the correlation detector (104) and the peak value searcher (105). When the peak value searcher (105) searches the correlation coefficient peak value, the signal strength filter (103) calculates the signal strength exceeding a threshold value, and the correlation detector (104) calculates the correlation coefficient exceeding a threshold value, the signal decision device (106) gives an identification mark, indicating that the template waveform is identified.

2. A signal processing and recognition apparatus according to claim 1, characterized in that The quadrature equalizer (101) adopts linear proportional cross combination, the amplitude I2 of the I input signal after equalization is the amplitude of the I and Q signals multiplied by the coefficients a and b respectively, and I2=I*a+Q*b; the amplitude Q2 of the Q input signal after equalization is the amplitude of the I and Q signals multiplied by the coefficients c and d respectively, and Q2=I*c+Q*d.

3. A signal processing and recognition apparatus according to claim 2, characterized in that The quadrature combiner (102) is responsible for combining the I and Q quadrature signals into a signal, and internally includes N parallel implemented combiners and a selector, which can output the combination result of a combiner according to the application environment; the number and combination of the implemented combiners include: The two-choice combiner is characterized by periodically selecting the I or Q signal to be directly output, and the selection is based on the calculation of the AC energy of the I and Q input signals in a period of time, and the signal with larger AC energy is output; The linear combiner is characterized by outputting the amplitude of the I and Q input signals multiplied by the coefficients m and n respectively, and then dividing the sum of the two coefficients, that is, F=(I2*m+Q2*n) / (m+n); wherein the two coefficients are periodically updated, and the updating method is that the new coefficient is equal to the ratio of the AC energy of the input signal to the differential noise of the input signal; Square root combiner characterized in that the output amplitude is F=(I2 2 ×m 2 +Q2 2 ×n 2 ) 1 / 2 / (m+n); where the two coefficients are periodically updated, the updating method is that the new coefficients are equal to the ratio of the AC energy of the input signal to the differential noise of the input signal.

4. A signal processing and recognition apparatus according to claim 1, characterized in that The signal strength filter (103) records the AC amplitude of the input signal in a period of time in real time, and calculates the swing strength of the signal in real time.

5. A signal processing and recognition apparatus according to claim 1, wherein The correlation detector (104) records the input signal data R1~R n in real time with a sliding length of n n , and performs convolution operation to obtain a correlation value; calculates the square root of the product of the autocorrelation value of the template and the autocorrelation value of the signal, and divides the two to obtain the correlation coefficient C of the signal, C=Σ(R i ×M i ) / ((Σ(R i -ΣR i / n) 2 ) 1 / 2 ×(ΣM i 2 ) 1 / 2 ); the target template sequence can be multiple, and the correlation coefficients of each can be calculated simultaneously.

6. A signal processing and recognition apparatus according to claim 1, wherein The peak value searcher (105) searches the local maximum or minimum value by saving the correlation coefficient values in recent times.

Images