A communication baseband signal generator with continuously variable symbol rate

Abstract

The present invention belongs to the field of spacecraft measurement and control communication technology, and in particular relates to a communication baseband signal generator with a continuously variable symbol rate, comprising: a shaping filter for baseband shaping an IQ symbol sequence to convert the IQ symbol sequence into a baseband waveform that meets the requirements; an interpolation controller for determining the interpolation points and fractional intervals through the input symbol rate control word; and a fractional interpolator for applying fractional interpolation technology to the baseband waveform at the interpolation point to generate a baseband waveform at the DA sampling frequency. The present invention can adapt to any baseband shaping filter, and baseband waveforms with different roll-off coefficients can generate baseband waveforms with any symbol rate. The symbol rate step accuracy generated by the present invention can be easily controlled below 1 Hz, and can even achieve an accuracy better than 10 ‑6 The Hz step accuracy can well meet the needs of fine adjustment of symbol rate in various wireless communication scenarios such as spacecraft measurement and control communications, satellite communications, etc.

Description

Technical Field

[0001] The present invention belongs to the technical fields related to satellite communication, spacecraft measurement and control communication, and wireless communication, and in particular relates to a communication baseband signal generator with continuously variable symbol rate. Background Art

[0002] In wireless communications, such as satellite communications and spacecraft telemetry and control communications, the baseband waveform of the physical layer signal is the fundamental unit of information. Its pattern directly determines the bandwidth occupied by the communication link and the physical channel rate. Subject to ITU radio regulations and frequency allocation, it typically requires reasonable design to ensure that it occupies the least possible frequency bandwidth without generating inter-symbol interference. This is a crucial design step in wireless communications. For example, the CCSDS protocol, widely used in spacecraft telemetry and control communications, and the DVB-S2 / S2X protocol used in satellite communications both provide detailed definitions and descriptions of baseband waveforms.

[0003] The transmission rate of a baseband waveform is known as the symbol rate. This symbol rate determines the bandwidth occupied by the baseband signal. In communication systems, it is often necessary to flexibly adjust the baseband symbol rate. This primarily involves the following three scenarios: 1) Multiple access protocols require the symbol rate to be adjusted based on frequency allocation. For example, in MF-TDMA multiple access, the appropriate symbol rate must be set based on the bandwidth specified by the randomly accessed time-frequency resources. 2) Different communication scenarios require different symbol rates. For example, at different satellite elevation angles, different symbol rates must be switched based on signal-to-noise ratios to ensure reliable communication. 3) In cognitive radio, changes in idle frequencies also require corresponding symbol rate adjustments to accommodate bandwidth changes. However, due to the need for baseband shaping filtering to suppress out-of-band power in the baseband signal and reduce bandwidth occupancy, existing symbol rates are typically adjusted between several fixed rates. For example, a set of symbol rates that are doubled is a common symbol rate combination. This coarse-grained, step-by-step adjustment approach makes it difficult to achieve optimal communication power and bandwidth efficiency, and it also hinders the full utilization of frequency resources within varying frequency bands. Summary of the Invention

[0004] In view of the problem in the prior art that coarse-grained step-type symbol rate adjustment is difficult to achieve optimal power efficiency and bandwidth efficiency, the purpose of the present invention is to overcome the above-mentioned defects of the prior art and propose a communication baseband signal generator with continuously variable symbol rate.

[0005] In order to achieve the above object, the present invention discloses a communication baseband signal generator with continuously variable symbol rate, comprising:

[0006] The shaping filter is used to perform baseband shaping on the IQ symbol sequence and convert the IQ symbol sequence into a baseband waveform that meets the requirements;

[0007] an interpolation controller for determining interpolation points and fractional intervals by inputting a symbol rate control word; and

[0008] The fractional interpolator is used to generate a baseband waveform at the DA sampling frequency by applying a fractional interpolation technique to the baseband waveform at an interpolation point.

[0009] Preferably, the input of the shaping filter is: an IQ symbol sequence x obtained by upsampling the modulated IQ symbols twice or four times. IQ (m), the output is: the formed baseband waveform x s (m), the sampling frequency is f i , the interval period is T i , and satisfy f i ≤f s ,T i ≥T s , where f s is the DA sampling frequency, T s is the sampling interval.

[0010] Preferably, when the shaping filter adopts a root raised cosine filter SRRC, its frequency domain response H(f) is:

[0011]

[0012] Where, f represents the frequency, f N =R s / 2 is the Nyquist frequency, R s is the symbol rate; α is the roll-off factor, which is set according to the specific communication standard adopted.

[0013] Preferably, the interpolation controller is implemented using a digitally controlled oscillator NCO, and its input is a symbol rate control word SR_ctrl, which satisfies the following formula:

[0014]

[0015] Digitally controlled oscillator NCO at f s The following accumulation operation is performed at the working frequency:

[0016] ph(k)=[ph(k-1)+SR_ctrl]mod1

[0017] Wherein, ph(k) and ph(k-1) are the accumulated values ​​of the register at the kth and k-1th times, respectively, and ph(0) = 0;

[0018] The output is the fractional interval μ used for each interpolation of the fractional interpolator k , satisfying the following formula:

[0019] μ k =ph(k);

[0020] The interpolation point is for each f s sampling points.

[0021] Preferably, update x once when ph(k-1)+SR_ctrl≥1 s (m), that is, x s (m) walks an interval period T i .

[0022] Preferably, the fractional interpolator adopts third-order polynomial interpolation of FARROW structure, and has four parallel finite impulse response filter FIR branches, namely FIR1, FIR2, FIR3 and FIR4. Each FIR branch has the same structure and has 4 tap coefficients.

[0023] Preferably, the input of the fractional interpolator is: the shaped baseband waveform x s (m) and fractional interval μ k ; Output is: at DA sampling frequency f s The baseband waveform y(k) under the symbol rate is R s , satisfying the following formula:

[0024]

[0025] Where l is 0 to 3, corresponding to FIR1, FIR2, FIR3, and FIR4 respectively; v1 to v4 correspond to the outputs of FIR1, FIR2, FIR3, and FIR4 respectively, and complete the following operations respectively:

[0026]

[0027]

[0028]

[0029]

[0030] Wherein, i represents the i-th tap, and b1(i) to b4(i) correspond to the i-th tap coefficients of FIR1, FIR2, FIR3, and FIR4, respectively.

[0031] Compared with the prior art, the advantages of the present invention are:

[0032] The present invention can adapt to any baseband shaping filter, and baseband waveforms with different roll-off coefficients can generate baseband waveforms with any symbol rate. The symbol rate step accuracy generated by the present invention can be easily controlled below 1 Hz, and can even achieve better than 10 -6 The Hz step accuracy can well meet the needs of fine adjustment of symbol rate in various wireless communication scenarios such as spacecraft measurement and control communications, satellite communications, etc. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 It is a block diagram of a communication baseband signal generator with continuously variable symbol rate;

[0034] Figure 2 This is the time domain waveform of the root raised cosine shaping filter with a roll-off factor of 0.25;

[0035] Figure 3 It is a fraction interval relationship diagram;

[0036] Figure 4 shows the baseband waveform and spectrum before and after interpolation at an 11Msps symbol rate. Figure 4(a) is a waveform scatter plot before and after interpolation, and Figure 4(b) is a baseband waveform after interpolation. Figure 4(c) is a baseband spectrum before interpolation, and Figure 4(d) is a baseband spectrum after interpolation.

[0037] Figure 5 shows the baseband waveform and spectrum before and after interpolation at a 7.5Msps symbol rate. Figure 5(a) is a waveform scatter plot before and after interpolation, and Figure 5(b) is a baseband waveform after interpolation. Figure 5(c) is a baseband spectrum before interpolation, and Figure 5(d) is a baseband spectrum after interpolation.

[0038] Figure 6 shows the baseband waveform and spectrum before and after interpolation at a symbol rate of 7.500001Msp. Figure 6(a) shows the waveform scatter plot before and after interpolation, and Figure 6(b) shows the baseband waveform after interpolation. Figure 6(c) shows the baseband spectrum before interpolation, and Figure 6(d) shows the baseband spectrum after interpolation.

[0039] Figure 7 shows the baseband waveform and spectrum before and after interpolation at a symbol rate of 7.50000001Msps. Figure 7(a) is a scatter plot of the waveform before and after interpolation, and Figure 7(b) is the baseband waveform after interpolation. Figure 7(c) is the baseband spectrum before interpolation, and Figure 7(d) is the baseband spectrum after interpolation. DETAILED DESCRIPTION

[0040] We have invented a communication baseband signal generator that can achieve continuously variable symbol rate. It uses fractional interpolation technology to shape baseband waveforms with arbitrary symbol rate and roll-off factor. The symbol rate step accuracy can be easily controlled below 1Hz, truly realizing continuously variable symbol rate.

[0041] The technical solution of the present invention is described in detail below with reference to the accompanying drawings and embodiments.

[0042] Example

[0043] The embodiment of the present invention provides a communication baseband signal generator with continuously variable symbol rate, such as Figure 1 It should be emphasized that this application emphasizes the design concept of the structural block diagram and does not limit the implementation method using software or hardware.

[0044] It mainly consists of three parts: 1) The shaping filter SRRC in the upper left corner is responsible for baseband shaping of the IQ symbol sequence and converting the IQ symbols into a baseband waveform that meets the requirements; 2) The interpolation controller NCO in the lower left corner is responsible for controlling the update of the interpolator input data and generating the fractional intervals required for interpolation; 3) Except for 1) and 2), the remaining part is the fractional interpolator, namely the FARROW interpolator, which generates the baseband waveform at the DA (digital-to-analog converter) sampling rate through interpolation.

[0045] (2) The shaping filter uses a root raised cosine filter (SRRC), which is a finite impulse response filter. Its specific response in the frequency domain is as follows:

[0046]

[0047] Among them, f N =R s / 2 is the Nyquist frequency. α is the roll-off factor, which can be set according to the specific communication standard used. For example, in DVB-S2 / S2X, the possible values ​​are 0.35, 0.25, 0.20, 0.15, 0.10, and 0.05. The smaller the roll-off factor, the smaller the bandwidth occupied by the signal. Different roll-off factors only affect the symbol waveform style and have no effect on subsequent waveform generation at any symbol rate, making them applicable to all devices. Figure 2 The time domain waveform of SRRC when α=0.25 is shown, and its span is 8 symbol periods. The shaping filter here can also be other types of shaping filters, such as Gaussian shaping filter, etc. The type of shaping filter has no effect on subsequent processing.

[0048] (3) The specific shaping and filtering process is as follows:

[0049] First, the modulated IQ symbols are upsampled twice or four times to obtain x IQ (m), the upsampling process is to directly fill zero after the IQ symbol, 1 zero is added for 2 times upsampling, and 3 zeros are added for 4 times upsampling, and then sent to the shaping filter with the same oversampling multiple for shaping filtering to obtain the shaped baseband waveform x s (m).

[0050] (4) The interpolation controller is implemented using a digitally controlled oscillator (NCO). The interpolation points and fractional intervals are determined by the input symbol rate control word SR_ctrl. This is a key module in the entire framework. The specific implementation process is as follows:

[0051] (5) Let the sampling frequency of DA be f s , the sampling interval is T s Let the formed baseband waveform x s The sampling frequency of (m) is f i , the interval period is T i If 2 times oversampling is used, f i =2R s If 4 times oversampling is used, f i =4R s , R s is the symbol rate, and satisfies f i ≤f s ,T i ≥T s Therefore, the symbol rate control word can be expressed as:

[0052]

[0053] SR_ctrl is a constant less than 1. When the DA sampling rate is determined, its accuracy directly determines the step accuracy of the symbol rate: Taking a 2x oversampling rate as an example, R s =SR_ctrl·f s / 2, usually in hardware implementation, a register with a bit width is used to represent SR_ctrl. Let the bit width be N, then the step accuracy of the symbol rate is f s / 2 N+1 .

[0054] NCO in f s The following accumulation operation is performed at the working frequency:

[0055] ph(k)=[ph(k-1)+SR_ctrl]mod1

[0056] Wherein ph(k) and ph(k-1) are the accumulated values ​​of the register at the kth and k-1th times respectively, and ph(0)=0.

[0057] (6)x s Updates to (m):

[0058] when

[0059] ph(k-1)+SR_ctrl≥1

[0060] Update once xs (m), that is, x s (m) One clock cycle T i .

[0061] (7) Interpolation points and fractional intervals μ k Determination of:

[0062] Because f i ≤f s , the entire interpolation belongs to the upsampling process, so each f s The sampling points of are all interpolation points, and the fractional interval of each interpolation is:

[0063] μ k =ph(k)

[0064] Interpolation interval μ k With T s and T i See the relationship Figure 3 .like Figure 3 As shown, the horizontal axis is the time axis, and the interval between every two real columns is T i , the interval between each two virtual columns is T s , the real column and the imaginary column at time zero coincide, and the vertical axis represents μ k The relationship curve as a whole presents a periodic triangular tooth shape, and the slope is SR_ctrl / T s .

[0065] (8) Figure 1 As shown in the figure, the fractional interpolator adopts a third-order polynomial interpolation with a FARROW structure. There are four FIR (finite impulse response) filter branches, namely FIR1, FIR2, FIR3 and FIR4. The FIR filter of each branch has 4 tap coefficients, and the structure of each FIR branch is consistent, as shown in the expanded structure of FIR1 in the figure.

[0066] (9) The fractional interpolator uses the x obtained in the previous step s (m) and μ k The final interpolated y(k) is obtained through the following process:

[0067] FIR1, FIR2, FIR3, and FIR4 complete the following operations respectively:

[0068]

[0069]

[0070]

[0071]

[0072] The final output of the fractional interpolator is obtained by the following operations:

[0073]

[0074] Among them, l is 0 to 3, corresponding to FIR1, FIR2, FIR3 and FIR4 respectively.

[0075] The tap coefficients of the third-order polynomial interpolator are shown in the following table:

[0076] Table 1. Tap coefficients of third-order polynomial interpolator

[0077] i <![CDATA[b1]]> <![CDATA[b2]]> <![CDATA[b3]]> <![CDATA[b4]]> 0 1 / 6 0 -1 / 6 0 1 -1 / 2 1 / 2 1 0 2 1 / 2 -1 -1 / 2 1 3 -1 / 6 1 / 2 -1 / 3 0

[0078] (10) The final y(k) is the symbol rate R s , the sampling rate is converted to the DA sampling rate f s The communication baseband signal can be directly sent to the DA chip for digital-to-analog conversion and subsequent RF module processing. The symbol rate is controlled by SR_ctrl. If a double oversampling rate is used in the shaping filter process, R s =SR_ctrl·f s / 2, if the shaping filter process uses a four-fold oversampling rate, then R s =SR_ctrl·f s / 4. When f s After the symbol rate is determined, the symbol rate step accuracy is completely determined by SR_ctrl. By selecting the bit width of the control word, it is easy to achieve a symbol rate step accuracy much better than 1Hz. For example, if the control word width is 32 bits, the double oversampling rate is formed, and the DA sampling rate is 100Msps, the symbol rate step accuracy is about 0.012Hz. If a 48-bit control word is used, the symbol rate step accuracy can reach about 1.78×10 -7 Hz!

[0079] (11) The following is a specific example of using this method to generate a communication baseband waveform with an arbitrary symbol rate: the DA sampling rate is 100Msps, the SRRC shaping filter roll-off coefficient is 0.25, the span is 8, and the oversampling rate is 4. This method is used to generate communication baseband signals with symbol rates of 11Msps, 7.5Msps, 7.5Msps+1sps, and 7.5Msps+0.01sps, respectively. The baseband waveforms and corresponding spectra before and after interpolation are given, as shown in Figures 4 to 7, respectively. Figure 4 shows the baseband waveforms and spectra before and after interpolation at an 11Msps symbol rate, where Figure 4(a) is a waveform scatter plot before and after interpolation, and Figure 4(b) is a baseband waveform after interpolation; Figure 4(c) is a baseband spectrum before interpolation, and Figure 4(d) is a baseband spectrum after interpolation.

[0080] Figure 5 shows the baseband waveform and spectrum before and after interpolation at a 7.5Msps symbol rate. Figure 5(a) is a waveform scatter plot before and after interpolation, and Figure 5(b) is a baseband waveform after interpolation. Figure 5(c) is a baseband spectrum before interpolation, and Figure 5(d) is a baseband spectrum after interpolation.

[0081] Figure 6 shows the baseband waveform and spectrum before and after interpolation at a symbol rate of 7.500001Msp. Figure 6(a) shows the waveform scatter plot before and after interpolation, and Figure 6(b) shows the baseband waveform after interpolation. Figure 6(c) shows the baseband spectrum before interpolation, and Figure 6(d) shows the baseband spectrum after interpolation.

[0082] Figure 7 shows the baseband waveform and spectrum before and after interpolation at a symbol rate of 7.50000001Msps. Figure 7(a) is a scatter plot of the waveform before and after interpolation, and Figure 7(b) is the baseband waveform after interpolation. Figure 7(c) is the baseband spectrum before interpolation, and Figure 7(d) is the baseband spectrum after interpolation.

[0083] The baseband waveforms of these four rates all involve non-integer multiple interpolation, and the baseband waveforms before and after interpolation remain consistent, verifying the effectiveness of this method. As can be seen from the spectrum diagram, the image suppression of the baseband waveform generated by the present invention is better than 45dB.

[0084] Finally, it should be noted that the above embodiments are intended only to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the embodiments, it should be understood by those skilled in the art that modifications or equivalent substitutions to the technical solutions of the present invention do not depart from the spirit and scope of the technical solutions of the present invention and are intended to be encompassed by the claims of the present invention.

Claims

1. A communication baseband signal generator with continuously variable symbol rate, characterized in that: include: The shaping filter is used to perform baseband shaping on the IQ symbol sequence and convert the IQ symbol sequence into a baseband waveform that meets the requirements; An interpolation controller, configured to determine an interpolation point and a fractional interval according to an input symbol rate control word; and A fractional interpolator is used to generate a baseband waveform at the DA sampling frequency by applying a fractional interpolation technique to the baseband waveform at the interpolation point; The input of the shaping filter is: the IQ symbol sequence x obtained by upsampling the modulated IQ symbol twice or four times IQ (m), the output is: the formed baseband waveform x s (m), the sampling frequency is f i , the interval period is T i , and satisfy f i ≤f s ,T i ≥T s , where f s is the DA sampling frequency, T s is the sampling interval; When the shaping filter adopts a root raised cosine filter SRRC, its frequency domain response H(f) is: Where, f represents the frequency, f N =R s / 2 is the Nyquist frequency, R s is the symbol rate; α is the roll-off factor, which is set according to the specific communication standard adopted; The interpolation controller is implemented using a digitally controlled oscillator (NCO), and its input is the symbol rate control word SR_ctrl, which satisfies the following formula: Digitally controlled oscillator NCO at f s The following accumulation operation is performed at the working frequency: ph(k)=[ph(k-1)+SR_ctrl]mod1 Wherein, ph(k) and ph(k-1) are the accumulated values ​​of the register at the kth and k-1th times, respectively, and ph(0) = 0; The output is the fractional interval μ used for each interpolation of the fractional interpolator k , satisfying the following formula: μ k =ph(k); The interpolation point is for each f s sampling points.

2. The communication baseband signal generator with continuously variable symbol rate according to claim 1, characterized in that: Update x once when ph(k-1)+SR_ctrl≥1 s (m), that is, x s (m) walks an interval period T i .

3. The communication baseband signal generator with continuously variable symbol rate according to claim 1, characterized in that: The fractional interpolator adopts third-order polynomial interpolation of FARROW structure, and has four parallel finite impulse response filter FIR branches, namely FIR1, FIR2, FIR3 and FIR4. Each FIR branch has the same structure and has 4 tap coefficients.

4. The communication baseband signal generator with continuously variable symbol rate according to claim 3, characterized in that: The input of the fractional interpolator is: the shaped baseband waveform x s (m) and fractional interval μ k ; Output is: at DA sampling frequency f s The baseband waveform y(k) under the symbol rate is R s , satisfying the following formula: Where l is 0 to 3, corresponding to FIR1, FIR2, FIR3, and FIR4 respectively; v1 to v4 correspond to the outputs of FIR1, FIR2, FIR3, and FIR4 respectively, and complete the following operations respectively: Wherein, i represents the i-th tap, and b1(i) to b4(i) correspond to the i-th tap coefficients of FIR1, FIR2, FIR3, and FIR4, respectively.

Images