CN101237435B - A method and device for reducing the peak-to-average ratio of multi-carrier signals - Google Patents

Abstract

The present invention relates to a system and method for sending digital signals, more specifically, a method and device for reducing the peak-to-average ratio of multi-carrier signals, including a device for reducing the peak-to-average ratio of multi-carrier signals, including: a hard limiting module , two delay modules, a carrier power estimation module, a matched filter processing module and two summation and accumulation modules. Applying the Coordinate Rotation Digital Computing (CORDIC) mode to the hard clipping process reduces the complexity and resource consumption of hardware implementation. The matched filter processing module is implemented by converting the peak noise signal to baseband with a matched filter bank. The invention adjusts the gain of the adjustable filter coefficients of the filter bank of the corresponding carrier according to the power estimation value of the baseband signal of each carrier, and performs digital intermediate frequency matching filtering and peak clipping processing on the combined multi-carrier signal, avoiding the need for different power carriers Signal distortion imbalance caused by using the same matched filter gain.

Description

A method and device for reducing the peak-to-average ratio of multi-carrier signals

technical field

The present invention relates to a system and method for sending digital signals, more specifically, a method and device for reducing the peak-to-average ratio of signals entering a power amplifier in a communication system, and in particular to a method for reducing the peak-to-average ratio of multi-carrier combined signals Peak clipping technology. the

Background technique

The transmitter of the wireless base station in the communication system uses a power amplifier to transmit signals to compensate for the signal attenuation caused by the propagation distance. The input signal with a large peak-to-average ratio reduces the efficiency of the power amplifier. For multi-carrier combined signals, such as code division multiple access systems, this problem is particularly prominent. the

In mobile communication systems, the peak-to-average ratio of the signal entering the power amplifier is usually reduced by peak clipping technology to improve the efficiency of the power amplifier, but it will also bring a certain degree of signal distortion, out-of-band spectrum expansion or adjacent channel interference. This degrades emission performance. In code division multiple access systems (such as WCDMA and cdma2000), the error vector magnitude is usually used to measure the degree of signal distortion, and the adjacent channel power leakage ratio is used to characterize the degree of out-of-band spectrum extension of the transmitter. the

Baseband clipping and digital IF clipping are two main types of clipping techniques in communication systems. Baseband clipping is performed before the pulse shaping filter, so it does not bring any out-of-band spectrum spread or adjacent channel power Leakage, its disadvantage is that the peak-to-average ratio of the peak-cut signal will increase after passing through the digital intermediate frequency multi-stage interpolation filter, so that the suppression of the peak-to-average ratio of the signal is not good enough. In addition, baseband clipping causes large signal distortion, which is manifested as a significant drop in error vector magnitude performance. Under the premise of satisfying the error vector magnitude performance, baseband peak clipping is difficult to make the signal peak-to-average ratio after digital intermediate frequency processing meet the expected index requirements. the

Digital IF clipping means that the baseband signal is clipped after passing through the digital IF multi-stage interpolation filter. Commonly used IF clipping includes hard clipping and matched filter clipping. The advantage of IF hard clipping technology is that the hardware implementation is relatively simple, but the disadvantage is that it directly performs digital IF hard clipping processing, which will cause serious adjacent channel power leakage. However, if the adjacent channel power leakage ratio performance index is to be met, it is difficult for the mid-frequency hard peak clipping technology to achieve the expected peak-to-average ratio performance. the

The peak clipping technology used in the invention patent application with the Chinese patent application number 03804978.3 belongs to the digital intermediate frequency matched filter peak clipping technology. This patent application is designed for the peak clipping of multi-carrier signals. First, a baseband with a digital intermediate frequency corresponding to the sampling frequency is designed. filter, shift the frequency of the baseband filter to the corresponding carrier frequency, and then add the frequency-shifted filters to obtain a combination filter, and use the combination filter to perform matched filtering on the peak cancellation signal. When the sampling frequency of the input combined digital intermediate frequency signal is relatively high, such as 12 times or more, the number of taps of the filter is very high, which increases the cost and difficulty of hardware implementation (FPGA or ASIC). the

In addition, although this technical solution considers the appropriate adjustment of the gain of the matched filter coefficients of each carrier to obtain better performance than only using the scaler for uniform gain adjustment, it does not specify how to match each carrier. The gain adjustment of the filter coefficients, and how to adjust the matched filter coefficients of each carrier in real time, especially how to adjust the gains of each matched filter in system implementation. the

Contents of the invention

In order to overcome the problems existing in the prior art, the present invention proposes a method and device for reducing the peak-to-average ratio of multi-carrier signals, and the present invention is specifically implemented as follows:

A device for reducing the peak-to-average ratio of multi-carrier signals, comprising: a hard limiting module, two delay modules, a carrier power estimation module, a matched filter processing module and two summation and accumulation modules;

The hard limiting module processes the input multi-carrier combined signal according to the threshold, and adopts the coordinate rotation digital calculation mode to generate the hard limited signal, and the hard limited module further includes: an amplitude calculation unit of the multi-carrier combined signal , the comparison with the threshold parameter and the decision unit and the limit unit that generates the hard limit signal according to the threshold parameter, wherein the coordinate rotation digital calculation mode is: (1) use the coordinate rotation calculation vector mode to obtain each input sampling signal Amplitude, save the rotation parameters calculated by each step; (2) use the coordinate rotation calculation vector mode to realize the transformation from Cartesian coordinates to polar coordinates; (3) according to the given multi-carrier combined signal, use the coordinate rotation calculation mode to rotate Finally obtain the amplitude and rotation gain of the input sampling signal; (4) compare the product of the amplitude obtained in step (1) with the product of the rotation gain in the threshold and step (3): if the amplitude obtained in step (1) is not greater than The product of the rotation gain in the threshold and step (3), then the hard-limited signal is the input multi-carrier combined signal; otherwise, the ratio of the rotation gain in the threshold to the step (3) and 0 are used as initial parameters using the coordinate The rotation calculation mode reverses the rotation, and calculates the hard clipping signal according to the rotation parameters saved during the coordinate rotation calculation vector mode calculation, where the coordinate rotation calculation mode derotation rotation parameters used in each stage must be aligned with the coordinate rotation vector mode;

The one summing and accumulating module generates a multi-carrier peak cancellation signal by subtracting the hard-limited signal and the multi-carrier combined signal processed by a delay module;

The carrier power estimation module estimates the power of each carrier according to the multi-carrier baseband signal; the matched filter processing module performs matched filtering on the peak cancellation signal of each carrier;

The one summing and accumulating module adds the filtered multi-carrier peak cancellation signal to the multi-carrier combination signal processed by another delay module to generate a multi-carrier peak-shaving signal. the

The iterative device is used for performing multi-stage peak-shaving processing on multi-carrier combined signals. the

The matched filter processing module further includes: two frequency mixing units, at least one matched filter bank, and a summation and accumulation unit; the one frequency mixing unit uses the negative frequency of the corresponding frequency of each carrier to down-convert the peak cancellation signal Obtain the corresponding baseband signal; the matched filter bank performs filtering and downsampling, gain adjustment, upsampling and filtering processing on each baseband signal; the first mixing unit uses the corresponding frequency of each carrier to perform matching filtering on each path The output signal of the device group is subjected to up-conversion processing; the summing and accumulating unit sums and accumulates the output results after frequency conversion on each circuit to obtain a filtered multi-carrier peak cancellation signal. the

The matched filter bank includes: two filtering and 2 times downsampling units, an adjustable gain filter and two 2 times upsampling and filtering units; wherein the adjustable gain filter is based on the estimated value of the corresponding carrier power Make gain adjustments to the filter. The gain-tunable filter may be an approximate square filter. the

A method for reducing the peak-to-average ratio of multi-carrier signals, comprising the steps of:

Step 1, according to multi-carrier combination signal and threshold value, produce hard clipping signal; Described step 1 further comprises: (1) obtain the magnitude of each input sampling signal with coordinate rotation calculation vector mode, save the rotation that each step calculates Parameters; (2) Use coordinate rotation to calculate the vector mode to realize the transformation from Cartesian coordinates to polar coordinates; (3) According to the given multi-carrier combined signal, use the coordinate rotation calculation mode to rotate to obtain the amplitude and rotation of the input sampling signal Gain; (4) compare the product of the amplitude obtained in step (1) with the threshold and the rotation gain in step (3): if the amplitude obtained in step (1) is not greater than the threshold and the rotation in step (3) The product of the gain, the hard-limited signal is the input multi-carrier combined signal; otherwise, the ratio of the threshold to the rotation gain in step (3) and 0 are used as the starting parameters to reverse the rotation using the coordinate rotation calculation mode, and according to the coordinate The rotation parameters saved during the calculation of the rotation calculation vector mode are used to calculate the hard clipping signal, and the rotation parameters used in each stage of the derotation in the coordinate rotation calculation mode must be aligned with the coordinate rotation vector mode. the

Step 2. Generate a peak cancellation signal according to the delayed multi-carrier combined signal and the hard-limited signal;

Step 3, estimating the power of each carrier according to the multi-carrier baseband signal;

Step 4, performing matched filter processing on the multi-carrier peak cancellation signal;

Step 5, summing and accumulating the filtered peak cancellation signal and the multi-carrier combined signal to obtain a multi-carrier peak-shaving signal. the

The method for reducing the peak-to-average ratio of multi-carrier signals further includes: performing multi-stage peak-shaving iterative processing on the multi-carrier combined signal. the

Said step 1 further includes:

(1) Calculate the multi-carrier combined signal using the coordinate rotation calculation mode to obtain the amplitude, phase and rotation gain of each input sampling signal;

(2) Compare the magnitude to the product of the threshold and the rotation gain:

If the amplitude is not greater than the product of the threshold and the rotation gain, the hard-limited signal is an input multi-carrier combination signal;

Otherwise, the coordinate rotation calculation vector mode uses the ratio and phase of the threshold and the rotation gain as a starting point to obtain a hard-limited signal with the coordinate parameter. the

Said step 4 further includes:

(1) Use the negative frequency corresponding to each carrier frequency to down-convert the peak cancellation signal to obtain the corresponding baseband signal;

(2) Perform filtering and downsampling, adjustable filtering, upsampling and filtering processing on each baseband signal;

(3) Adjust the gain of the adjustable filter according to the estimated value of the carrier power;

(4) Up-convert the output signals of each matched filter bank with the frequency corresponding to each carrier;

(5) The output after frequency conversion of each channel is summed and accumulated to obtain a filtered multi-carrier peak cancellation signal. the

According to the power estimation value of each carrier baseband signal, the invention adjusts the gain of the adjustable filter coefficient of the filter bank of the corresponding carrier, and performs digital intermediate frequency matching filtering and peak clipping processing on the combined multi-carrier signal, avoiding the difference in power The carrier uses the same matched filter gain to cause signal distortion and imbalance. The present invention can achieve a better peak clipping effect under the condition of the same error vector magnitude and adjacent channel power leakage ratio, that is, the peak-to-average ratio of the combined multi-carrier signal after peak clipping can be lower. the

The hard clipping method adopted by the present invention is different from the traditional implementation method, and the coordinate rotation digital calculation (CORDIC) mode is applied to the hard clipping process, which reduces the complexity and resource consumption of hardware implementation; the matched filter processing module passes the peak value The conversion of the cancellation signal to the baseband is realized by a matched filter bank, which greatly reduces the multiplier resources required by the traditional matched filter method, and the adjustable filter design method in the matched filter bank improves the effect of peak clipping. the

In digital IF processing, if you want to obtain a higher sampling rate while taking into account the saving of hardware resources, you can use a two-stage half-band interpolation filter. In addition, design a matched filter bank according to the specific IF sampling rate. If you want to obtain lower peak-to-average ratio performance under the conditions of error vector magnitude and adjacent channel power leakage ratio,

The multi-stage iterative peak-shaving processing properly used in the present invention adopts a sequential processing method for digital signal processing without any feedback processing module, and is convenient to be implemented by pipeline in an actual hardware system. the

Description of drawings

Fig. 1 is the schematic diagram of the position of multi-carrier signal formation and multi-carrier peak clipping module in the present invention;

Fig. 2 makes the specific embodiment of the present invention, the amplitude-frequency response of the RRC filter after compensating CIC gain in the specific realization structure shown in Fig. 1;

Fig. 3 is the basic structure that the multi-carrier peak clipping that is used to reduce multi-carrier signal peak-to-average ratio of the present invention realizes;

Fig. 4 (a), (b) are the hard clipping realization method in the structure shown in Fig. 3;

Fig. 5 is the concrete realization of the matched filter processing of basic structure shown in Fig. 3;

Fig. 6 is the specific implementation structure of the corresponding matched filter bank in the matched filter processing module shown in Fig. 5;

Fig. 7 is the magnitude-frequency response comparative performance of RRC and square filter of the present invention;

Fig. 8 is another kind of specific realization of the basic structure shown in Fig. 3;

FIG. 9 is a CCDF curve of the multi-carrier combined signal before and after the multi-carrier peak clipping of the present invention. the

Detailed ways

The present invention is described in detail with reference to the accompanying drawings. the

In the base station transmitter of the wireless communication system, peak-shaving technology is used to reduce the peak-to-average ratio of the multi-carrier combined signal entering the power amplifier, so as to improve the efficiency of the power amplifier. Fig. 1 is a schematic diagram of the location of multi-carrier signal formation and multi-carrier peak clipping modules in the present invention. For each carrier, its baseband complex signal is pulse-shaped and filtered through a compensated root-raised cosine interpolation filter (CRRC), and passed through a half-band interpolation filter (HBF) to increase the sampling rate of the digital intermediate frequency signal, and then through cascaded integration Comb interpolation filter (CIC) to enhance the image suppression of the signal in the frequency domain, and mix with the numerically controlled oscillator to form a single-carrier digital intermediate frequency signal output, and finally combine the digital intermediate frequency signals of each carrier to produce a multi-carrier combined output Signal. the

In order to reduce the cost of hardware storage tables, the NCO frequencies of the four carriers can be set to be symmetrical, for example, f ₁ , f ₂ , f ₃ and f ₄ are set to f ₁ , f ₂ , -f ₂ and -f ₁ respectively, In this way, the NCOs of f ₁ and -f ₁ can share a look-up table (LUT). Furthermore, if a higher sampling rate and reduced hardware implementation complexity are required, two-stage half-band interpolation filters can be used in digital IF processing.

In the interpolation and filter processing of digital intermediate frequency signals, although the use of CIC filters reduces the cost of hardware implementation and provides good image rejection performance, it will also cause uneven gain in the band. In order to make the entire digital intermediate frequency interpolation and filtering process have better in-band gain flatness, a compensated RRC filter (CRRC) is designed to compensate the in-band gain unevenness of the CIC filter. the

Figure 2 shows the amplitude-frequency response of the CRRC filter designed according to the example in Figure 1. Figure 2 also shows the amplitude of the convolution results of the RRC filter and the entire digital IF filter (including CRRC, HBF, and CIC) before and after compensation. From the frequency response characteristics, it can be seen that if the RRC does not compensate the in-band gain loss of the CIC, the in-band flatness performance of the convolution of the entire digital IF filter is poor. the

FIG. 3 is the basic structure of the implementation of multi-carrier peak clipping for reducing the peak-to-average ratio of multi-carrier signals according to the present invention. The multi-carrier peak clipping device in Fig. 3 is composed of a hard clipping module, two delay modules, a carrier power estimation module, a matched filter processing module and two summation and accumulation modules. the

The hard limiting module performs hard limiting processing on the input multi-carrier combined signal according to the threshold parameter to generate a hard limiting signal. The delay module performs delay processing on the multi-carrier combined signal. The multi-carrier combined signal after the delay module is subtracted from the hard-limited signal to generate a peak cancellation signal. The one summing and accumulating module subtracts the hard-limiting signal and the multi-carrier combined signal processed by a delay module to generate a multi-carrier peak cancellation signal; the matched filter processing module performs matched filtering on the peak cancellation signal to obtain the filtered The multi-carrier peak cancellation signal. The carrier power estimation module is used for estimating the baseband signal power of each carrier for estimation. The other summing and accumulating module is for summing and accumulating the multi-carrier combined signal after the delay module and the filtered multi-carrier peak cancellation signal to generate a multi-carrier peak clipping signal. the

The hard limiting module further includes an amplitude calculation unit of the multi-carrier combined IQ signal, a comparison and judgment unit with a threshold parameter, and a limiting unit for generating a hard limited signal according to the threshold parameter. the

The matched filter processing module includes a frequency mixing unit for down-converting the peak cancellation signal to obtain a corresponding baseband signal with the negative frequency corresponding to each carrier frequency; filtering and downsampling, adjustable filtering, and upsampling for each baseband signal The matched filter bank for filtering, adjusts the gain adjustment unit of the adjustable filter according to the estimated value of the carrier power; the frequency mixing unit for up-converting the output signal of each matched filter group with the corresponding frequency of each carrier; the frequency conversion of each channel A summation accumulation unit whose outputs are summed. the

The method for reducing the peak-to-average ratio of a multi-carrier signal according to the present invention includes generating a hard clipping signal according to the multi-carrier combined signal and a threshold parameter; generating a peak cancellation signal according to the delayed multi-carrier combined signal and the hard clipped signal; The power of each carrier is estimated according to the multi-carrier baseband signal; the multi-carrier peak cancellation signal is processed by matched filtering; the filtered peak cancellation signal and the multi-carrier combination signal are summed and accumulated to obtain a multi-carrier peak-shaving signal. the

和旋转方向参数d_n，该幅度结果为信号幅度 

的K倍(K来自于CORDIC算法的处理增益)，这里d_n取决于y_n的符号(若y_n＜0时d_n取1，否则d_n取-1)，每级的旋转角度分别为d_n*arctan 2^-n；(b)图为根据Thr/K和保存的d_n参数进行CORDIC反旋转产生硬限幅输出信号(Thr*cosφ_k和Thr*sinφ_k，  ${\mathrm{\φ}}_k=\arctan \frac{Q_k^{\mathrm{sum}}}{I_k^{\mathrm{sum}}}$ )的过程，每级的旋转角度分别为-d_n*arctan2^-n，反旋转时每级的d_n与正向旋转相同。  The hard-limiting processing further includes: calculating the amplitude of the multi-carrier combined IQ signal, comparing and judging with threshold parameters, and generating a hard-limiting signal according to the threshold parameters. Wherein Figure 4 is a schematic diagram of the implementation process of the hard clipping method in the structure shown in Figure 3, (a) Figure shows the process of using the coordinate rotation digital calculation (CORDIC) mode to calculate the amplitude of the multi-carrier combined signal to obtain the signal amplitude

and the rotation direction parameter d _n , the amplitude results in the signal amplitude

K times (K comes from the processing gain of the CORDIC algorithm), where d _n depends on the sign of y _n (if y _n <0, d _n takes 1, otherwise d _n takes -1), and the rotation angles of each level are respectively d _n *arctan 2 ^-n ; (b) The figure shows that CORDIC reverse rotation is performed according to Thr/K and saved d _n parameters to generate hard-limited output signals (Thr*cosφ _k and Thr*sinφ _k , ${\mathrm{\&phi;}}_k=\arctan \frac{Q_k^{\mathrm{sum}}}{I_k^{\mathrm{sum}}}$ ) process, the rotation angle of each level is -d _n *arctan2 ^-n respectively, and the d _n of each level during the reverse rotation is the same as the forward rotation.

The multi-carrier peak clipping device of the present invention can not only be applied to a multi-carrier communication system, but can also be applied to a single-carrier communication system through proper modification and tailoring according to hardware resources and performance requirements in actual communication system design. If only a single carrier needs to be supported, the parts related to up-conversion and down-conversion can be cut out, which can greatly reduce the consumption of hardware resources. the

Fig. 5 shows an implementation example of matched filter processing adopted according to the present invention. The matched filter processing further includes: respectively using the negative frequency corresponding to each carrier frequency to down-convert the peak cancellation signal to obtain the corresponding baseband signal; to filter, down-sample, and adjustable filter each baseband signal through a matched filter bank respectively. , up-sampling and filtering processing; adjust the gain of the adjustable filter of the corresponding channel matched filter bank according to the estimated value of the carrier power; use the corresponding frequency of each carrier to perform up-conversion on the output of each channel matched filter bank; up-convert the frequency of each channel The final output results are summed and accumulated to obtain a filtered multi-carrier peak cancellation signal. the

Figure 6 shows an implementation example of a matched filter bank that adopts 12 times digital intermediate frequency sampling corresponding to a multi-carrier peak clipping module according to the present invention. The gain adjustment unit of the filter is composed of two 2 times upsampling and filtering units, and the gain adjustment unit performs gain adjustment according to the corresponding carrier power estimation value. The adjustable-gain filter in the matched filter bank is designed as an approximate square filter (rather than a traditional RRC filter) to improve the performance of multi-carrier peak clipping. the

Fig. 7 shows the comparative performance of the amplitude-frequency response of the RRC and the square filter of the present invention. It can be seen from Fig. 7 that the occupied bandwidth of the square filter is larger than that of the RRC filter, which is beneficial to improve the peak clipping performance. the

A mathematical model of a four-carrier communication system is considered in the present invention. For the convenience of description and representation, different mathematical symbols are used below to represent the signals of each stage or module, and i, j and k represent the carrier index, baseband sampling index and intermediate frequency sampling index respectively. the

1) Baseband IQ signal: I _{i, j} ^bb , Q _{i, j} ^bb

2) Intermediate frequency signals corresponding to each carrier after interpolation filtering: I _{i, k} ^dif , Q _{i, k} ^dif

3) Multi-carrier combined signal: I _k ^sum , Q _k ^sum

4) Threshold parameter: Thr

5) Hard clipping signal: I _k ^hard_clip , Q _k ^hard_clip

6) Multi-carrier peak cancellation signal: I _k ^peak_noise , Q _k ^peak_noise

7) Peak cancellation signals corresponding to each carrier after down-conversion: I _{i, k} ^down_shift , Q _{i, k} ^down_shift

8) Peak cancellation signals corresponding to each carrier after matched filtering: I _{i, k} ^filter_bank , Q _{i, k} ^filter_bank

9) Peak cancellation signals corresponding to each carrier after up-conversion: I _{i, k} ^up_shift , Q _{i, k} ^up_shift

10) Filtered multi-carrier peak cancellation signal: I _k ^{final_peak_noise} , Q _k ^{final_peak_noise}

11) Multi-carrier clipping signal: I _k ^clipped , Q _k ^clipped

Each carrier baseband IQ signal undergoes multi-stage interpolation filtering (CRRC, HBF and CIC) to generate digital intermediate frequency sampling signals I _{i, k} ^dif and Q _{i, k} ^dif , which are mixed with the corresponding carrier frequency and added to generate a multi-carrier combined signal I _k ^sum and Q _k ^sum . The four-carrier combined signal before the four-carrier peak clipping module can be expressed as:

$I_k^{\mathrm{sum}}+j*Q_k^{\mathrm{sum}}=\underset{i=1}{\overset{4}{\mathrm{\&Sigma;}}}(I_{i,k}^{\mathrm{dif}}+j*Q_{i,k}^{\mathrm{dif}})*{\mathrm{<figure-callout\; id="2"\; label="e\; j"\; filenames="H07102958420070201E000022.png,H07102958420070201E000031.png"\; state="\{\{state\}\}">e</figure-callout>}}^j$ (Formula 1)

The hard clipping in FIG. 3 of the present invention can be realized by using the CORDIC algorithm. First, use the CORDIC vector mode to obtain the amplitude A _k for each input sampling signal I _k ^sum and Q _k ^sum $(A_k=K*\sqrt{{\left(I_k^{\mathrm{sum}}\right)}^2+{\left(Q_k^{\mathrm{sum}}\right)}^2})$ and phase φ _k ; compare A _k with Thr*K; if A _k is less than or equal to Thr*K, the hard-limited signal is an input multi-carrier combined signal; if A _k is greater than Thr*K, use CORDIC rotation mode A hard-limited signal is obtained with Thr/K and φ _k (calculated from the CORDIC vector mode) as starting parameters.

The hard clipping shown in Fig. 3 of the present invention can also be realized by using the modified CORDIC algorithm, which can save 1/3 of the hardware circuit resources and reduce the cost of hardware implementation. Through Fig. 4 and the following description, it is easy to understand the method for realizing the hard clipping of the present invention. Firstly, calculate the amplitude A _k of each input sampling signal by using the CORDIC vector mode, and save the calculated d _n of each step for subsequent CORDIC inverse rotation, without calculating the phase term φ _k . The CORDIC vector mode realizes the transformation from Cartesian coordinates to polar coordinates, and the formula is expressed as:

$\left\{\begin{array}{c}{x}_{\mathrm{no}+1}=x_{\mathrm{no}}-d_{\mathrm{no}}\&CenterDot;{\mathrm{the\; y}}_{\mathrm{no}}\&CenterDot;2^{-\mathrm{no}}\\ {\mathrm{the\; y}}_{\mathrm{no}+1}={\mathrm{the\; y}}_{\mathrm{no}}+d_{\mathrm{no}}\&CenterDot;x_{\mathrm{no}}\&Center\; Dot;2^{-\mathrm{no}}\end{array}\right.\mathrm{no}\&Greater\; Equal;0$ (Formula 2)

In formula 2, d _n depends on the sign of y _n :

$d_{\mathrm{no}}=\left\{\begin{array}{cc}1& {\mathrm{the\; y}}_{\mathrm{no}}<0\\ -1& {\mathrm{the\; y}}_{\mathrm{no}}\&Greater\; Equal;0\end{array}\right.$ (Formula 3)

given $x_0=I_k^{\mathrm{sum}}$ and ${\mathrm{the\; <figure-callout\; id="0"\; label="y"\; filenames="H07102958420070201E000012.png,H07102958420070201E000052.png"\; state="\{\{state\}\}">y</figure-callout>}}_0=Q_k^{\mathrm{sum}},$ After CORDIC rotation, we get:

${\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; K</span>}<span\; class="notranslate">\; \&Center\; Dot;</span>\sqrt{{{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; +</span>{{\mathrm{<span\; class="notranslate">\; the\; <figure-callout\; id="0"\; label="y"\; filenames="H07102958420070201E000012.png,H07102958420070201E000052.png"\; state="\{\{state\}\}">y</figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}$

y _n+1 ＝0 (Formula 4)

$\mathrm{<span\; class="notranslate">\; K</span>}<span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}}{\overset{\mathrm{<span\; class="notranslate">\; <figure-callout\; id="1"\; label="N"\; filenames="H07102958420070201E000012.png,H07102958420070201E000021.png"\; state="\{\{state\}\}">N</figure-callout></span>}}{\mathrm{<span\; class="notranslate">\; \&Pi;</span>}}}\sqrt{\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; 2</span>}}^{<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; no</span>}}}$

In Equation 4, N is the series of CORDIC rotation (usually N can be set to 8 to obtain good precision), x _n+1 is the amplitude calculated by input sampling signal I _k ^sum and Q _k ^sum after CORDIC, and A _k Compared with Thr*K (K is the gain of CORDIC rotation); if A _k is less than or equal to Thr*K, the hard-limited signal is an input multi-carrier combined signal; if A _k is greater than Thr*K, x ₀ and y ₀ The CORDIC derotation is performed with Thr/K and 0 as the starting parameters respectively, and the hard-limited signal is calculated according to the _dn parameter saved during the calculation of the CORDIC vector mode. The _dn used by each stage of CORDIC derotation must be aligned with the CORDIC vector pattern.

$\left\{\begin{array}{c}{x}_{\mathrm{no}+1}=x_{\mathrm{no}}+d_{\mathrm{no}}\&Center\; Dot;{\mathrm{the\; y}}_{\mathrm{no}}\&CenterDot;2^{-\mathrm{no}}\\ {\mathrm{the\; y}}_{\mathrm{no}+1}={\mathrm{the\; y}}_{\mathrm{no}}-d_{\mathrm{no}}\&Center\; Dot;x_{\mathrm{no}}\&CenterDot;2^{-\mathrm{no}}\end{array}\right.\mathrm{no}\&Greater\; Equal;0$ (Formula 5)

Given x ₀ =Thr/K and y ₀ =0, then:

x _n+1 ＝Thr·cosφ _k

y _n+1 ＝Thr·sinφ _k (Formula 6)

$\mathrm{<span\; class="notranslate">\; K</span>}<span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}}{\overset{\mathrm{<span\; class="notranslate">\; <figure-callout\; id="1"\; label="N"\; filenames="H07102958420070201E000012.png,H07102958420070201E000021.png"\; state="\{\{state\}\}">N</figure-callout></span>}}{\mathrm{<span\; class="notranslate">\; \&Pi;</span>}}}\sqrt{\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; 2</span>}}^{<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; no</span>}}}$

In Equation 6, x _n+1 and y _n+1 are the hard clipping signals I _k ^hard_clip and Q _k ^hard_clip corresponding to the carrier combined signal I _k ^sum and Q _k ^sum , and φ _k is the value of I _k ^sum and Q _k ^sum phase.

The multi-carrier peak cancellation signal is obtained by subtracting the delayed multi-carrier combination signal from the hard-limited signal:

$\left\{\begin{array}{c}{I}_k^{\mathrm{peak}\_\mathrm{noise}}=I_k^{\mathrm{hard}\_\mathrm{clip}}-I_k^{\mathrm{sum}}\\ Q_k^{\mathrm{peak}\_\mathrm{noise}}=Q_k^{\mathrm{hard}\_\mathrm{clip}}-Q_k^{\mathrm{sum}}\end{array}\right.$ (Formula 7)

In the carrier power estimation module in Figure 3, the power of the baseband signal corresponding to each carrier is calculated by formula 8:

$P_i^{\mathrm{est}}=\frac{1}{N_{\mathrm{samples}}}\underset{\mathrm{no}=1}{\overset{N_{\mathrm{samples}}}{\mathrm{\&Sigma;}}}({I_{i,\mathrm{no}}}^2+{Q_{i,\mathrm{no}}}^2),i=\mathrm{1,2,3,4}$ (Formula 8)

In Equation 8, N _samples is the number of baseband samples used for power estimation, and the corresponding amplitude gain A _i ^est is obtained according to the P _i ^est look-up table (LUT).

$A_i^{\mathrm{est}}=\sqrt{P_i^{\mathrm{est}}},i=\mathrm{1,2,3,4}$ (Formula 9)

Use the negative frequency of the carrier frequency to down-convert the peak cancellation signal to obtain the corresponding baseband signal:

$I_{i,k}^{\mathrm{down}\_\mathrm{shift}}+j*Q_{i,k}^{\mathrm{down}\_\mathrm{shift}}=(I_k^{\mathrm{peak}\_\mathrm{noise}}+j*Q_k^{\mathrm{peak}\_\mathrm{noise}})*e^{-2\mathrm{\&pi;}{f}_{i}k},i=\mathrm{1,2,3,4}$ (Formula 10)

The baseband signals output by the down-conversion are respectively subjected to matched filter processing through the matched filter bank shown in Figure 6, and the matched filter bank processing includes adjusting the adjustable filter gain according to the estimated carrier power to obtain I _{i, k} ^filter_bank and Q _{i, k} ^filter_bank :

$\left\{\begin{array}{c}{I}_{i,k}^{\mathrm{filter}\_\mathrm{bank}}=\mathrm{filter}\_\mathrm{bank}\left(I_{i,k}^{\mathrm{down}\_\mathrm{shifted}}\right)*A_i^{\mathrm{est}}\\ Q_{i,k}^{\mathrm{filter}\_\mathrm{bank}}=\mathrm{filter}\_\mathrm{bank}\left(Q_{i,k}^{\mathrm{down}\_\mathrm{shifted}}\right)*A_i^{\mathrm{est}}\end{array}\right.$ (Formula 11)

Use the corresponding carrier frequency to down-convert the peak cancellation signal to obtain the corresponding baseband signal:

$I_{i,k}^{\mathrm{up}\_\mathrm{shift}}+j*Q_{i,k}^{\mathrm{up}\_\mathrm{shift}}=(I_{i,k}^{\mathrm{filter}\_\mathrm{bank}}+j*Q_{i,k}^{\mathrm{filter}\_\mathrm{bank}})*{\mathrm{<figure-callout\; id="2"\; label="e"\; filenames="H07102958420070201E000022.png,H07102958420070201E000031.png"\; state="\{\{state\}\}">e</figure-callout>}}^{2\mathrm{\&pi;}{f}_{i}k},i=\mathrm{1,2,3,4}$ (Formula 12)

After up-conversion, the peak cancellation signals corresponding to each carrier are summed and accumulated to generate a filtered multi-carrier peak cancellation signal:

$\left\{\begin{array}{c}{I}_k^{\mathrm{final}\_\mathrm{peak}\_\mathrm{noise}}=\underset{i=1}{\overset{4}{\mathrm{\&Sigma;}}}{I}_{i,k}^{\mathrm{up}\_\mathrm{shift}}\\ Q_k^{\mathrm{final}\_\mathrm{peak}\_\mathrm{noise}}=\underset{i=1}{\overset{4}{\mathrm{\&Sigma;}}}{Q}_{i,k}^{\mathrm{up}\_\mathrm{shift}}\end{array}\right.$ (Formula 13)

The delayed multi-carrier combined signal and the filtered multi-carrier peak cancellation signal are added to obtain the multi-carrier peak clipping signal:

$\left\{\begin{array}{c}{I}_k^{\mathrm{clipped}}=I_k^{\mathrm{sum}}+I_k^{\mathrm{final}\_\mathrm{peak}\_\mathrm{noise}}\\ Q_k^{\mathrm{clipped}}=Q_k^{\mathrm{sum}}+Q_k^{\mathrm{final}\_\mathrm{peak}\_\mathrm{noise}}\end{array}\right.$ (Formula 14)

FIG. 8 shows a specific embodiment of the present invention, another specific implementation of multi-carrier peak clipping processing of the basic structure shown in FIG. 3 . The difference between the specific implementation of Fig. 8 and Fig. 3 is that the multi-stage iterative peak clipping operation can be performed on the multi-carrier combined signal according to the system performance requirements and hardware resources during system design, so that the system error vector magnitude and adjacent Under the condition of the channel power leakage ratio index, a better peak clipping effect is achieved, that is, a smaller peak-to-average ratio of the multi-carrier peak clipping signal is guaranteed. In Fig. 8, the multi-carrier peak clipping modules at all levels can also use different threshold parameters according to needs, so as to achieve better peak clipping effects. the

FIG. 9 shows the comparison of the CCDF curves of the signals after the four-carrier WCDMA digital IF combined signal passes through the peak clipping module of the present invention. It can be seen from Fig. 9 that not only the peak-to-average ratio decreases significantly after peak clipping, but also drops from 9.8dB (0.01% CCDF) to 5.7dB before peak clipping, and the signal CCDF curve also becomes steeper. Theoretical analysis and laboratory test results show that the steep CCDF curve after peak clipping is beneficial to improve the efficiency of the power amplifier. Compared with similar peak-shaving technologies, the sampling multi-carrier peak-shaving technology of the present invention uses less hardware resources and obtains lower peak-to-average ratio performance under the condition that the system error vector magnitude and adjacent channel power leakage ratio are guaranteed to be met , it is easier to obtain higher power amplifier efficiency. the

The present invention can be appropriately modified and tailored according to hardware resources and performance requirements in actual communication system design. If only a single carrier needs to be supported, the parts related to up-conversion and down-conversion can be cut out, which greatly reduces the consumption of hardware resources. In digital intermediate frequency processing, if you want to obtain a higher sampling rate and take into account the saving of hardware resources, you can use a two-stage half-band interpolation filter. If it is desired to obtain a lower peak-to-average ratio performance under the condition of satisfying the error vector magnitude and adjacent channel power leakage ratio, a multi-level iterative peak-shaving processing method can be appropriately used to achieve it. In the scheme provided by the present invention, the digital signal processing adopts a sequential processing mode, and does not involve any feedback processing module, so it is very convenient to realize in the actual hardware system. the

Here, the present invention has been described in detail through specific implementation examples. The description of the above embodiments is provided in order to enable those skilled in the art to make or apply the present invention. Various modifications of these embodiments are easy for those skilled in the art understand. The invention is applicable to code division multiple access systems such as WCDMA and cdma2000. the

Claims (8)

1. A device for reducing the peak-to-average ratio of multi-carrier signals, characterized in that, comprising: Hard clipping module, two delay modules, carrier power estimation module, matched filter processing module and two summation and accumulation modules; The hard limiting module processes the input multi-carrier combined signal according to the threshold, and adopts the coordinate rotation digital calculation mode to generate the hard limited signal, and the hard limited module further includes: an amplitude calculation unit of the multi-carrier combined signal , the comparison and judgment unit with the threshold parameter and the clipping unit that generates the hard clipping signal according to the threshold parameter, wherein the coordinate rotation digital calculation mode is: (1) obtain each input sampling signal with the coordinate rotation digital calculation mode Amplitude, save the rotation parameters calculated by each step; (2) use the coordinate rotation digital calculation mode to realize the transformation from Cartesian coordinates to polar coordinates, and calculate and input the given multi-carrier according to the given multi-carrier combination signal After combining the signals, the amplitude and rotation gain of the input sampling signal obtained after the rotation of the coordinate rotation digital calculation mode; (3) the product of the amplitude obtained in the step (1) and the threshold value and the rotation gain in the step (2) is carried out Comparison: if the amplitude obtained in step (1) is not greater than the product of the threshold value and the rotation gain in step (2), then the hard-limited signal is the input multi-carrier combination signal; otherwise, the threshold and step (2) The ratio of the rotation gain and 0 is the initial value to carry out the reverse rotation of the coordinate rotation digital calculation mode, wherein the ratio of the threshold and the rotation gain in step (2) and 0 are the initial values of the parameter x ₀ and the parameter y ₀ respectively ,

and

It is a multi-carrier combination signal, and calculates the hard-limited signal according to the rotation parameters saved during the calculation of the coordinate rotation digital calculation mode, wherein the coordinate rotation digital calculation mode derotates the rotation parameters used in each stage and must be aligned with the coordinate rotation digital calculation mode; The one summing and accumulating module subtracts the hard-limited signal and the multi-carrier combined signal processed by a delay module to generate a multi-carrier peak cancellation signal; The carrier power estimation module estimates the power of each carrier according to the multi-carrier baseband signal, and sends it to the matched filter processing module; The matched filter processing module performs matched filter on the peak cancellation signal of each carrier; The other summing and accumulating module adds the filtered multi-carrier peak canceling signal to the multi-carrier combination signal processed by another delay module to generate a multi-carrier peak clipping signal. 2. The device for reducing the peak-to-average ratio of multi-carrier signals as claimed in claim 1, further comprising: The iterative device is used for multi-stage peak-shaving processing of multi-carrier combined signal. 3. The device for reducing the peak-to-average ratio of multi-carrier signals as claimed in claim 1 or 2, wherein the matched filter processing module further comprises: Two frequency mixing units, at least one matched filter bank, and a summing and accumulating unit; The frequency mixing unit uses the negative frequency corresponding to each carrier frequency to down-convert the peak cancellation signal to obtain a corresponding baseband signal; The matched filter bank performs filtering and downsampling, gain adjustment, upsampling and filtering processing on each baseband signal; The other frequency mixing unit performs up-conversion processing on the output signals of each matched filter bank with the frequency corresponding to each carrier; The summing and accumulating unit sums and accumulates the frequency-converted output results of each circuit to obtain a filtered multi-carrier peak cancellation signal. 4. The device for reducing the peak-to-average ratio of multi-carrier signals as claimed in claim 3, wherein the matched filter bank further comprises: Two filtering and 2x downsampling units, gain adjustable filter and two 2x upsampling and filtering units; The adjustable gain filter performs gain adjustment on the filter according to the estimated value of the corresponding carrier power. 5. the device that reduces multi-carrier signal peak-to-average ratio as claimed in claim 4, is characterized in that, The gain-tunable filter may be an approximate square filter. 6. A method for reducing the peak-to-average ratio of multi-carrier signals, comprising the steps of: Step 1. Generate a hard clipping signal according to the multi-carrier combined signal and the threshold; Said step 1 further includes: (1) Obtain the amplitude of each input sampling signal with the coordinate rotation digital calculation mode, and save the rotation parameters obtained by each step of calculation; (2) Realize the conversion from Cartesian coordinates to polar coordinates with the coordinate rotation digital calculation mode, according to the given multi-carrier combination signal, after calculating and inputting the given multi-carrier combination signal, through the coordinate rotation digital calculation mode The amplitude and rotation gain of the input sampling signal obtained after rotation; (3) Compare the magnitude obtained in step (1) with the product of the threshold and the rotation gain in step (2): If the amplitude obtained in the step (1) is not greater than the product of the rotation gain in the threshold value and the step (2), then the hard limiting signal is an input multi-carrier combining signal; Otherwise, take the ratio of the threshold value and the rotation gain in step (2) and 0 as the starting value to carry out the reverse rotation of the coordinate rotation digital calculation mode, wherein the ratio of the threshold value to the rotation gain in step (2) and 0 are respectively starting values for parameter x ₀ and parameter y ₀ , and

It is a multi-carrier combined signal, and calculates the hard-limited signal according to the rotation parameters saved during the calculation of the coordinate rotation digital calculation mode, wherein the coordinate rotation digital calculation mode derotates the rotation parameters used by each stage must be aligned with the coordinate rotation digital calculation mode; Step 2, subtracting the delayed multi-carrier combined signal and the hard-limited signal to generate a peak cancellation signal; Step 3, estimating the power of each carrier according to the multi-carrier baseband signal, and sending it to the matched filter processing module; Step 4, the matched filter processing module performs matched filter processing on the multi-carrier peak cancellation signal; Step 5, summing and accumulating the filtered peak cancellation signal and the delayed multi-carrier combined signal to obtain a multi-carrier peak-shaving signal. 7. The method for reducing the peak-to-average ratio of multi-carrier signals as claimed in claim 6, further comprising: Multi-stage peak clipping iterative processing is performed on the multi-carrier combined signal. 8. The method for reducing the peak-to-average ratio of multi-carrier signals as claimed in claim 6 or 7, wherein said step 4 further comprises: (1) carry out down-conversion to the peak cancellation signal with the negative frequency of the corresponding frequency of each carrier to obtain the corresponding baseband signal; (2) Filtering and down-sampling processing are performed on each baseband signal; (3) Adjust the gain of the adjustable filter according to the estimated value of the carrier power, and the adjustable filter performs adjustable filtering processing on each baseband signal after filtering and down-sampling processing; (4) performing upsampling and filtering processing on the signal processed by the adjustable filter; (5) Up-convert the output signal of the up-sampling and filtering process with the frequency corresponding to each carrier; (6) The output after frequency conversion of each channel is summed and accumulated to obtain a filtered multi-carrier peak cancellation signal.

Images