WO2009010010A1

WO2009010010A1 - A method and device of clipping process of signal

Info

Publication number: WO2009010010A1
Application number: PCT/CN2008/071670
Authority: WO
Inventors: Andrey Vorobyev; Igor Punkov
Original assignee: Huawei Technologies Co., Ltd.
Priority date: 2007-07-17
Filing date: 2008-07-17
Publication date: 2009-01-22
Also published as: CN101076008B; CN101076008A

Abstract

A method of clipping process of signal includes the following steps: in the signal, detecting the peak value point signal which is over the preset threshold; generating a clipping noise according to the peak value point signal over the threshold and the threshold, and making the clipping process according to the clipping noise. A device of clipping process of signal includes a peak value detecting module and a clipping process module. By using the invention, the peak average rate of the signal in the system is effectively reduced, and the outputted power and efficiency of the base station power amplifier are increased on the basis of satisfying the protocol.

Description

Signal clipping method and device

Technical field

Embodiments of the present invention relate to the field of communications technologies, and in particular, to a method and a device for clipping a signal in a communication system. Background technique

In a conventional wireless communication system, a base station uses a power amplifier to communicate with user terminals distributed within a predetermined service area. And especially in CDMA (CDMA Code Division)

Multiple Access, Code Division Multiple Access) /WCDMA (Wireless CDMA, Wireless

In a CDMA) / OFDM (Orthogonal Frequency Division Multiplex) system, a downlink signal is usually a combination of a plurality of user signals. The composite envelope of this signal results in a high PAR (Peak Average Rate) of the final signal, especially for multi-carrier WCDMA/CDMA/OFDM signals. The high PAR will strictly limit the linearity requirements of the power amplifier, so that the power amplifier used by the base station must amplify the signal with high PAR and send the amplified signal, which reduces the power and efficiency of the base station amplifier output.

In order to reduce the average power loss of the power amplifier, it is necessary to perform peak (clipping) processing on the signal waveform of the multi-carrier. The simplest direct clipping method in the prior art is the hard cutting method, which directly cuts the amplitude of the signal waveform and keeps the phase unchanged. This method has little effect on EVM (Error Vector Magnitude), but the disadvantage is that the clipping method will have sharp edges and sharp peaks in the signal, sudden changes in clipping and short duration of clipping edges. Significant out-of-band spectral anomaly signals, such as spectral distortion, adjacent band interference, spectrum spreading, etc., are generated, which reduces the transmission quality of the signal.

Another clipping method in the prior art is carrier phase shifting processing. The method uses backwards and multiple iterations to optimize the optimal initial relative configuration of each carrier, so that the combined peaks of multiple carriers are reduced to The lowest, the purpose of multi-carrier clipping. This method does not affect the original spectrum characteristics, and can also reduce PAR, but the disadvantage is that the delay is long, the implementation is complicated, and it is necessary to go through multiple optimizations, and at different time slots and different code channel numbers. Underneath, the phase of each carrier needs to be re-adjusted, which is not conducive to receiving device processing.

Summary of the invention

Embodiments of the present invention provide a method and apparatus for clipping a signal to reduce the peak-to-peak ratio of the signal to improve the output power and efficiency of the base station power amplifier.

In order to achieve the above object, an embodiment of the present invention provides a clipping processing method for a signal, including:

Detecting a peak point signal in the input signal that exceeds a preset threshold;

Clipping noise is generated based on the peak point signal and the threshold exceeding the threshold, and the input signal is clipped according to the clipping noise.

Embodiments of the present invention also provide a signal clipping processing apparatus including a peak detection module and a clipping processing module:

The peak detecting module detects a peak point signal of the input signal that exceeds a preset threshold;

The clipping processing module generates clipping noise according to a peak point signal and a threshold detected by the peak detecting module exceeding a threshold, and performs clipping processing on the input signal according to the clipping noise.

Embodiments of the present invention have the following advantages over the prior art:

In the embodiment of the invention, the peak clipping method is used to perform clipping processing on the peak point signal exceeding the preset threshold, thereby effectively reducing the peak-to-level ratio of the signal in the system, and improving the output power of the base station power amplifier on the basis of satisfying the protocol. And efficiency, effectively saving system resources.

DRAWINGS

1 is an architectural diagram of a signal clipping processing method in Embodiment 1 of the present invention; FIG. 2 is a schematic diagram of signal peaks and preset threshold values in Embodiment 1 of the present invention; FIG. 3 is a first embodiment of the present invention; Flow chart for mid-peak detection; 4 is a flow chart showing the update of the peak processing queue in the first embodiment of the present invention; FIG. 5 is a schematic structural view of the clipping filter module in the first embodiment of the present invention; and FIG. 6 is a peak clipping processing in the first embodiment of the present invention. Schematic diagram

7 is a schematic structural view of a clipping filter module according to Embodiment 2 of the present invention; FIG. 8 is a schematic structural view of a multi-stage clipping processing in Embodiment 3 of the present invention; FIG. 9 is a signal cutting in Embodiment 4 of the present invention. A structural diagram of a wave processing device. detailed description

The embodiments of the present invention will be further described below in conjunction with the accompanying drawings and embodiments. In the first embodiment of the present invention, an architecture diagram of a signal clipping method is shown in FIG. 1 , taking the three carrier signals in the original signal as an example, mainly including two steps of carrier combining processing and clipping noise generation. The multi-carrier signal processing method is the same as described in the following steps.

Specifically, the carrier combining process of the signal includes the following steps:

Shape Filtering: This process can be done in the Shape Filtering module shown in Figure 1. The purpose is to convert the data to be transmitted into a signal suitable for transmission on the channel, reducing inter-symbol interference.

Variable Rate Processing: This process can be done in the Upsampling module as shown in Figure 1 to increase the sample rate of the signal.

Carrier Frequency Shift: This process can be multiplied with the Oscillator Freq Shift module after the signal has undergone variable rate processing as shown in Figure 1.

Combined processing: After the above steps, the three signals are added to obtain the combined multi-carrier signal.

Specifically, the process of generating clipping noise includes the following steps:

Weighting factor calculation: Calculate the share of different signals in the clipping noise signal according to the power of different signals, and send the combined signal to the Clipping Noise Generation module. The stronger the power of the signal, the greater the share of the resulting clipping noise signal. Clipping noise generation: This process can be done in the Clipping Noise Generation module as shown in Figure 1.

Clip Filtering Process: This process can be done in the Optimum Clipping Filtering module shown in Figure 1 to filter the clipping noise generated in the Clipping Noise Generation module.

After obtaining the combined multi-carrier signal and the clipped filtered clipping noise signal, the delay module adjusts the delay difference of the two signals after different processing. After the delay difference is eliminated, the two signals are superimposed and combined to obtain the clipped multi-carrier signal.

In the first embodiment of the present invention, in the peak clipping algorithm architecture, the clipping noise generation of the Clipping Noise Generation module and the clipping processing of the Optimum Clipping Filtering module are performed. In the description, a method of peak clipping is proposed, which only clips the peak signal above the threshold, that is, for the signal shown in Figure 2, only three peak signals above the threshold. Process it. The threshold value can be set in advance according to actual needs. The clipping process includes main processes such as peak detection, clipping noise generation, clipping noise queue update, storage compensation queue update, and signal output after clipping.

Figure 3 shows the peak detection process during this process:

Step s301, detecting a peak value, the implementation process is shown in the upper part of FIG.

First, calculate the mode of each point in the signal according to Input_I and Input_Q, and then compare the value of the mode with a preset threshold Limit. When the threshold is higher than the threshold, the mode of the signal point is detected at the same time. If the mode of the previous signal point and the next signal point are higher than the same, the signal point is a peak point, and the next processing of the peak point above the threshold value is required. If the mode of the signal point is lower than the threshold, or the mode of the signal point is higher than the threshold but not the peak point, continue to detect the next signal point, and output the signal point to the Compensation Memory Queue ). The input signals Input_I and Input_Q in this step may be the original signal, or may be the superposition of the original signal and the compensation signal of the current signal by the clipping noise generated by the previous several clippings.

Step s302, output clipping noise (noise), the implementation process is as shown in the lower right part of FIG. Shown.

When the detected modulus of a signal point (Current) is higher than the threshold Limit and is the peak value, the value of Divider Ksi (the ratio of the threshold value to the peak value) is calculated, Divider Ksi = Limit/Current;

Based on the Divider Ksi value, the value of the clipping noise is calculated. In the figure, Peak-I and Peak-Q are the respective signal points of the peak of the input signal whose modulus exceeds the peak value of the threshold. The method for calculating the value of clipping noise based on the peak value of the input signal is:

The real part of the clipping noise Pnoise — 1 = the real part of the peak signal Peak — I χ ( Divider Ksi - 1 );

The imaginary part of the clipping noise Pnoise— Q = the imaginary part of the peak signal Peak— Q ( Divider

Ksi - 1).

At the same time, according to Divider Ksi calculation and output signal to Compensation Memory Queue (compensation queue;):

The real part of the output signal = the real part of the signal Peak-I Divider Ksi;

The imaginary part of the output signal = the real part of the signal Peak- Q Divider Ksi.

Step s303, output Peaks_Queue (peak sequence) and Filter_Coeff_Index_ Queue (filter coefficient index sequence) update signal.

The content in the Peaks-Queue is the clipping noise sequence, and the content in the Filter-Coeff-Index-Queue is the index value of the clipping filter coefficients, which is used to find the filter coefficient table.

After the peak is detected and the Queue update signal is output, the update of the Queue is required. The update process of the Queue in the first embodiment of the present invention is as shown in FIG. 4:

After receiving the Queue update signal described in Figure 3, the Index-Queue Register is updated, and the contents of the Peaks-Queue and Filter-Coeff-Index-Queue are corresponding, and need to be updated synchronously. .

Specifically, as shown in the left half of FIG. 4, the received Pnoise_I and Pnoise_Q are stored in Max 0 and output Max-0, and the next time Pnoise-I and Pnoise-Q are received, the queue pointer is moved. Store in Max 1 and output Max-1 until all Max-Nmax are output. When the queue is full, if the new value is received again, the queue is shifted, and the earliest stored in the queue is discarded. The value entered.

For Filter-Coeff-Index-Queue, when new Pnoise-I and Pnoise Q are detected, as shown in the right half of Figure 4, the new value in Filter-Coeff-Index-Queue (ie the value of Index 0 Ind) — 0 ) The initial value is assigned to Order/2, where Order is the order of the clipping filter. Each time a new signal is received, the original value of Filter_Coeff_Index_Queue is decremented by 1 each time, when decremented to 0. Stop decreasing.

Figure 5 shows a schematic diagram of the processing of the clipping filter, including the index portion, the filter coefficient table, the multiply accumulator portion, and the Compensation Memory Queue portion.

The index part includes a clipping noise queue (Peaks-Queue), a clipping filter coefficient index queue (Filter-Coeff-Index-Queue), and the queue length is M. The value of the queue length M is selected according to actual needs, generally depends on The probability of occurrence at the peak.

The filter coefficient table stores the clipping filter coefficients. When the FIR (Constant Impulse Response) filter is used, the filter coefficient corresponding to the FIR filter is used, and the coefficient table length is the filter order Order/ 2, Order is the order of the clipping filter.

The multiply accumulator is calculated as follows: The values in the Peaks-Quene are multiplied by the corresponding filter coefficients Coeff and accumulated. The filter coefficient Coeff corresponding to the different Peaks values in the queue is obtained according to the Clipping Filter Coefficient Index Queue (Filter_Coeff_Index_Queue). The specific accumulation method is: For the i-th value in the two queues, Coeff_i is obtained by searching the filter coefficient table according to Ind_i in the clipping filter coefficient index queue; calculating MAX_i*Coeff_i; The MAX_i*Coeff-i values are accumulated and output as a post-compensation signal.

The peak clipping filter processing flow is shown in FIG. 5 and FIG. 6, wherein the curve in FIG. 6 represents the time domain of the FIR filter, and the straight line in the vertical direction of the time t-axis represents each sample point of the signal, and the dotted line is connected. The sample points are symmetric about the center signal. The clip processing queue shift direction is from right to left, that is, X (i + order) is the latest input signal point. As shown in Fig. 6, if the processed signal sample point is the peak point, the peak value of the peak point and the subsequent order/2 signals will be sampled during the clipping process. The effect is affected, so pre-compensation (compensation for the previous order/2 signal samples) and post-compensation (compensation for the rear order/2 signals) are required. The latest input signal point in Figure 6 is X(i + order), and the post-compensation is performed before the peak queue detection is sent, that is, the compensation of all the peak points before the current sample point is superimposed. In the process of post-compensation, according to the central symmetry principle of the FIR filter, the signal samples before the peak point are pre-compensated.

Each time an X(i+order) is input, the Peaks-Queue content is multiplied by the corresponding filter coefficient Coeff and accumulated as a post-compensation signal, superimposed with the current input X(i+order), and then sent to the signal peak detection module. Wherein, the corresponding filter coefficient Coeff is searched in the filter coefficient table by the index value in the Filter_Coeff_Index_Queue. When the index value is 0, the filter coefficient Coeff is set to 0. The signal peak detection module output signal is sent to the compensation queue.

Filter— Coeff— Index—The index value in the Queue is also used to calculate the index of the input C(k) to be compensated in the Compensation Memory Queue, k=i+order-2*Index-l„ Compensation The queue stores the Order signals to be compensated, and the outputs of the M complex multipliers are added as the pre-compensation signal and the corresponding input C(k) values in the queue, and the result is stored back in the compensation queue. Each input X(i) +order), output C(i) in the compensation queue as the output signal after clipping, and compensate the queue for shifting.

In the first embodiment described above, the pre-compensation compensates for the Order Compensated Input C stored in the compensation queue. The pre-compensation signal comes from the output of M complex multipliers, where the inputs of the M complex multipliers are Peaks-Queue. Therefore, the calculation of the pre-compensation signal depends on the clipping noise of the previous M post-compensated peak signals. That is, the calculation of the compensation signal of the current input signal needs to first calculate the clipping of the previous M post-compensated peak signals. noise. This results in greater resource consumption and complexity in computational and logical implementations.

To this end, Embodiment 2 of the present invention proposes an optimized clipping processing method. Use the original input X to generate the corresponding all or several of the clipping noise. That is, peak detection is performed on the original input X, and several pieces of clipping noise are generated in the same manner as in FIG. A plurality of post-compensation signals for the current input signal are calculated based on the plurality of clipping noises. These several post-compensation signals are used as prediction values for the real post-compensation signal, Add to the current input signal X and detect if the signal peak point exceeds a preset threshold. See Figure 7, the structure is similar to Figure 5, except that Max T ~ Max N of the original peak queue is used as the predicted value of the real post-compensation signal, superimposed on the current input signal X. The next steps are similar to those described in Figure 5. The peak detected signal is input to the compensation queue, and each clipped noise is multiplied by the corresponding filter coefficient as a pre-compensation signal and added to the corresponding value in the compensation queue. The signal in the updated compensation queue is output as the clipped signal. Because of the new optimized clipping method, the calculation and logic are more convenient, and the performance remains unchanged.

To further optimize the clipping performance, the peak clipping algorithm can perform multiple levels of processing to optimize the output PAR. In the third embodiment of the present invention, a multi-stage clipping processing architecture is shown in FIG. 8. In the application, the number N of clipping processing can be determined according to system resources and processing performance.

Embodiments 1 to 3 of the present invention propose a clipping algorithm processing method for a multi-carrier signal.峰值Using the peak clipping method, only the peak exceeding the threshold is clipped, and the clipping noise is generated by detecting and clipping the peak. Only the wavelet filter impulse response generated by peak clipping noise is stored and calculated, which greatly reduces the number of multipliers and saves logic resources. At the same time, on the basis of reducing resources, multi-level clipping can be used to optimize performance.

Embodiment 4 of the present invention provides a signal clipping processing apparatus, as shown in FIG. 9, comprising: an input module 10, a peak detecting module 20, and a clipping processing module 30.

The input module 10 is configured to send the multi-carrier signal that needs to be processed to the peak detecting module 20.

The peak detecting module 20 is configured to perform peak detection on the multi-carrier signal sent by the input module 10, and notify the clipping processing module 30 to perform clipping processing when detecting a peak signal exceeding the threshold. The original input signal sent by the input module 10 or the original input signal compensated by the post-compensation signal generated by the compensation queue sub-module 34.

The clipping processing module 30 is configured to perform a clipping process on the signal detected by the peak detecting module 20 as a peak point exceeding the threshold to obtain a signal after the clipping process.

Specifically, the clipping processing module 30 further includes a clipping noise generating sub-module 31, a clipping noise queue sub-module 32, a filter coefficient index queue sub-module 33, and a compensation queue. Submodule 34 and output submodule 35.

The clipping noise generating sub-module 31 is configured to generate clipping noise according to the signal of the peak point exceeding the threshold detected by the peak detecting module 20, and generate an update signal to be sent to the clipping noise index sub-module 32 and the filter coefficient index queue. Sub-module 33, and generates a compensation signal to send to compensation queue sub-module 34.

The clipping noise queue sub-module 32 is configured to update the internally stored clipping noise queue according to the content sent by the clipping noise generation sub-module 31, and discard the earliest input value in the queue when the queue is full.

The filter coefficient index queue sub-module 33 updates the clipping filter coefficient index queue according to the content sent by the clipping noise generation sub-module 31, and the clipping in the clipping filter coefficient index queue and the clipping noise queue sub-module 32 The noise queue sub-module corresponds; when the queue is full, the oldest input value in the queue is discarded.

The compensation queue sub-module 34 is configured to perform compensation according to the clipping noise queue stored in the clipping noise queue module 32, the filter coefficient index queue stored by the filter coefficient index queue sub-module 33, and the clipping noise generation sub-module 31. The signal generation compensation queue compensates for the input signal of the initial input module 10 and the signal output by the peak detection module 20. And outputting the clipping processing signal to the output sub-module 35. Specifically, according to the clipping filter coefficient index of the filter coefficient index queue sub-module 33, the filter coefficient table is obtained to obtain a filter coefficient corresponding to the clipped filter coefficient index; the clipping noise queue sub-module 32 is The clipping noise is multiplied by the corresponding filter coefficient and accumulated as a post-compensation signal, and the detected signal, which is superimposed with the input signal in the input module 10, is sent to the peak detecting module 20 as a new input. In addition, each clipping noise in the clipping noise queue sub-module 32 is multiplied by a corresponding filter coefficient to be a pre-compensation signal, and is added to the corresponding value of the compensation queue in the module, and the compensation queue in the module is updated. . Each time a new compensation signal is input to the end of the compensation queue in the module, a signal is output from the compensation queue leader to the output sub-module 35.

The output sub-module 35 is configured to output the signal after the clipping process.

The clipping processing device described in the fourth embodiment only stores and calculates the clipping response of the clipping filter generated by the peak clipping noise, greatly reduces the number of multipliers, and saves logic resources. At the same time, on the basis of reducing resources, multi-level clipping can be used to optimize performance.

Through the description of the above embodiments, those skilled in the art can clearly understand that the present invention can be implemented by hardware or by software plus necessary general hardware platform. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which may be stored in a non-volatile storage medium (which may be a CD-ROM, a USB flash drive, a mobile hard disk, etc.), including several The instructions are for causing a computer device (which may be a personal computer, server, or network device, etc.) to perform the methods described in various embodiments of the present invention.

In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the scope of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and scope of the present invention are intended to be included within the scope of the present invention.

Claims

Rights request

A method for clipping a signal, comprising:

2. The method of clipping a signal according to claim 1, wherein the step of detecting a peak point exceeding a preset threshold in the detection signal is specifically:

Whether the mode of the peak point exceeds the threshold in the composite signal of the current input signal or the clipping noise generated by the previous clipping on the current input signal and the current input signal.

3. The method of clipping a signal according to claim 1, wherein said generating a clipping noise based on a peak point signal and a threshold exceeding a threshold, and clipping said signal according to said clipping noise The steps specifically include:

Generating a clipping noise queue, a clipping filter coefficient index queue, and a compensation queue according to the peak point signal and the threshold exceeding the threshold;

Updating the clipping noise queue and the clipping filter coefficient index queue; updating the compensation queue according to the updated clipping noise queue and the clipping filter coefficient index queue, and outputting the compensation queue by the compensation queue Wave processed signal.

4. The method of clipping a signal according to claim 3, wherein the step of generating a clipping noise queue comprises:

Obtaining a peak point signal and a threshold exceeding the threshold, and a ratio of the threshold to a peak point mode exceeding the threshold;

And clipping noise is generated according to the peak point signal exceeding the threshold and the ratio; and the clipping noise is input to the clipping noise queue.

The method for processing a clipping of a signal according to claim 3, wherein the step of generating a clipping index of the clipping filter coefficient comprises:

Obtaining a clipping filter coefficient index corresponding to the generated clipping noise; inputting a clipping filter coefficient index corresponding to the clipping noise into a clipping filter coefficient index queue.

The method for processing a signal clipping according to claim 3, wherein the step of generating a compensation queue comprises:

If the modulus of the detected signal is lower than the threshold, or the modulus of the detected signal is higher than the threshold but not the peak point, the signal is directly input to the compensation queue;

If the modulus of the detected signal exceeds the threshold and is a peak point, the peak point signal exceeding the threshold and the ratio are calculated and input to the compensation queue.

7. The method of clipping a signal according to claim 3, wherein the step of updating the clipped noise queue and the clipped filter coefficient index queue comprises:

And storing the newly obtained clipping noise to the tail of the queue of the clipping noise queue, storing the clipping filter coefficient index corresponding to the newly obtained clipping noise to the tail of the queue of the clipping filter coefficient index queue When the clipping noise queue or the queue of the clipping filter coefficient index queue is full, the signal of the head of the team is discarded.

8. The method of clipping a signal according to claim 3, wherein said updating said compensation queue is performed based on said updated clipping noise queue and a clipping filter coefficient index queue, said compensation The queue output clipping processed signal specifically includes: according to the clipping filter coefficient index in the clipping filter coefficient index queue, the query filter coefficient table is obtained corresponding to each of the clipping filter coefficient indexes Filter coefficient

The clipping noise generated by the current input signal compensated by the post-compensation signal, and/or the clipping noise generated by the current input signal, is multiplied by a corresponding filter coefficient and accumulated as a post-compensation signal, After superimposing with the current input signal as a new input signal, detecting whether the peak value of the signal exceeds the threshold;

And multiplying the clipping noise by a corresponding filter coefficient as a pre-compensation signal, and adding the corresponding value in the compensation queue to update the compensation queue;

Each time a new compensation signal is input to the end of the compensation queue, a signal is output from the compensation queue head as a signal output after clipping.

9. The method of clipping a signal according to claim 1, wherein after outputting the clipped signal, the method further comprises: The clipped signal is used as an input signal, and one or more clipping processes are performed.

10. A signal clipping processing device, comprising: a peak detection module and a clipping processing module:

The clipping processing device of claim 10, wherein the clipping processing module further comprises:

a clipping noise generating sub-module, generating clipping noise according to a signal detected by the peak detecting module exceeding a peak point of the threshold, and generating an update signal to be sent to the clipping noise index sub-module and a filter coefficient index queue Module, and generating a compensation signal to send to the compensation queue sub-module;

a clipping noise queue sub-module, updating the stored clipping noise queue according to the content sent by the clipping noise generating sub-module;

a filter coefficient indexing queue sub-module, updating a clipping filter coefficient index queue according to the content sent by the clipping noise generating sub-module, wherein the clipping filter coefficient index queue corresponds to the clipping noise queue sub-module;

a compensation queue sub-module, generating a compensation queue according to the content stored in the clipping noise queue module and the compensation signal sent by the clipping noise generating sub-module, compensating the initial input signal, and outputting the clipping processing signal to Output submodule;

Output sub-module, output the signal after clipping.

12. The clipping processing apparatus of claim 11, further comprising: an input module for receiving an input signal and accepting compensation by the compensation queue sub-module for its input signal.