KR101613731B1 - Apparatus and method for extracting reference cancellation pulse in multi-carrier system and peak cancellation crest factor reduction system for the same - Google Patents

Abstract

 A reference cancellation pulse extracting apparatus and method for extracting a reference cancellation pulse for counting an offset signal from an input signal in real time based on the reference cancellation pulse, and a PC-CFR system therefor are disclosed. According to this apparatus and method, based on the peak position, size, and phase information of an input signal, sampling is performed with respect to each peak and each sampling signal is normalized so that the magnitude and phase of the peak point become the reference, and the normalized signals are accumulated And a reference offset pulse is generated by obtaining an average value. The reference offset pulse has a peak size of 1 and has the same frequency characteristics as the input signal.

CFR, PAPR, offset pulse, coefficient, PC-CFR, peak

Description

BACKGROUND OF THE INVENTION 1. Field of the Invention [0001] The present invention relates to a reference canceling pulse extracting apparatus, a method thereof, and a PC-CFR system for the same. 2. Description of the Related Art [

The present invention relates to a PC-CFR (Peak Cancellation Crest Factor Reduction) structure for lowering a Peak to Average Power Ratio (PAPR) in a multi-carrier system. More particularly, A reference offset pulse extracting device for extracting a reference offset pulse for counting the offset pulse from the input signal and generating an offset pulse based on the extracted reference offset pulse, and a PC- CFR system.

In order to improve the efficiency of the wireless communication system, there is a need to reduce the PAPR of the signal. To this end, a commonly used method is the Crest Factor Reduction (CFR). The CFR reduces the size of the signal within a range that does not affect the frequency band when a large-sized signal occurs instantaneously in the baseband signal. Among the various CFR implementation methods, PC-CFR is used which is excellent in performance and easy to implement. PC-CFR is a method of removing a peak of a signal using a cancellation pulse generated by CPG.

The driving process of the conventional PC-CFR will be described as follows.

Conventionally, a cancel pulse is added to a position of a signal exceeding a predetermined threshold to prevent the threshold from being exceeded. The offset pulses are generated from a cancellation pulse generator (CPG). The CPG selects a peak from an input signal (a signal having a larger PAPR) than a threshold value, and calculates a peak based on the position, And produces an offset pulse. The CPG uses the coefficients of the offset pulses (for example, 255 (0 to 254) filter coefficients when the length of the offset filter is 255) in generating the offset pulses.

One CPG can produce one offset pulse for peak signal 1. Also, since one CPG in operation can not generate a canceling pulse for another peak signal 2 while generating one canceling pulse, the next CPG is allocated to the peak signal 2. In other words, another CPG (CPG2) is needed to remove another peak signal 2 that may appear while CPG1 produces a canceling pulse for peak signal 1, and the number of CPGs can be tailored to the system, The number of CPGs can not be increased without increasing the number of CPGs. For example, if there are four CPGs for generating offset pulses, for example, if there are four CPGs, up to four CPGs can be operated at one time, and there are four offset pulses that can operate continuously. The reason why there are many GPGs is to process the peaks at the same time when the peaks are detected at various positions as described above. The cancellation pulses generated by the four CPGs are all added together (so that each offset pulse does not overlap because the time domain is different) so that the peaks in the original signal are removed (offset) by the opposite magnitude. In other words, the peak is removed by subtracting the offset pulse at the corresponding position in order to remove the peak occurring in the input signal.

In order to generate the offset pulse in the CPG, it is necessary to know the cancellation pulse coefficient and to know the upper information about the input signal. That is, as the upper information of the input signal, it is necessary to know the bandwidth of the signal, the number of the currently used carriers, the frequency position, and so on. The coefficients of the offset pulses are calculated or stored in the processor and are provided to the CPG according to the conditions of the input signal.

The process of calculating the coefficients of the offset pulses will be described as follows.

The offset pulse is generated in the CPG with the position, size, and phase information of the peak. The offset pulse has the same frequency characteristic as the input signal and has a center peak, which is added to the peak signal in reverse to attenuate the peak signal.

First, the offset pulse for one carrier is obtained. The spectrum of the offset pulses is obtained in the form of a low-pass filter so as to be a carrier having a frequency of zero. This is so that adding the offset pulses does not affect the spectrum. Assuming that the coefficient of the offset pulses of one carrier is CP (n) (n: 0 to 20) (this signal is such that the center point n is 10 and the peak is 1) The coefficients of the canceling pulses of two carriers of f1 and f2 are obtained as follows.

Exp (i * f2 * (n-10) ") of the tone signal C2 having the frequency f2 and the tone signal C1 having the frequency f1 as "exp (i * f1 * * 2 * pi / fs) ", C1 and C2 are Tone signals having frequencies of f1 and f2, respectively, when n is 10.

If we say "C3 (n) = C1 (n) + C2 (n)", C3 becomes a two-tone signal with frequency f1 and f2. (N) = CP (n) = C3 (n) * CP (n), CP2 (n) is a 2-carrier signal with frequencies f1 and f2, The size is 2. In the time domain, this signal has the maximum value when n = 10, and its value is 2. By dividing this signal by the maximum value of 2 and making it the signal with the maximum value of 1, it becomes the desired offset pulse. When 255 samples are taken for this offset pulse, 255 (0 to 254) filter coefficients (i.e., the coefficients of the offset pulses) are obtained.

However, conventionally, every time the state of the input signal is changed (whenever the upper information is changed), the processor has to recreate the coefficients of the offset pulses through the above process and provide them to the CPG every time. If the processor calculates the coefficients of the offset pulses each time the state of the input signal is changed, not only the burden on the CFR control of the processor is increased, but also an additional Putting a processor is also inefficient.

It is an object of the present invention to extract a reference offset pulse for the counting of the offset pulse from the input signal in real time without knowing the upper information about the input signal (the bandwidth of the signal, the number of carriers and the frequency position, etc.) And to provide a PC-CFR system for the reference cancellation pulse extracting apparatus and method therefor.

According to one aspect of the present invention, there is provided a reference cancellation pulse extracting apparatus and method for extracting a reference cancellation pulse for counting cancellation pulses from an input signal in real time, A CFR system is disclosed. According to this apparatus and method, based on the peak position, size, and phase information of an input signal, sampling is performed with respect to each peak and each sampling signal is normalized so that the magnitude and phase of the peak point become the reference, and the normalized signals are accumulated And a reference offset pulse is generated by obtaining an average value. The system is a PC-CFR system that reduces the size of a corresponding signal within a range that does not affect a frequency band when a signal having a large magnitude instantly occurs in a baseband signal. The system includes a peak position and size of an input signal, , Sampling each of the peaks and normalizing each sampling signal so that the magnitude and phase of the peak point are the reference, accumulating the normalized signals to obtain a mean value, and generating a reference offset pulse, An offset pulse coefficient extracting unit for extracting a coefficient of the offset pulse; A peak detector for detecting a peak of an input signal exceeding a threshold value; At least one offset pulse generator for generating offset pulses for the peaks detected by the peak detector, respectively, based on the coefficients of the offset pulses; A summation unit for summing at least one offset pulse generated by the at least one offset pulse generator; And an adder for subtracting the summation pulse summed by the summation unit from the signal delayed by the original input signal to attenuate the pulse with the corresponding offset pulse.

According to the present invention, there is no need to calculate the coefficients of the offset pulses every time the frequency characteristics of the input signal are changed, and it is possible to cancel out the input signal without upper information (bandwidth of the signal, It is possible to generate a pulse coefficient, and there is an advantage that the configuration of the CFR is simplified.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings. In the following description, well-known functions or constructions will not be described in detail if they obscure the subject matter of the present invention.

1 is a diagram illustrating a structure of a PC-CFR system in which the present invention can be implemented.

The peak detector 12 detects a peak value Peak larger than a threshold value (| Sin | > | Vth |) when an input signal having a large PAPR (which is a multicarrier signal with intervals) is received and outputs each peak to the offset pulse generator CPG) 13 to generate a cancel pulse for each peak in the CPG 13. [

While one CPG 13 operates on the first peak 1 that is larger than the threshold value to produce one offset pulse (CPG Busy), for example, 254 samples from the peak 1 (when the offset pulse length is 255, 0 to 254 Peak 1 is assumed to be the 0 < th > sample), it is impossible to generate another offset pulse for the peaks 2 and 3. Therefore, when the peak 1 is assigned to CPG1, And peak 3 is assigned to CPG2 and CPG3, respectively. If there are 4 CPGs making offset pulses, up to 4 CPGs can be operated at a time within 255 samples to produce four offset pulses in succession. That is, when one peak larger than the threshold value is generated within 255 samples, the CPG1 is allocated to the CPG1, and if there are several peaks larger than the threshold value within 255 samples, the respective peaks are sequentially assigned to the CPG1,2,3,4. Although four CPG configurations are taken as an example, eight CPGs, four CPGs (one stage), four CPGs (two stages), sixteen CPGs, four CPGs (one stage), and four CPGs ) + .... + Four CPGs (four stages) are possible.

The peak detector 12 outputs peak information for the peak 1 (e.g., the 0th sample), the peak 2 (e.g., the 50th sample), and the peak 3 (e.g., the 150th sample) generated within 255 samples to the corresponding CPG 13 Peak 1 indicates CPG1, Peak 2 indicates CPG2, and Peak 3 indicates CPG3.

In each CPG 13, offset pulses 1 to 3 for the peaks 1 to 3 generated within 255 samples are generated using the coefficients of the offset pulses (filter coefficients of 0 to 254). At this time, the coefficients of the offset pulses (for example, 255 (0 to 254) filter coefficients) are provided in the offset pulse coefficient extracting unit 20. The process of generating the offset pulses to attenuate the peaks in the original signal will be described later.

In the present invention, a canceling pulse coefficient extracting unit (a reference cancellation pulse extracting apparatus of the present invention) 20 first extracts an offset pulse from an input signal (which is different from canceling pulses 1 to 3 generated in CPG1 to 3, (Hereinafter, referred to as a 'reference offset pulse'), and 255 samples (for example, 0 to 254) of filter coefficients (that is, coefficients of the offset pulses) . At this time, the reference offset pulses generated by the offset pulse extractor 20 are generated in accordance with the characteristics of the input signal, and the coefficients of the reference offset pulses are used by the CPG 13 to generate the offset pulses 1 to 3. The reference offset pulse has a peak size of 1 and has the same spectral characteristics as the input signal.

The process of obtaining the reference offset pulses is as follows.

The normalization unit 21 of the offset pulse coefficient extraction unit 20 receives the position, size, and phase information of the peak from the peak detection unit 12, and cuts out a signal of a predetermined length centered on the peak, And normalized to be a reference (see Fig. 2). For example, a signal having a size of 255 samples centered on the peak 1. Likewise, for the peaks 2, 3, ..., N (N is the number of the peak), a signal having a 255-sample size centered at each of the peaks 2, 3, ..., N is generated. For example, a peak 1 (for example, a size of 10, a phase of 6 + 8j (I = 6, Q = 8) (For example, the size is '8') is the 150th sample, the peak 4 (for example, the size is 7.5) is the 350th sample, ..., the peak N is M (M is the position of the peak as the sample number) (For example, the size is 7), the left and right 127 samples are taken centering on the 0th, 50th, 150th, 350th, ... Mth samples, Sampling signals 1, 2, 3 ... N). Then, the sampling signal is divided by the peak value so that the peak of each sampling signal is 1. Divide the sampling signal 2 by the size 9, divide the sampling signal 3 by the size 8, and divide the sampling signal 4 by the size 7.5, for example, by dividing the sampling signal 1 (255 sample signals taking 127 samples on the left and right sides for the peak 1) . When the sampling signal is divided by the magnitude of the peak, the peak point size is 1 and the signal is normalized to the same phase as the input signal.

The averaging unit 22 sufficiently accumulates (adds) the normalized signal (the signal having the peak point size of 1 and the phase of the input signal) to obtain an average value (see Fig. 2). This average value is a reference offset pulse.

Specifically, a desired offset pulse can be obtained by cutting out a portion determined to be a peak point in a continuously input signal (255 sample signals obtained by taking 127 samples on the left and right sides of the peak) and matching the peak of the sampling signal by 1 have. Assuming that the input signal is s (n) and the position of the selected peak is n1, n2, n3, n4, ... and the length of the reference offset pulse is 255, the absolute magnitude of s (n) 3 is the same as FIG. s (n1) is a peak signal having a length of 255 from s (n1-127) to s (n1 + 127). s (n1-127) ... s (n1 + 127)} / s (n1) to divide the peak of s (n1) n3, ..., and if all of them are added, they are incremented by 1 at the peak point and finally canceled out at the other intervals. The signal normalized with the size of 1 divided by the peak value is accumulated A reference cancellation pulse as shown in the following Equation 1 can be obtained.

Equation 1 is obtained by adding 10 intervals from s (n1) to s (n10) where a peak occurs and averaging to obtain a reference offset pulse having a length of 255 called cp (m). Here, we assume that the peak is 10, but it will converge as we create a large number of averages. This is as shown in Fig.

The reference offset pulses obtained through this process have a peak size of 1 and have the same spectral characteristics as the input signal. The reference cancellation pulse must be converged to zero at both ends so that no discontinuity occurs when the reference cancellation pulse is used to cancel the signal. To this end, a filter (e.g., Kaiser Filter 23) If the signals generated by averaging these multiple peak intervals are multiplied by a signal (Kaiser filter coefficient) such as a Kaiser filter, the peak point remains unchanged and the size of the side-robe portion (0 to 254) filter coefficients (i.e., the coefficients of the offset pulses) can be obtained by taking 255 samples as reference offset pulses, and the coefficients of the offset pulses can be obtained for each CPG (13).

Referring back to FIG. 1, the offset pulse generator 13 generates offset pulses for the peaks detected by the peak detector 12. Peak 1 (0th sample) assigns CPG1, Peak 2 (50th sample) assigns CPG2, and Peak 3 (150th sample) assigns CPG3 when peaks 1 through 3 occur within 255 samples. Each CPG generates offset pulses 1 to 3 based on the coefficients of the offset pulses.

The summing unit 14 adds all of the offset pulses 1 to 3 generated in each CPG 13. At this time, the offset pulses 1 to 3 do not overlap each other because the time domains are different. The offset pulse 1 for the 0th sample, the offset pulse 2 for the 50th sample, and the offset pulse 3 for the 150th sample.

The delay unit 11 delays the original signal in consideration of the time for CFR processing.

The adder 15 adds (or subtracts) the delayed original signal and the offset pulses 1 to 3 to the opposite magnitudes. That is, the offset pulse 1 is subtracted for the peak 1 (0th sample), the offset pulse 2 is subtracted for the peak 2 (50th sample), and the offset pulse 3 is subtracted for the peak 3 (150th sample).

For example, if the threshold value of the CFR is defined as 7 and the signal 41 shown in FIG. 4 is input to the CFR input (6 + 8j (I = 6, Q = 8)) Since the signal is 10 at the peak point, the offset pulse should be subtracted so that it is reduced by 3. Since the reference cancellation pulse having the size of 1 in the peak is already known (the reference cancellation pulse is a value obtained in real time from the input signal by the cancellation pulse coefficient extracting unit 20, and the size is 1 at the peak) (42) to match the peak point (43). In practice, both the input signal and the canceling pulse are complex signals, so that the phase is equalized at the peak.

Here, the offset pulse is generated in the CPG 13, and the phase is the same as the input signal 6 + 8j and the size is 3. (1.8 + 2.4j) (42) can be obtained by multiplying the reference offset pulse having (6 + 8j) * (3/10) = 1.8 + 2.4j .

Therefore, it is possible to extract the reference offset pulses from the offset pulse coefficient extracting unit 20, obtain the coefficients of the offset pulses from the reference offset pulses, and inform the respective CPGs 13 of the offset canceling pulses. However, In each CPG 13, a reference offset pulse may be N times (N is a positive integer) to generate a canceling pulse.

Although the present invention has been described in connection with certain embodiments thereof, it should be understood that various changes and modifications may be made therein without departing from the spirit and scope of the invention as will be apparent to those skilled in the art to which the invention pertains. something to do. It is also contemplated that such variations and modifications are within the scope of the claims appended hereto.

BRIEF DESCRIPTION OF THE DRAWINGS Figure 1 is a diagram illustrating the structure of a PC-CFR system in which the present invention may be implemented.

FIG. 2 illustrates a process of generating a reference offset pulse according to an embodiment of the present invention; FIG.

3 illustrates a normalized signal in a time domain in accordance with an embodiment of the present invention.

4 is a view for explaining a CFR process according to an embodiment of the present invention;

DESCRIPTION OF THE REFERENCE NUMERALS

11: Delay unit 12: Peak detection unit

13: offset pulse generator 14: summing unit

15: adder 20: offset pulse coefficient extracting unit

Claims (13)

An apparatus for extracting reference offset pulses for the counting of offset pulses in a multi-carrier system, A normalization unit for sampling each of the plurality of sampling signals around each peak based on the peak position, size, and phase information of the input signal, and normalizing each sampling signal so that the size and phase of the peak point are the reference; And And an averaging section for generating a reference offset pulse by accumulating the signals normalized by the normalizing section and obtaining an average value. The method according to claim 1, Further comprising a filter for multiplying the filter coefficient of the reference cancellation pulse to reduce the size of both side lobe portions while keeping the peak point of the reference cancellation pulse intact. 3. The method of claim 2, The filter may be a Kaiser filter, a reference offset pulse extractor. 4. The method according to any one of claims 1 to 3, Wherein the reference cancellation pulse has a peak size of 1 and has the same frequency characteristics as the input signal. 5. The method of claim 4, Wherein the sampling is performed with a signal having a size of 255 samples centered on a peak. 5. The method of claim 4, Wherein the normalizing unit normalizes each sampling signal to a peak size of 1. CLAIMS 1. A method for extracting a reference offset pulse for a coefficient of a cancellation pulse in a multi- a) sampling each center of each peak based on the peak position, size, and phase information of the input signal, and normalizing each sampling signal so that the size and phase of the peak point are the reference; And b) generating a reference offset pulse by accumulating the normalized signals to obtain an average value. 8. The method of claim 7, c) multiplying the filter coefficient of the reference cancellation pulse to reduce the size of both side lobe portions while leaving the peak point of the reference cancellation pulse intact. 9. The method of claim 8, The reference cancellation pulse has a peak size of 1 and has the same frequency characteristics as the input signal, The method of claim 1, wherein in step (a), a signal of a 255-sample size centered on a peak is sampled, and each sampling signal is normalized to a peak size of 1. A PC-CFR (Peak Cancellation Crest Factor Reduction) system that reduces the size of a signal within a range that does not affect the frequency band when a signal having a large magnitude instantly occurs in the baseband signal, Based on the peak position, size, and phase information of the input signal, sampling is performed with respect to each peak, and each sampling signal is normalized so that the magnitude and phase of the peak point become a reference. By accumulating the normalized signals and obtaining an average value, An offset pulse coefficient extraction unit for generating a pulse and extracting a coefficient of the offset pulse based on the reference offset pulse; A peak detector for detecting a peak of an input signal exceeding a threshold value; At least one offset pulse generator for generating offset pulses for the peaks detected by the peak detector, respectively, based on the coefficients of the offset pulses; A summation unit for summing at least one offset pulse generated by the at least one offset pulse generator; And And an adder for subtracting the summation pulse summed by the summation unit from the signal delayed from the original input signal to attenuate the pulse with the corresponding offset pulse. 11. The method of claim 10, Wherein the canceling pulse coefficient extracting unit further has a function of multiplying the filter coefficient of the reference cancellation pulse and reducing the size of both side lobe parts while keeping the peak point of the reference cancellation pulse as it is. 12. The method of claim 11, The reference cancellation pulse has a peak size of 1 and has the same frequency characteristics as the input signal, Wherein the canceling pulse coefficient extracting unit cuts and samples a signal having a size of 255 samples centered on a peak and normalizes each sampling signal to a size of a peak to 1. The PC- 13. The method according to any one of claims 10 to 12, Wherein the input signal is a multi-carrier signal having an interval.

Images