KR20080057122A - Apparatus and method for transmitting data using a plurality of carriers - Google Patents

Abstract

An apparatus and a method for transmitting data using plural carriers are provided to design a sequence used for a specific channel of OFDM(Orthogonal Frequency Division Multiplexing) by utilizing a sequence generated from a time domain. In a method for selecting sequences used for a specific channel of OFDM(Orthogonal Frequency Division Multiplexing), a sequence index among plural sequence indexes which identify a sequence with an L length is determined as a mother sequence index. At least one the rest sequence index is determined using an integer multiplication value on half of a period and the mother sequence index. Sequences corresponding to at least two sequence indexes among the mother sequence index and the rest sequence index are selected as the sequences for the specific channel(S10).

Description

Apparatus and method for transmitting data using a plurality of carriers}

1 shows a structure of a downlink subframe of the IEEE 802.16 system.

FIG. 2 is a diagram illustrating a subcarrier set for transmitting a preamble transmitted in a zeroth sector in an IEEE 802.16 system.

3 and 4 illustrate various methods of including the P-SCH and the S-SCH in a radio frame.

5A is a diagram illustrating a transceiver structure to which an embodiment of the present invention is applied.

FIG. 5B is a flowchart illustrating a method of designing a time domain of a sequence having a low correlation ratio and maintaining reasonable correlation characteristics according to an embodiment of the present invention.

6 is a diagram illustrating autocorrelation characteristics of the CAZAC sequence.

7 shows one method of configuring a P-SCH.

8 is a flowchart illustrating a method of generating a P-SCH according to the second embodiment.

9 is a diagram illustrating a subcarrier to which a P-SCH of the LTE standard is mapped.

FIG. 10 is a block diagram showing a Frank sequence of length 36 in the time domain.

FIG. 11 is a block diagram illustrating a result of generating a sequence by repeating twice in the time domain according to the second embodiment.

12 is a diagram showing the result of S1703.

13 is a diagram showing the results of step S1704-1.

FIG. 14 illustrates a result of performing a cyclic shift to the right of the result of FIG. 12.

15 is a procedure flowchart for explaining the third embodiment.

16 is a view comparing constellations of a sequence in which a DC component is removed and a sequence in which a DC component is not removed according to the present embodiment.

17 illustrates a method of designing a sequence in the frequency domain.

18 to 19 show the structure of a receiving end according to the fourth embodiment.

20A to 24 are diagrams for describing a sequence according to the present embodiment.

The present invention relates to a method for transmitting and receiving data in an orthogonal frequency division multiplexing (OFDM) based communication system. It is about a design method

Hereinafter, OFDM and OFDMA and SC-FDMA techniques used in the present invention will be described.

Recently, the demand for high-speed data transmission is increasing, and OFDM is a suitable method for such a high-speed transmission, and has been recently adopted as a transmission method of various high-speed communication systems. Hereinafter, orthogonal frequency division multiplexing (OFDM) will be described. The basic principle of OFDM is to divide a high-rate data stream into a large number of low-rate data streams, which are transmitted simultaneously using multiple carriers. Each of the plurality of carriers is called a subcarrier. Since orthogonality exists between the multiple carriers of the OFDM, the frequency components of the carriers can be detected at the receiving side even if they overlap each other. The high data rate data stream is converted into a plurality of low data rate data streams through a serial to parallel converter, and each subcarrier is included in the parallel data streams. After multiplying, the respective data strings are summed and sent to the receiving side. OFDMA is a multiple access method for allocating subcarriers according to a transmission rate required by multiple users in the OFD.

Hereinafter, the conventional Single Carrier-FDMA (SC-FDMA) method will be described. The SC-FDMA method is also called a DFT-S-OFDM method. The conventional SC-FDMA technique is a technique mainly applied to uplink. In this scheme, spreading is first applied to a DFT matrix in a frequency domain before generating an OFDM signal, and then the result is modulated by a conventional OFDM scheme and transmitted. . Several variables are defined to describe the SC-FDMA technique. N denotes the number of subcarriers transmitting the OFDM signal, Nb denotes the number of subcarriers for any user, F denotes a discrete Fourier transform matrix, or DFT matrix, s denotes a data symbol vector, x Denotes a vector in which data is distributed in the frequency domain, and y denotes an OFDM symbol vector transmitted in the time domain.

In SC-FDMA, the data symbols s are distributed using a DFT matrix before transmission. This is expressed by the following formula.

는, 데이터 심볼(s)을 분산시키기 위해서 사용된 N_b 크기의 DFT 행렬이다. 이렇게 분산된 벡터(x)에 대하여 일정한 부 반송파 할당 기법에 의해 부 반송파 매핑(subcarrier mapping)이 수행되고, IDFT 모듈에 의해 시간영역으로 변환되어 수신 측으로 전송하고자 하는 신호가 얻어진다. 상기 수신 측으로 전송되는 전송신호는 아래 식과 같다. In Equation 1

It is a DFT matrix of size N _b used to disperse the data symbol (s). Subcarrier mapping is performed on the distributed vector x by a constant subcarrier allocation scheme, and a signal to be transmitted to the receiver is obtained by converting to the time domain by the IDFT module. The transmission signal transmitted to the receiving side is as follows.

는 주파수 영역의 신호를 시간 영역의 신호로 변 환하기 위해 사용되는 크기 N의 DFT 행렬이다. 상술한 방법에 의해 생성된 신호 y는, 순환 전치(cyclic prefix)가 삽입되어 전송된다. 상술한 방법에 의해 전송 신호를 생성하여 수신 측으로 전송하는 방법을 SC-FDMA 방법이라 한다. DFT 행렬의 크기는 특정한 목적을 위해 다양하게 제어될 수 있다.In Equation 2

Is a DFT matrix of size N that is used to convert a signal in the frequency domain into a signal in the time domain. The signal y generated by the above-described method is transmitted by inserting a cyclic prefix. The method of generating a transmission signal and transmitting it to the receiving side by the above-described method is called an SC-FDMA method. The size of the DFT matrix can be variously controlled for a specific purpose.

The above description is based on the DFT or IDFT operation. However, for convenience of description, the following description uses a Discrete Fourier Transform (DFT) or Fast Fourier Transform (FFT) operation. If the number of input values of the DFT operation is a power of 2, it will be apparent to those skilled in the art that the FFT operation can be performed instead of the DFT operation. Therefore, the following FFT operation is also applicable to the DFT operation.

In an OFDM system, a plurality of OFDM symbols form one frame, which is transmitted in units of frames, and first transmits a preamble every frame or several frame intervals. In this case, the number of OFDM symbols constituting the preamble varies depending on the system. For example, in the OFDMA-based IEEE 802.16 system, a preamble composed of one OFDM symbol is transmitted first in every downlink frame. The preamble is provided to a communication terminal for the purpose of synchronization, cell search, channel estimation, etc. in a communication system.

1 shows a structure of a downlink subframe of the IEEE 802.16 system. As shown, the preamble consisting of one OFDM symbol is transmitted in advance every frame, and the preamble is used for time and frequency synchronization, cell search, and channel estimation.

FIG. 2 is a diagram illustrating a subcarrier set for transmitting a preamble transmitted in a zeroth sector in an IEEE 802.16 system. Some of the given bandwidths are used as guard bands. In addition, when the number of sectors is three, each sector inserts a sequence at intervals of three subcarriers and transmits 0 by inserting 0 into the remaining subcarriers.

Hereinafter, a conventional sequence used in the preamble will be described. The sequence used in the preamble is shown in Table 1 below.

index IDcell Sector Sequence (hex) 0 0 0 A6F294537B285E1844677D133E4D53CCB1F182DE00489E53E6B6E77065C7EE7D0ADBEAF One One 0 668321CBBE7F462E6C2A07E8BBDA2C7F7946D5F69E35AC8ACF7D64AB4A33C467001F3B2 2 2 0 1C75D30B2DF72CEC9117A0BD8EAF8E0502461FC07456AC906ADE03E9B5AB5E1D3F98C6E . . . . . . . . . . . .

The sequence is determined by the sector number and the IDcell parameter value. Each defined sequence is converted to a binary signal in ascending order and mapped to a subcarrier through BPSK modulation. In other words, the proposed hexadecimal sequence is converted into a binary sequence (Wk), and then Wk is mapped from the Most Significant Bit (MSB) to the Least Significant Bit (LSB). (0 maps to +1 and 1 maps to -1. For example, in the 0th segment with index 0, Wk for hexadecimal 'C12' is 110000010010 ..., so the converted binary code value is -1. -1 +1 +1 +1 +1 +1 -1 +1 +1 -1 +1 ...)

The above-described sequence according to the prior art is a sequence capable of maintaining a low peak-to-average power ratio (PAPR) when converting to a time domain while maintaining some correlation characteristics among the types of sequences that can be constituted by binary codes. Was found through computer simulation. If the structure of a system changes or adapts to another system, you have to find a new sequence.

Hereinafter, a sequence used in the latest ^{3GPP (3 rd Generation Partnership Project)} (Long Term Evolution) LTE technology newly proposed.

Various sequences are also used in LTE systems. Hereinafter, a sequence used in the channel of LTE will be described.

In general, the first thing the terminal performs to communicate with the base station is to perform synchronization with the base station in a synchronization channel (hereinafter, referred to as 'SCH') and perform a cell search.

A series of processes for synchronizing with a base station and acquiring a cell ID to which a terminal belongs is called a cell search. In general, the cell search includes an initial cell search performed when the initial UE is powered on, and a neighbor cell where the UE in a connection or idle mode searches for an adjacent base station. It is classified as a neighbor cell search.

The chronization channel (SCH) may have a hierarchical structure. For example, a primary SCH (P-SCH) and a secondary SCH (S-SCH) may be used.

The P-SCH and the S-SCH may be included in a radio frame by various methods. 3 and 4 illustrate various methods of including the P-SCH and the S-SCH in a radio frame. In the LTE system, the SCH may be configured according to the structure of FIG. 3 or 4 according to various situations.

In FIG. 3, the P-SCH is included in the last OFDM symbol of the first sub-frame. In addition, the S-SCH is included in the last OFDM symbol of the second subframe.

In addition, in FIG. 4, the P-SCH is included in the last OFDM symbol of the first subframe, and the S-SCH is included in the second OFDM symbol at the end of the first subframe.

The LTE system may acquire time and frequency synchronization using the P-SCH. In addition, the S-SCH may include a cell group ID, frame synchronization information, antenna configuration information, and the like.

Hereinafter, a method of configuring a P-SCH proposed in the 3GPP LTE system will be described.

P-SCH transmits through the 1.08 MHz band around the carrier frequency (carrier frequency). This corresponds to 72 subcarriers. At this time, the interval of each subcarrier (subcarrier) is 15kHz. This value is determined because 12 subcarriers are defined as one resource block (RB) in the LTE system. In this case, 72 subcarriers correspond to 6 RBs.

P-SCH used in a communication system using a plurality of orthogonal subcarriers such as OFDM or SC-FDMA preferably satisfies the following conditions.

First, auto-correlation and cross-correlation characteristics in the time domain for a sequence forming a P-SCH should be good for the detection of excellent performance on the receiving side.

Second, complexity due to synchronization detection should be low.

Third, it is desirable to have a Nx repetition structure for good frequency offset estimation performance.

Fourth, the Peak-to-Average Power Ratio (PAPR) (or CM) should be low.

Fifth, if the P-SCH can be utilized for channel estimation, it is desirable that the frequency response has a constant value. That is, in terms of channel estimation, a flat response in the frequency domain is known to have the best channel estimation performance.

Although various sequences have been proposed in the related art, there is a problem that the various conditions described above are not satisfied.

The present invention has been proposed to improve the above-described prior art, and the present invention provides a sequence having excellent correlation characteristics.

Summary of the Invention

According to the present invention a method for selecting a plurality of sequences is provided. CLAIMS What is claimed is: 1. A method of selecting sequences for a particular channel in a multi-carrier system, comprising: determining an arbitrary sequence index of a plurality of sequence indexes identifying a sequence of length L as a mother sequence index; Determining at least one remaining sequence index using an integer multiple of half of the value corresponding to one period in the generation formula for the sequence and the parent sequence index; And selecting sequences corresponding to at least two sequence indexes of the parent sequence index and the remaining sequence indexes as sequences for the particular channel.

In addition, in a method for generating sequences for a specific channel in a multi-carrier system, a random sequence index among a plurality of sequence indexes for identifying a sequence of length L is determined as a mother sequence index, and the parent Determining at least one remaining sequence index using the sequence index and the sequence length; Selecting at least two sequence indices of the parent sequence index and the remaining sequence indexes; Generating a sequence corresponding to the selected sequence indices in a frequency domain or a time domain; Performing data processing to remove a direct current (DC) component on the generated sequences; And converting the sequence on which the data processing is performed into a time domain sequence.

In addition, the method for calculating the cross-correlation values of the plurality of sequences and the received signal received from the transmitting side, the cross-correlation value of the received signal and a first sequence of the plurality of sequences, the first even sequence Obtaining a first result that is a real part of the cross-correlation values and a second result that is an imaginary part, and a third result that is a real part and a fourth result that is an imaginary part of the cross correlation values for the odd-numbered first sequence; Obtaining a cross-correlation value between the received signal and the conjugate complex sequence of the first sequence, obtaining a fifth result that is a real part and a sixth result that is an imaginary part of the cross correlation values for the even-numbered conjugate complex sequence, and an odd number Obtaining a seventh result that is a real part and an eighth result that is an imaginary part of a cross correlation value for the first conjugate complex sequence; And adding or subtracting two of the first to eighth results to calculate a real part or an imaginary part of the cross-correlation value for at least one of the plurality of sequences.

Work of the present invention Example

The construction, operation and effects of the invention will be embodied in accordance with one embodiment of the invention described below.

This document largely proposes the first to fourth embodiments.

Hereinafter, the contents of the first to fourth embodiments relate to contents generated, transmitted, and received in the time domain to the frequency domain. For convenience of description, the description may be based on the time domain or the frequency domain, but the present disclosure is not limited thereto.

The first embodiment proposes a method for generating a sequence for a specific channel using a sequence generated in the time domain. The second embodiment proposes a method for generating a sequence for a specific channel using a sequence having excellent correlation characteristics. The third embodiment proposes a method of generating a sequence for a specific channel using a plurality of sequences. In addition, the fourth embodiment proposes a reception method for a sequence generated according to the third embodiment.

First Example

The first embodiment includes a method for generating a sequence in the time domain, a method for converting a time domain sequence to a frequency domain sequence, a method for converting a frequency domain sequence to a time domain, a data processing method for eliminating DC components, and a repetition characteristic in the time domain. We propose a method for generating a sequence having

The sequence generated according to the first embodiment may be used for various channels. For example, it can be used for channels for data and control signals, can be used for channels for retransmission (eg, ACK / NAK signals), and can also be used for channels for frequency and time synchronization. Hereinafter, a method of generating a sequence for channels for frequency and / or time synchronization is described for convenience of description, but the present invention is not limited to this specific example.

For example, in the case of transmitting information on a corresponding channel without performing time synchronization, a specific instantaneous correlation output plays a role of acquiring the corresponding information in the time synchronization concept. Zero-delayed correlation output is performed in the same procedure.

5A is a diagram illustrating a structure of a transmitting and receiving side to which an embodiment of the present invention is applied. The transmitting side is described as follows. When input data is input, channel coding is performed to add additional bits to prevent the data from being distorted in the channel. The channel coding may be a turbo code or an LDPC code. The operation of mapping the bit stream on which the channel coding is performed to a predetermined symbol is performed, and the symbol mapping may be performed by QPSK, 16 QAM, or the like. The data symbol undergoes subchannel modulation mapped to a plurality of subcarriers transmitting the symbol. The signal subjected to the subchannel modulation is carried on the carrier in the time domain through the IFFT, and is transmitted to the wireless channel through filtering and analog conversion. The receiving side reverses the operation of the transmitting side.

5B is a flowchart illustrating a method of designing a time domain of a sequence maintaining a reasonable correlation property and having a low PAPR according to an embodiment of the present invention. A method of generating a preamble according to an embodiment of the present invention with reference to FIG. 5B is as follows. First, a step (S101) of generating and inserting a sequence of length N in the time domain is performed, and a step (S103) of converting to a frequency domain sequence through an N-point FFT operation is performed, and according to the requirements of the communication system. Inserting a direct current (DC) subcarrier and a guard subcarrier (S105), applying a PAPR attenuation technique to the sequence according to the result of the performing (S107), and N-point Inverse Fast Fourier A transform (S109) may be performed through a transform operation. As described above, it is obvious that the DFT or FFT may be selectively used depending on the N value.

Hereinafter, with reference to the accompanying drawings will be described in detail each step of the preferred embodiment according to the present invention.

Hereinafter, a method (S101) of forming and inserting a sequence of length N according to one embodiment of the present invention will be described.

One embodiment of the present invention proposes a method of forming and inserting a sequence showing good correlation characteristics in a time domain and maintaining a constant amplitude value. To this end, the present embodiment generates a sequence of a specific length in the time domain, and inserts the generated sequence in the time domain.

Hereinafter, the preferable conditions required for the sequence used in the present embodiment, that is, the sequence generated and inserted in the time domain will be described. As described above, in order to increase the efficiency of the transmitting amplifier, it is preferable to transmit a sequence for reducing the PAPR, so that the sequence used in the present embodiment has a constant amplitude value in the time domain as described above. It is preferable to have. In addition, it is preferable that the size of the sequence is small in the frequency domain as well as in the time domain.

Most communication methods limit the maximum value of power available in the allocated frequency band while allocating a certain frequency band to a specific transmission / reception side. That is, there is a constant spectrum mask in a general communication method. Therefore, even if the amplitude of the signal is not constant in the frequency domain even if the sequence is transmitted in a constant amplitude in the time domain, when boosting the sequence in the frequency domain may cause problems beyond the spectral mask. have.

In the case of generating and inserting a sequence in the frequency domain, since the size in the frequency domain can be easily controlled, the constraint by the spectral mask will not be large. However, the present embodiment should consider the power value in the frequency domain as the sequence is generated and inserted in the time domain rather than the frequency domain. In addition, if the channel value in the frequency domain is known in advance, it is preferable to perform power allocation according to whether the channel is good or bad, but the power of the used subcarrier is difficult because the channel is not known in advance due to the use purpose of the preamble. It is common to have a constant. Since the FFT module used in the OFDM system uses an N point FFT module, the sequence preferably has a length of N.

In addition, for the above-mentioned frequency-flat characteristics, when using the sequence as a specific channel and using it for channel estimation (for example, in case of P-SCH in LTE), The reference signal has a frequency-flat characteristic is optimal.

In addition to the PAPR characteristic, the sequence used in the present embodiment preferably has excellent correlation characteristics so as to easily detect and distinguish a signal. The excellent correlation property refers to having an excellent autocorrelation property and an excellent crosscorrelation property.

In addition, the sequence is preferably generated to facilitate the synchronization at the receiving end receiving the sequence. The synchronization means frequency synchronization and time synchronization. In general, when a specific pattern is repeated in one OFDM symbol in the time domain, frequency synchronization and time synchronization can be easily obtained.

Therefore, in the sequence according to the present embodiment, it is more preferable that a specific pattern is repeated within one OFDM symbol in the time domain. For example, in the sequence generation and insertion step according to the present embodiment, a preamble sequence having two identical patterns may be inserted in one OFDM symbol generated by the N-point FFT module. There is no limit to the method of creating a sequence of a specific length by repeating the same pattern in the time domain. First, the following example is possible. If an N-point FFT or DFT is a problem, a sequence of length N / 2 is generated and inserted twice by repeating it, thereby forming a preamble sequence of total length N. In addition, if a sequence of length N / 4 is generated and inserted twice by repeating it, a preamble sequence of total length N / 2 can be constructed. The N / 2 preamble sequence thus constructed has a length of N / 2 in the frequency domain. In this case, a sequence of N lengths can be generated later by adjusting the interval of the sequence in the frequency domain.

Meanwhile, the present invention may use a sequence that is not necessarily repeated in the time domain. In this case, if the repetition is not repeated, the above-described repetitive operation can be omitted. In other words, it is also possible to generate only N length sequences in the time domain without repetition. In addition, the sequence used in this step may use a sequence such as a Golay sequence or a binary sequence sequence in addition to the CAZAC sequence mentioned below.

According to one embodiment of the present invention, there are various sequences that can be selected in consideration of the above-described conditions. Hereinafter, as an example of the present embodiment, a method of forming and inserting a sequence of length 1024 in the time domain using a CAZAC sequence will be described. According to the present embodiment, in the following Equation 3, M = 1 (M is a natural number primed with 1024) and a length of 1024 CAZAC (Constant Amplitude Zero Auto Correlation) [David C. Chu, “Polyphase Codes with Good Periodic Correlation Properties” , Information Theory IEEE Transaction on, vol. 18, issue 4, pp. 531-532, July, 1972] Create and insert sequences in the time domain.

In Equation 3, n = 0, 1, 2, ..., N-1.

As described above, the specific pattern used in the present embodiment can be repeated, and the CAZAC sequence can repeat the specific pattern by adjusting the N value. That is, a length 1024 sequence may be generated by generating a CAZAC sequence by repeating M twice with M = 1 and N = 512 in Equation 1 above.

6 is a diagram illustrating autocorrelation characteristics of the CAZAC sequence. As described above, it is preferable that the sequence used in the present embodiment has excellent correlation characteristics. The autocorrelation characteristics in the time domain with respect to the CAZAC sequence show that the ideal autocorrelation characteristics are shown. In conclusion, it can be seen that the CAZAC sequence is one of sequences satisfying various conditions required by the present embodiment.

Hereinafter, a method (S103) of converting a time-domain sequence into a frequency-domain sequence through an FFT operation will be described according to an embodiment of the present invention.

A method for converting a time-domain sequence into a frequency-domain sequence for modification in a frequency domain in accordance with a predetermined standard in an OFDM system, wherein an N-point FFT is performed on a sequence of length N generated in the time domain as shown in Equation 4 below. Change to an area sequence.

In the above equation, k = 0, 1, 2, ..., N-1 is satisfied.

As described above, the time domain sequence generated in the time domain is converted into the frequency domain sequence A _k by Equation 4 above.

Hereinafter, the step (S105) of inserting the DC subcarrier and the guard subcarrier according to an embodiment of the present invention.

In general, certain OFDM communication methods may require the insertion of DC subcarriers and the insertion of certain guard subcarriers. If a DC subcarrier protection subcarrier has to be inserted in order to meet a predetermined standard of a specific OFDM communication method, step S105 is performed. Insertion of the DC subcarrier means inserting data 0 into a subcarrier having a frequency of 0 in the frequency domain in order to solve the problem caused by the DC offset in the RF terminal of the transmission and reception. In addition, the insertion of the guard subcarrier refers to inserting the guard subcarriers to reduce Adjacent Channel Interference (ACI).

In addition, in this step, when mapping to a subcarrier in the frequency domain, it is also possible to change and map the subcarrier position of the corresponding signal. For example, at least one subcarrier may be cyclically shifted to map or randomly mapped. The present invention also includes the present invention, but the position in the frequency domain is preferably not changed. An embodiment of the present invention describes a case where the position in the frequency domain of the generated signal does not change.

Hereinafter, according to an embodiment of the present invention, step S107 of applying the PAPR attenuation technique to the sequence in which the previous steps are performed will be described.

As described above, due to the DC subcarrier and the protection subcarrier insertion, the time domain signal may be modified to increase the PAPR. In this embodiment, the PAPR attenuation technique may be performed again to reduce the increased PAPR. In this case, it is preferable to apply the PAPR attenuation technique while minimizing the change in the magnitude level of the codes of the frequency domain sequence. The frequency domain sequence in which this step is performed is a value already known by the transmitting and receiving side, and can be used later as a reference signal for other uses (for example, channel estimation).

Hereinafter, the step (S109) of converting the sequence into a time domain sequence through the IFFT operation of the present invention will be described.

The above step is a method of generating a final time-domain preamble sequence, which is performed as shown in Equation 5 below, and the generated sequence may be used for performing synchronization and detecting and distinguishing signals.

2nd Example

The following second embodiment relates to a method of applying any one of the CAZAC sequences described in the first embodiment to a P-SCH of an LTE system. More specifically, after repeating a Frank sequence, which is one of the CAZAC sequences in the time domain, a P-SCH is generated through a method of processing data in the frequency domain.

The second embodiment proposes a sequence that satisfies all of the requirements that the P-SCH must satisfy by using a Frank sequence.

The Frank sequence used in the second embodiment is one of the above-described CAZAC sequences, and has a constant envelop in both time and frequency domains, and has an ideal auto-correlation characteristic. [R. L. Frank and S. A. Zadoff, “Phase shift pulse codes with good periodic correlation properties,” IRE Trans. Inform. Theory, vol. IT-8, pp. 381-382, 1962.]

On the other hand, it should be noted that in the case where the P-SCH and the S-SCH are multiplexed in the FDM scheme in LTE, a discussion of configuring a P-SCH using a frank sequence has already been proposed. However, the method proposed by the second embodiment is to propose a P-SCH which shows better performance than the conventionally proposed method when the P-SCH and the S-SCH are multiplexed by the TDM scheme.

Hereinafter, the proposed P-SCH configuration method will be described and compared with the second embodiment.

The Frank sequence may be represented by Equation 6 below.

는 하기 수학식 7과 같다. remind

Is shown in Equation 7 below.

In Equations 6 and 7, N denotes a length N of a Frank sequence and must satisfy N = m ² . Also, r is a natural number smaller than m while being smaller than m.

For example, when N = 4, the sequences according to Equation 6 have a constellation map such as BPSK. In addition, when N = 16, it has the same constellation as QPSK. A Frank sequence is generated in the time domain for N = 16 and r = 1, as shown in Table 2 below, and the sequence converted into the frequency domain is shown in Table 3 below.

In phase Quadrature 0 0 One One -One 0 2 0 -One 3 One 0 4 -One 0 5 One 0 6 -One 0 7 One 0 8 0 -One 9 -One 0 10 0 One 11 One 0 12 One 0 13 One 0 14 One 0 15 One 0

In phase Quadrature 0 One 0 One 0 One 2 -sqrt (1/2) sqrt (1/2) 3 -sqrt (1/2) sqrt (1/2) 4 0 One 5 0 One 6 sqrt (1/2) sqrt (1/2) 7 sqrt (1/2) -sqrt (1/2) 8 -One 0 9 0 One 10 sqrt (1/2) -sqrt (1/2) 11 -sqrt (1/2) sqrt (1/2) 12 0 -One 13 0 One 14 -sqrt (1/2) -sqrt (1/2) 15 sqrt (1/2) -sqrt (1/2)

The results in Table 2 are eventually the same as the QPSK modulation, and the results in Table 3 have a constant amplitude.

For example, when the number of actually used subcarriers (used subcarriers) is 16, using the result of Table 3, regardless of whether or not to use a scalable bandwidth (band scalable bandwidth) Can be.

When an operation such as a timing acquisition is performed in the time domain by a method of cross-correlation, the BPSK or M-PSK scheme (perform phase rotation on the constellation to include the desired information) In this case, the complexity of calculating the correlation value is reduced. That is, the complexity of the calculation is reduced because the correlation value is calculated by simple complex addition using a simple code converter rather than complex operations.

In addition, since the Frank sequence is a CAZAC sequence, it shows excellent correlation characteristics in both the time domain and the frequency domain.

In addition, since the Frank sequence has a constant value in both the time and frequency domains, the PAPR is low and has an optimal condition when used for channel estimation.

For example, a signal vector r = [r (0) r (1)… received in the time domain for N = 16, r = 1. r (15)], and the equation for calculating a correlation value with a known signal (a = [a (0) a (1)… a (15)] ^H ) for timing synchronization acquisition is as follows. . In this case, the signal a is shown in Table 2.

In direct operation according to Equation 8, a total of 15 complex multiplications and 15 complex additions are required to calculate one R (d).

However, the correlation value may be calculated by performing addition by changing only the real or imaginary value of the received signal to be multiplied due to the characteristic of the Frank sequence a. Therefore, the operation can be completed by only 15 complex additions without complex multiplication.

In general, one complex multiplication operation has about eight times the complexity of one complex addition operation.

The already proposed method configures the P-SCH using the advantages of the above-described Frank sequence. In other words, a P-SCH of the FDM scheme that maps to 64 subcarriers using a Frank sequence of length 16 has been proposed.

7 shows a configuration method of the P-SCH already proposed. First, a Frank sequence of length 16 is inserted at two frequency index intervals in the frequency domain. In other words, the sequence of Table 3 is inserted at two frequency index intervals. The meaning of two frequency index intervals means that the m-th sequence is inserted into the k-th subcarrier, the sequence is not inserted into the k + 1-th subcarrier, and the m + 1-th sequence is inserted into the k + 2nd subcarrier. it means.

By copying and extending the sequences inserted at two frequency index intervals in the frequency domain, a sequence as shown in FIG. 7 mapped to a total of 64 subcarriers can be obtained. The sequence of FIG. 7 is composed of a form inserted twice at intervals of two samples in the time domain and repeated twice.

The second embodiment of the present invention improves the configuration method of the above-described P-SCH in the following parts.

First, a sequence according to the configuration method of the proposed P-SCH has a value of 0 in the time domain, and thus has a very bad PAPR characteristic. The second embodiment improves this property.

In addition, the number of subcarriers that can be used in the current LTE is 72, since the proposed method uses only 64, there is a loss in frequency diversity. The second embodiment improves this property.

In addition, in order to solve the problem of the DC carrier (that is, the 0 th carrier), the proposed method inserts a sequence into an odd subcarrier rather than an even subcarrier. In other words, data is inserted into a subcarrier having an odd frequency index. Observing the sequence generated according to this method in the time domain has a fatal problem of deforming to a form other than QPSK in the time domain, which is an advantage of the Frank sequence. That is, a problem arises in that the complexity of complex arithmetic increases. The second embodiment improves this property.

8 is a flowchart illustrating a method of generating a P-SCH according to the second embodiment. 5B and 8, it can be seen that the second embodiment is basically based on the method of the first embodiment.

Steps S1701 to S1705 shown in FIG. 8 are described in detail below.

9 is a diagram illustrating a subcarrier to which a P-SCH of the LTE standard is mapped. The P-SCH of the LTE standard is mapped to 73 subcarriers (including a DC carrier) around a DC carrier.

The second embodiment provides a sequence of a structure that is repeated twice in the time domain to generate a sequence to be mapped to 73 subcarriers (including a DC carrier) required by the LTE standard. That is, a sequence having a 2x repetition structure in the time domain is proposed. In this case, when processing for the DC subcarrier is performed, 71 length frank sequences of 72 length frank sequences are used.

At this time, the sequence repeated twice in the time domain is preferably a Frank sequence. In addition, the length of the Frank sequence is 36, and the variable r in the equation (6) is preferably 1. If the length is 36, the Frank sequence has constellations such as 6-PSK.

The reason why the length of the Frank sequence is selected as 36 in the second embodiment is that a sequence to be mapped to 73 subcarriers must be made. That is, when the sequence is generated by repeating the 36-length sequence twice, it conforms to the LTE standard.

Of course, if you do not want a repeating form, you can choose a length of 64 for the LTE system. In addition, when generating the P-SCH by repeating four times, a Frank sequence of length 16 may be used.

Hereinafter, S1701 of FIG. 8 will be described.

Generate a Frank sequence with N _pre (length of initial sequence for generating P-SCH) = 36. At this time, the variable r in Equation 6 is set to 1.

FIG. 10 is a block diagram showing a Frank sequence of length 36 in the time domain. The sequence of FIG. 10 may be represented by a (i), i = 0, 1, ..., 35. Table 4 below shows the real and imaginary values of a (i) according to the index i. Table 4 below shows some of the values of the sequence in the time domain.

Real Imag 0 One 0 One -cos (pi / 3) -sin (pi / 3) 2 -One 0 3 cos (pi / 3) sin (pi / 3) 4 cos (pi / 3) sin (pi / 3) 5 One 0 6 cos (pi / 3) sin (pi / 3) 7 cos (pi / 3) sin (pi / 3) 8 One 0

Hereinafter, step S1702 will be described.

When using a Frank sequence of length 36, the sequence is repeated twice in the time domain. FIG. 11 is a block diagram illustrating a result of generating a sequence by repeating twice in the time domain according to the second embodiment.

Some of the signals repeated twice as shown in FIG. 11 are shown in Table 5 below.

Real Imag 0 One 0 One -cos (pi / 3) -sin (pi / 3) 2 -One 0 3 cos (pi / 3) sin (pi / 3) 4 cos (pi / 3) sin (pi / 3) 5 One 0 6 cos (pi / 3) sin (pi / 3) 7 cos (pi / 3) sin (pi / 3) 8 One 0

The values of the sequences in Table 5 above are values in the time domain.

Hereinafter, S1703 will be described.

A Frank sequence (that is, a sequence repeated twice in the time domain) of length 72 generated in S1702 is converted into a frequency domain signal through a 72-point FFT or DFT transform. In this case, in the frequency domain, since the repetition is performed twice in the time domain, it appears that alternating insertion is performed from the even frequency index in the frequency domain. That is, a sequence is inserted into the even-numbered frequency index as shown in FIG. 12 is a diagram showing the result of S1703.

A table showing a part of the sequence value inserted into the even frequency index is as follows.

Real Imag 0 Sqrt (2) * 1 0 One 0 0 2 Sqrt (2) * cos (pi / 9) Sqrt (2) * sin (pi / 9) 3 0 0 4 Sqrt (2) * cos (3 * pi / 9) Sqrt (2) * sin (3 * pi / 9) 5 0 0 6 -Sqrt (2) * cos (3 * pi / 9) Sqrt (2) * sin (3 * pi / 9) 7 0 0 8 -Sqrt (2) * cos (pi / 9) -Sqrt (2) * sin (pi / 9) 9 0 0

Hereinafter, step S1704 will be described.

Step S1704 is a step for solving a problem according to the DC subcarrier. If the DC subcarrier part is not used in the communication standard to be used (that is, if 0 is transmitted through the DC subcarrier), it is preferable to perform the step S1704. The second embodiment proposes two methods for solving the DC subcarrier problem. First, step S1704-1 is described as a first method, and then step S1704-2 is described as a second method.

Hereinafter, step S1704-1 will be described.

In step S1704-1, the corresponding sequence located on the DC subcarrier is punctured (in other words, nullified to 0).

13 is a diagram showing the results of step S1704-1. That is, if the step S1704-1 is performed on the result of FIG. 12, the result of FIG. 13 may be obtained.

Some of the results of FIG. 13 may be displayed as shown in Table 7 below.

Real Imag 0 0 0 One 0 0 2 Sqrt (2) * cos (pi / 9) Sqrt (2) * sin (pi / 9) 3 0 0 4 Sqrt (2) * cos (3 * pi / 9) Sqrt (2) * sin (3 * pi / 9) 5 0 0 6 -Sqrt (2) * cos (3 * pi / 9) Sqrt (2) * sin (3 * pi / 9) 7 0 0 8 -Sqrt (2) * cos (pi / 9) -Sqrt (2) * sin (pi / 9)

Hereinafter, step S1704-2 will be described.

S1704-2 of the second embodiment is a step of mapping the sequence avoiding the DC subcarrier.

In step S1702, the sequence was repeated two times in the time domain. Accordingly, the result of S1703 also has a form of being inserted at two frequency index intervals in the frequency domain. In other words, the sequence is inserted into the even-numbered frequency index.

In this case, the second embodiment performs step S1704-2 to perform a circular shift of the generated sequence to the right or left. FIG. 14 illustrates a result of performing a cyclic shift to the right of the result of FIG. 12.

Some of the results in FIG. 14 can be represented by Table 8 below.

Real Imag 0 0 0 One Sqrt (2) * 1 0 2 0 0 3 Sqrt (2) * cos (pi / 9) Sqrt (2) * sin (pi / 9) 4 0 0 5 Sqrt (2) * cos (3 * pi / 9) Sqrt (2) * sin (3 * pi / 9) 6 0 0 7 -Sqrt (2) * cos (3 * pi / 9) Sqrt (2) * sin (3 * pi / 9) 8 0 0

Comparing the steps S1704-1 or S1704-2, it can be seen that the method of S1704-1 is more preferable. S1704-1 can calculate a correlation value through a simple operation using the signals of Table 5 already known. The specific method of calculating a correlation value is demonstrated below. However, in S1704-2, since the value of the time-domain sequence is changed by inserting the sequence at the odd-numbered index, it is difficult to obtain the correlation value through a simple operation. Of course, the reception side may be solved by shifting the carrier frequency by a subcarrier spacing between subcarriers, but this may cause a DC offset because the first subcarrier becomes a DC component. Therefore, the S1704-1 method is preferred. Of course, the frequency shift may be performed by multiplying a specific complex number in the time domain after reception. However, using the method of multiplying this particular complex number for simple correlation value calculation is a method of too low efficiency.

Hereinafter, step S1705 will be described. Step S1705 is an additional step for applying a 128 point FFT without performing down sampling on the receiving side.

Step S1705 may be useful when the receiving side does not support down sampling. As an example of an LTE system, a subcarrier spacing is 15 kHz. If a 128 point FFT (or DFT) is applied, the value of 128 samples is obtained in the time domain, which has a sampling frequency of 1.92 MHz. The receiving side may perform any one of the following operations after filtering the received signal to 1.08 MHz. The first is to use a sampling frequency of 1.92 MHz (the frequency corresponding to 128 samples) as it is, and the second is to use the down sampling to a sampling frequency of 1.08 MHz (the frequency corresponding to 72 samples). . The step S1705 is an additional step for using the sampling frequency of 1.92 MHz as it is without performing down sampling on the receiving side.

When upsample is to be performed, step S1705 upsamples a sequence generated at 1.08MHz (frequency corresponding to 72 samples) to 1.92MHz. The digital upsampling method is based on zero-padding on 56 (= 128-72) subcarriers and performing a 128 point IFFT on the result. The detailed sampling technique is a method that can be understood by anyone skilled in the art to which the present invention pertains, and thus, further description thereof will be omitted. For reference, the sequence of Table 7 or Table 8 may be used within the corresponding band, that is, the 1.08 MHz band during transmission.

Hereinafter, the operation at the receiving side that receives the P-SCH sequence proposed in the second embodiment will be described. That is, a cross-correlation method at the receiving side will be described.

Since the above-described example is repeated twice in the time domain, after determining a predetermined range of a received signal using auto-correlation, fine using cross-correlation for the determined range. The synchronization acquisition process can be performed. As a method of determining a predetermined range of a repeating received signal through an auto-correlation method, an existing method may be applied as it is, and thus, a method of reducing a computation amount through a cross-correlation method is described below. Explain only about.

A timing synchronization acquisition method using the cross correlation method is shown in Equation 9 below.

각각은, 시간 영역에서 알고 있는 P-SCH 시퀀스 값, 수신 신호, 부분적인 상관(partial correlation) 방법을 위한 수행하기 위한 M 값, FFT 크기, 검출된 시간 동기 위치를 나타낸다. In Equation 9

Each represents a P-SCH sequence value known in the time domain, a received signal, an M value to perform for a partial correlation method, an FFT size, and a detected time synchronization position.

When there is no repetition form in the P-SCH, if the frequency offset in the 2GHz band is 5ppm maximum, M = 2 shows sufficient performance. Therefore, in the above example, since the structure is repeated twice, Equation 9 shows sufficient performance when M = 1. Therefore, there is no need to apply a partial correlation method within the repeated section.

In the LTE system, based on Equation 9, when down-sampled a received signal at a 1.08 MHz sampling frequency (72 samples), and the P-SCH is 2 symbols within 10 ms, and averages for 5 ms interval to obtain time synchronization , The complexity of the calculation for time synchronization acquisition is as shown in Equation 10 below.

(72 complex multiplications + 72 complex additions + 2 complex powers) * 9600

In order to explain the correlation value calculation method according to the present invention, the Frank sequence of Table 4 will be described as an example.

If the received signal is r = (r (0) r (1) r (2),... , r (35)], the correlation between the received signal and the correlation value calculation in Table 4 is processed in parallel as follows.

First, the real value is processed as in Equation 11a.

Real value:

Real [r (0)]-Real [r (2)] + Real [r (5)] + Real [r (8)] + Real [r (11)] + Real [r (13)]-Real [ r (14)] + Real [r (15)]-Real [r (16)] + Real [r (17)]-Real [r (18)] + Real [r (20)] + Real [r ( 23)]-Real [r (26)] + Real [r (29)] + Real [r (31)] + Real [r (32)] + Real [r (33)] + Real [r (34) ] + Real [r (35)] + cos (pi / 3) * {-Real [r (1)]-Real [r (3)] + Real [r (4)] + Real [r (6)] -Real [r (7)]-Real [r (9)]-Real [r (10)]-Real [r (12)]-Real [r (19)]-Real [r (21)]-Real [r (22)]-Real [r (24)]-Real [r (25)]-Real [r (27)] + Real [r (28)] + Real [r (30)]} + sin ( pi / 3) * {-Imag [r (1)] + Imag [r (3)] + Imag [r (4)]-Imag [r (6)] + Imag [r (7)]-Imag [r (9)] + Imag [r (10)]-Imag [r (12)]-Imag [r (19)] + Imag [r (21)]-Imag [r (22)] + Imag [r (24 )] + Imag [r (25)]-Imag [r (27)]-Imag [r (28)] + Imag [r (30)]}

Imag value:

Imag [r (0)]-Imag [r (2)] + Imag [r (5)] + Imag [r (8)] + Imag [r (11)] + Imag [r (13)]-Imag [ r (14)] + Imag [r (15)]-Imag [r (16)] + Imag [r (17)]-Imag [r (18)] + Imag [r (20)] + Imag [r ( 23)]-Imag [r (26)] + Imag [r (29)] + Imag [r (31)] + Imag [r (32)] + Imag [r (33)] + Imag [r (34) ] + Imag [r (35)] + cos (pi / 3) * {-Imag [r (1)]-Imag [r (3)] + Imag [r (4)] + Imag [r (6)] -Imag [r (7)]-Imag [r (9)]-Imag [r (10)]-Imag [r (12)]-Imag [r (19)]-Imag [r (21)]-Imag [r (22)]-Imag [r (24)]-Imag [r (25)]-Imag [r (27)] + Imag [r (28)] + Imag [r (30)]}-sin ( pi / 3) * {-Real [r (1)] + Real [r (3)] + Real [r (4)]-Real [r (6)] + Real [r (7)]-Real [r (9)] + Real [r (10)]-Real [r (12)]-Real [r (19)] + Real [r (21)]-Real [r (22)] + Real [r (24) )] + Real [r (25)]-Real [r (27)]-Real [r (28)] + Real [r (30)]}

If the complexity according to equations (11a) and (11b) is indicated, it is as follows. Comparing Equation 12 and Equation 10 shows that there is a big difference in complexity.

((52 * 2) real addition + (2 * 2) real multiplication) * 9600

= (104 real addition + 4 real multiplication) * 9600

Also, since cos (pi / 3) = 1/2, this corresponds to a 1-bit shift in hardware implementation and can be ignored in terms of computation. In this case, the amount of calculation is as follows.

((51 * 2) real addition + (1 * 2) real multiplication) * 9600

= (102 real addition + 2 real multiplication) * 9600

Since sin (pi / 3) = sqrt (3) / 2 = 0.8660, this is approximated to 0.75 (= 1/2 + 1/4). In this case, because it can be implemented as a bit shift (bit shift), if the amount of calculation is ignored, the complexity is reduced as in the following equation.

((51 * 2) real addition + (1 * 2) real addition) * 9600

= (102 real addition) * 9600

On the other hand, since the '+' or '-' sign is simply implemented through a sign inverter, it is not included in the calculation amount.

In the above example, the P-SCH is configured twice in the time domain. However, since these specific numerical values are only examples for describing an example of the present invention, the present invention is not limited to these specific numerical values.

For example, the initial sequence may use a Frank sequence of length 16. That is, as a step S1701, a Frank sequence of length 16 is generated. In addition, as a step S1702, a Frank sequence of length 16 is repeated four times in the time domain. Also, as step S1703. Convert to frequency domain through 64 FFT. In this case, the sequence is inserted every four frequency indexes in the frequency domain. In step S1704, puncturing is performed at the DC carrier position, or a sequence is inserted avoiding the DC carrier. Thereafter, the signal may be converted into a time domain signal, and operation S1705 may be performed if necessary.

Third embodiment

The third embodiment below describes a method for generating a sequence used in various channels. More specifically, a method of generating a sequence used in a communication system for receiving a sequence using a correlation technique at a receiving end will be described.

Hereinafter, for convenience of description, a description will be made based on a sequence for frequency and time synchronization, but the sequence proposed in the third embodiment can be used in various communication fields such as signaling, control channel, and ACK / NACK.

In general, a correlation metric includes a delay component in an operation procedure for obtaining time synchronization. That is, it is expressed as (R (d)). However, unless the time synchronization is obtained, the correlation metric according to the delay is not necessary. In other words, if the present invention is applied to a channel related to time synchronization, the delay component (d) should be considered. If applied to a channel not related to time synchronization, the delay component need not be considered.

Hereinafter, various equations are proposed in consideration of the delay component (d). However, it is obvious that the proposed formula can be equally applied even when there is no delay component (d = 0). Therefore, the case where there is no delay component is not described separately.

Hereinafter, a method of generating a plurality of sequences for frequency and time synchronization will be described. That is, a description will be given based on the case where a plurality of sequences are used instead of a common sequence in one cell. A sequence used for frequency and time synchronization in a cell may be referred to as a primary sequence code (PSC). For example, when a P-SCH is designed using one common sequence in one cell, a cell common PSC is applied, and a P-SCH is used by using multiple sequences in one cell. In the case of designing a plurality of PSC is applied.

The third embodiment proposes a method of generating a plurality of sequences so that the receiving end can obtain correlation values for the plurality of sequences through one correlation operation.

When the P-SCH is designed using the Frank sequence described in Equation 6, a sequence of length 16 and a sequence of length 36 may be used. In this case, when the length N is 16, two kinds of Frank sequences are used because the variable m according to Equation 6 is '4'. In addition, when the length N is 36, two kinds of sequences are used because the variable m according to Equation 6 is '6'. In this case, a problem of not supporting three or more PSCs may occur.

Hereinafter, a method for generating a sequence for a synchronization channel that can be used in various communication systems is proposed, but a method for supporting several types of synchronization channels in one cell is proposed.

There is no limitation on the type of the various communication systems, and the following description will be made based on the LTE system for convenience of description.

The third embodiment describes a Zadoff-Chu sequence through Equation 15 to propose a method for generating a plurality of PSCs. The Zadoff chew sequence has already been described in Equation 3 above.

M is a natural number equal to or less than L and is a natural number displaced from L. For example, if L = 8, m is set to 1,3,5,7.

The third embodiment proposes a method for generating a plurality of sequences using the Zadoff-Chu sequences. It is preferable that the synchronization channel generated according to the sequence according to the third embodiment has a structure as shown in FIG. 9.

The sequence according to the third embodiment is generated according to the procedure of FIG. First, a plurality of sequence indexes are efficiently selected to generate a plurality of sequences (S10). When a plurality of sequence indexes are selected, a sequence is generated in the time domain according to the selected index (S20). In this case, it is more preferable to repeat N times in the time domain (S30). The generated time domain sequence is converted into a frequency domain sequence (S40). In the frequency domain, data processing for removing the DC component is performed (S51 or S52). When data processing for removing the DC component is performed, data processing for converting to a time domain sequence is performed (S60).

First, step S10 of efficiently selecting a plurality of sequence indices will be described. In step S10, the transmitting end determines one mother sequence index and determines the remaining sequence indexes. More specifically, when timing synchronization acquisition is performed at the receiving end, a parent sequence index and remaining sequence indexes for obtaining a correlation correlation value are determined through a small number of operations.

The number of PSCs used in a cell may vary. Hereinafter, an example of configuring a P-SCH using four PSCs will be described. If only three PSCs are needed, four PSCs may be selected and three PSCs may be used.

Hereinafter, a method of generating a sequence using a Zadoff-Chew sequence having a length of 36 or 32 will be described. In this case, a method of generating a P-SCH by repeating twice in the time domain will be described.

By using Equation 15, a Zadov chew sequence of length 36 or length 32 may be generated.

According to Equation 15, when the length L is 36, m representing the sequence index is 1, 5, 7, 11, 13, 17, 19, 23, 25, 29, 31, 35. In the case where the length L is 32, m representing the sequence index is 1, 3, 5, 7, 9, 11, 13, 15, 17, 19, 21, 23, 25, 27, 29, 31. .

When the length is 36, one of 1, 5, 7, 11, 13, 17, 19, 23, 25, 29, 31, and 35 is determined as the parent sequence index. In addition, when the length is 32, one of 1, 3, 5, 7, 9, 11, 13, 15, 17, 19, 21, 23, 25, 27, 29, 31 is determined as the parent sequence index index. Hereinafter, the parent sequence index is referred to as m _o , and the rest of the sequence index is referred to as m _i .

It is preferable that the relationship of the following expression (16a) is established between the parent sequence index mo and the remaining sequence index mi.

or

P _L refers to a value corresponding to one period corresponding to 2 * pi in a polyphase sequence. Typically, the value in the denominator of the phase component in a generation equation that produces a sequence is the value corresponding to this one period. In other words, in the case of polyphase sequence, it is related to the integer multiple of half of the period generated by the generation formula.

이다. Equations 16b and 16d are examples of applying Equation 16a to a specific sequence. In Equation 15, in the case of a Zadoff-Chu sequence, a value corresponding to one period is equal to the length L of the sequence. therefore,. In Equation 16b, the generation period is L. Applying the same method to the Frank sequence, the equation (16d) can be obtained. That is, in Equation 16d, a value corresponding to one period is

to be.

or

As shown in Equation 16b, when the mother sequence index m _o and the remaining sequence index m _i are determined, the operation at the receiving end is simplified.

For example, when a sequence is generated by selecting one m _o and three m ₁ , m ₂ , and m ₃ , the receiving end must obtain a cross-correlation value using four sequences. That is, after receiving an unknown signal, the receiving end obtains a cross correlation value between the unknown signal and the m _o , m ₁ , m ₂ , and m ₃ sequences stored in the receiving end, and according to the magnitude of the cross correlation value. It is necessary to distinguish whether the unknown signal is an m _o sequence, an m ₁ sequence, an m ₂ sequence, or an m ₃ sequence.

However, when an unknown signal according to the third embodiment is received, the magnitude of the cross correlation for the remaining sequences is determined by obtaining the magnitude of the cross correlation for any one of the m _o to m ₃ sequences. Specific operations at the receiving end will be described with reference to the fourth embodiment.

For example, when the sequence length L is 32, the parent sequence index m ₀ may be determined as '1'. In this case, by substituting 1 into m _o of the first equation of Equation 16b and substituting 32 into L, '15' may be obtained as m ₁ . In addition, by substituting 1 for mo of the second equation of Equation 16b and substituting 32 for L, '17' may be obtained as the m ₂ value. Also, when m _o and L are substituted in the first equation of Equation 16b, '31' may be obtained as the m ₃ value. In this case, m _o , m ₁ , m ₂ , and m ₃ may be determined as one index group.

In summary, when one parent sequence index is determined, an index group may be determined accordingly.

When the length is 32, m ₀ = 3, m ₁ = 13, m ₂ = 19, and m ₃ = 29 may be determined as one index group. Of course, other combinations are possible, and when using eight sequences, two groups can be selected in the same manner.

When the sequence length L = 36, m ₀ = 1, m ₁ = 17, m ₂ = 19, and m ₃ = 35 may be determined as one index group using Equation 16b. In addition, m ₀ = 5, m ₁ = 13, m ₂ = 23, and m ₃ = 31 may be determined as one index group.

In the case of a prime number with L = 37, m ₀ = 1 and m ₁ = 36 can be determined as one group, or m ₀ = 2 and m ₁ = 16 can be determined as one group.

When the length L is odd, Equation 16b is simply summarized as in Equation 16c below.

When sequences corresponding to the sequence indices selected by Equation 16c are used, all correlation operations are completed in one correlation operation as in the case of Equation 16b.

Equation 16c is only a subset of Equations 16a to 16b. That is, Equation 16c is a special form of Equations 16a to 16b.

The sequences selected according to the third embodiment may be a polyphase sequence composed of all CAZAC sequences or exponential functions in addition to the Zadoff-Chu sequences. For example, a Frank sequence or the like is possible. However, in the case of the Frank sequence, the equations (16b) and (16c) are modified as in the following equations (16d) and (16e).

Equations 16d to 16e are also only a subset of Equations 16a. That is, it is a special form of Equation 16a.

or

The sequences selected according to the third embodiment may be truncated Zadoff chew sequences. In the case of generating a Zaddo-Chu sequence, a larger number of sequences is obtained by setting the length of the sequence to a prime number. In this case, some of the bits can be cut out to form a truncated Zadoff chew sequence. For example, after generating a sequence of length 37, length 1 may be abandoned to generate a sequence of length 36.

According to Equation 16c, two sequence index groups processed at one time may be generated. For example, in the case of a Zadoff chew sequence of length 37, (1-36), (2-35), (3-34), (4-33), (5-32), (6-31), (7-30), (8-29), (9-28), (10-27), (11-26), (12-25), (13-24), (14-23), (15 -22), (16-21), (17-20), (18-19) can be set to the index group.

Since Equation 16c is only a special form of Equation 16b, sequence indices following Equation 16c have properties corresponding to sequence indices according to Equation 16b.

As described above, all sequence indices may be selected according to Equation 16a, or may be selected using another method. For example, a method of selecting a new sequence by selecting some sequence indexes and cyclically shifting any one of the selected sequence indexes by an arbitrary size according to Equation 16a is possible.

For example, select the sequence indexes '1' and '31' of length 32. In this case, a new sequence may be selected by performing a transition of the sequence corresponding to the sequence index '1' or '31' by half the length of the sequence. That is, a new third sequence may be selected by cyclically shifting the 32-length sequence corresponding to the sequence index '1' or '31' by 16.

The specific numerical values described above are merely examples of the present embodiment, and therefore, the content of the present invention is not limited to the specific numerical values described above.

Hereinafter, for convenience of explanation, the case where the sequence length (L) is 32 or 36 will be described, and when the length is 32, m ₀ = 1, m ₁ = 15, m ₂ = 17, and m ₃ = 31 A case where one index group is defined will be described. A case where the length is 36 will be described based on a case where m ₀ = 1, m ₁ = 17, m ₂ = 19, and m ₃ = 35 as one index group. .

Hereinafter, step S20 of generating a sequence in the time domain according to the selected sequence index will be described.

When using the equation (15), and a length of 36, and to generate a sequence for a _{m0 = 1, m 1 = 17} , m 2 = 97, m 3 = 35 a single index group. Table 9 shows the generated sequence.

The results in Table 9 relate to four sequences, one of which is shown in FIG. However, FIG. 10 relates to Frank sequence and the results of Table 9 relate to Zadoff-Chu sequence.

In addition, using Equation 15, a sequence result of one index group having a length of 32 and m0 = 1, m1 = 15, m2 = 17, and m3 = 31 can be obtained. Table 10 shows the generated sequences.

Hereinafter, the step of repeating N times in the time domain, that is, the step S30 will be described.

The step S30 may be omitted, and N may be freely determined.

Hereinafter, an example of repeating the results of Table 9 twice in the time domain will be described with reference to Tables 11 and 12. Table 11 and Table 12 are the results of repeating Table 9.

Hereinafter, an example of repeating the results of Table 10 twice in the time domain will be described with reference to Tables 13A and 13B. As indicated, the results in Table 10 are repeated once more.

Hereinafter, the step of converting the generated time domain sequence into the frequency domain sequence, that is, step S40 will be described.

When a sequence having a structure that is repeated twice in the time domain is converted into a frequency domain sequence, a sequence having a frequency component only in even-numbered frequency indexes in the frequency domain is generated according to the characteristics of the DFT operation.

Specifically, by converting the sequences of Tables 11 and 12 into sequences in the frequency domain, the sequences of Tables 14A and 14B can be obtained.

In addition, the sequence of Tables 15a and 15b can be obtained by converting the sequences of Tables 13a and 13b into sequences in the frequency domain.

 Hereinafter, a step (S51 or S52) of performing data processing for removing the DC component in the frequency domain will be described.

Since the step S51 is to perform DC puncturing, only the DC component is processed as '0' in Table 14a. That is, the results of Table 14a and Table 14b are shown in Table 16, and the results of Table 15a and Table 15b are shown in Table 17 below.

Table 16 and Table 17 show only the DC component for convenience, and the rest of the components are the same and are omitted.

m ₀ = 1 Real Imag m ₁ = 17 Real Imag m ₂ = 19 Real Imag m ₃ = 35 Real Imag 0 0 0 0 0 0 0 0 0 0 0 0

m ₀ = 1 Real Imag m ₁ = 15 Real Imag m ₂ = 17 Real Imag m ₃ = 31 Real Imag 0 0 0 0 0 0 0 0 0 0 0 0

Step S51 may be described based on the frequency domain as described above, or may be described based on the time domain as follows.

For example, a sequence of length 35 generated according to the present embodiment may be denoted as c (n). c (n) corresponds to a time domain sequence. The result of performing DC puncturing on the c (n) sequence may be referred to as d (n).

로 표시되고, d(n)은

와 같이 표시될 수 있다. In this case, c (n) is

Where d (n) is

It may be displayed as follows.

When the step S52 has a repetitive structure in the time domain, since frequency components are alternately generated in the frequency domain in the frequency domain, the corresponding sequence is shifted (or cyclically shifted) to prevent the presence of frequency components in the DC component during subcarrier mapping. Is a data processing method to remove the DC component.

Since the indices of the results of Tables 14a to 15b are adjusted, detailed results are omitted.

When data processing for removing the DC component is performed, data processing (S60) for converting to a time domain sequence is performed. Processing the results of Table 16 according to S60 yields the results of Tables 18a and 18b, and processing the results of Table 17 yields the results of Tables 19a and 19b.

16 is a view comparing constellations of a sequence in which a DC component is removed and a sequence in which a DC component is not removed according to the present embodiment. More specifically, when the parent sequence index (m _o ) is '1', the result of repeating the sequence of length 36 twice is the result of (a) of FIG. 16, and the result of repeating the sequence of length 32 twice This is the result of FIG.

In both cases, it can be seen that there are only 12 constellations, and even though DC puncturing is performed, the positions of the constellations are shifted by the punctured values to maintain 12 fixed constellations. Such a small number of features can reduce the amount of computation associated with correlation at the receiving end.

FIG. 17 is a diagram illustrating a method of designing a sequence in the frequency domain to have a form repeated twice in the time domain.

Zadoff-Chu sequences maintain ideal correlation in both time and frequency domains. Accordingly, the sequence may be generated in the time domain as shown in FIG. 15, and the sequence may be generated in the frequency domain as shown in FIG. 17. That is, the sequence may be generated through the step S70 of FIG. 17 instead of the step S20 to S40 of FIG. 15.

In other words, insert Zadov-Chu sequences right in the frequency domain. If the sequence is inserted at even-numbered frequency indexes by two spaces, the same result as in FIG. 15 can be obtained.

Hereinafter, an additional description will be given with respect to step S10 of FIG. 15. The above-described method of selecting a plurality of sequence indices is a method for allowing the receiving end to simply obtain a cross correlation value. However, since the Jard Chu sequence is basically a polyphase sequence, it is sensitive to a frequency offset.

Hereinafter, the frequency offset characteristic of the Zadoff chew sequence will be described.

20A, 20B, and 21 are diagrams showing correlation values when a normalized frequency offset of a Zadoff-Chu sequence having a length of 36 is 0.5. As shown, if there is an influence of the frequency offset, it is robust to the frequency offset when it is close to the low value of M or L. In consideration of this, four sequences can be calculated at one time. However, if only three are needed, the frequency offset is selected. For example, if the sequence length is 36, m0 = 1, m1 = 19, m2 = 35 or m0 = 1, m1 = 17, m2 = 35 or two calculations at the same time, giving up the gain of calculating four simultaneously. In consideration of the above, it is also possible to select m0 = 1, m1 = 35 and m2 = 5.

In summary, when only three sequences are selected according to Equation 16b without considering the frequency offset, it may be difficult to find an accurate correlation value according to the frequency offset. As a result of the experiment, when the difference between the sequence indices is 1/2 of the sequence length, an error may occur in the correlation value. For example, if m0 = 1 and m2 = 35, the difference does not occur because the difference between the two sequence indices is not L / 2 (= 18), but if m0 = 1 and m1 = 19, Since the difference is (L / 2), problems can arise. In this case, two sequence indexes among the three sequence indexes may be determined according to Equation 16b, and the other sequence index may determine a sequence offset having a characteristic more robust to the frequency offset in consideration of the frequency offset characteristic.

In conclusion, in selecting a plurality of sequence indices, it is possible to consider only Equation 16b, and to consider the characteristics of the frequency offset together with Equation 16b.

The above description relates to a method of selecting a plurality of sequence indices in consideration of a frequency offset. Hereinafter, a method of additionally considering a criterion other than the frequency offset will be described.

Hereinafter, a method of considering the sequence index m by further considering the characteristics of the PAPR will be described.

FIG. 22 shows cubic metric characteristics of the sequence index M. FIG. 22 reflects the result of DC puncturing. The PAPR characteristic can be seen by looking at the cubic metric characteristic that represents the third power of the Zadoff-Chu sequence. For example, repeating a sequence of length 35 twice in the time domain. Here, preferred sequence indices are 1,2,4,31,33,34. (CM 2.5dB or less)

Hereinafter, a method of considering a sequence index by considering correlation characteristics will be described.

For example, since the Zadoff-Chu sequence is a CAZAC sequence, it is preferable to select a sequence having ideal autocorrelation characteristics and excellent cross-correlation characteristics. For example, in the case of length 35, a combination of three sequences (1,2,34) or (1,33,34) in consideration of both the above-described equation (16c) and the characteristics of the frequency offset and the above-described PAPR is selected. Can be.

The characteristics of the cross correlation for the index combination of (1, 2, 34) proposed in this embodiment are as shown in FIG.

Hereinafter, the characteristics of the sequence of length 35 proposed in this embodiment will be described.

More preferably, a sequence of length 35 is used in the LTE system.

 The SCH signal is transmitted in six radio blocks (corresponding to 73 subcarriers including the DC component).

If all 73 subcarriers are used and a structure is repeated twice in the time domain, a sequence of length 36 may be used. All cases generated in the frequency domain or time domain are possible. The same is true if it is not repeated in the time domain or if it is repeated three or more times.

In this case, the receiving end of the 1.08 x MHz interpolator is required.

However, as mentioned above, when the above criteria are applied, the optimal index group is (1, 2, 35), and the cross correlation at this time is shown in FIG.

The index group of FIG. 24 has 40% cross correlation in the worst case.

In this case, it is preferable to use a sequence of length less than 36.

In this case, it is preferable to select an odd length sequence while approaching the original length to be generated, so that the length is preferably 35.

The length of 35 finds a combination with better correlation properties compared to the case of even lengths.

Fourth embodiment

The fourth embodiment will be described below at the receiving end.

There are certain rules among the sequences selected according to step S10 of FIG. 15 or FIG. 17, so that instead of calculating the cross correlation values for all sequences, the correlation values for the sequences corresponding to one mother sequence index are used. It is possible to derive a correlation value for the sequences corresponding to the sequence indices.

Method 1 of the fourth embodiment

First, Method 1 of the fourth embodiment describes a method of obtaining a cross-correlation value for sequences having a length of 36, m _o = 1, m ₁ = 17, m ₂ = 19, and m ₃ = 35.

The receiving end stores a sequence having a sequence index of '1' and calculates a correlation between the stored sequence and the received sequence. In this case, using the intermediate results generated in the process of obtaining the cross correlation between the received signal and the sequence having the sequence index of '1', while obtaining the cross correlation value between the received signal and the sequence having the sequence index of '17', The cross-correlation value between the received signal and the sequence having the sequence index '19' may be obtained while obtaining the cross correlation value between the received signal and the sequence having the sequence index '19'.

번째 지연(delay)의 상호 상관 값을 계산하는 기준으로 설명한다. 다시 말하면, 수신 신호를

이라고 할때, d번째 지연된 샘플인

와의 상관 값을 기준으로 설명한다. 여기서, 시퀀스 인덱스 m에 대한 상관 값의 결과는 수학식 17a와 같다. Hereinafter, in the fourth embodiment

A description will be given as a reference for calculating the cross correlation value of the first delay. In other words, the received signal

, The d-th delayed sample

It demonstrates based on the correlation value with. Here, the result of the correlation value for the sequence index m is expressed by Equation 17a.

With respect to the first embodiment for the description of the present invention, since m ₀ = 1, m ₁ = 17, m ₂ = 19, m ₃ = 35, the following relationship holds.

는 k가 짝수일 때

의 켤 레 공액(conjugate)과 같고, k가 홀수일 때

의 실수 부와 허수 부를 바꾸고 -1을 곱한 것과 같다. In addition,

Is when k is even

Equal to the conjugate of, where k is odd

Equivalent to multiplying real and imaginary parts of and multiplying by -1.

는 k가 짝수일 때

와 같고, k가 홀수일 때 실수부와 허수부를 바꾼 것의 공액과 같다. Also,

Is when k is even

Equal to the conjugation of changing real and imaginary parts when k is odd.

는 k가 짝수일 때

의 실수 부에만 '-1'을 곱한것과 같고, k가 홀수일 때

의 공액과 같다. Also,

Is when k is even

Is equal to multiplying '-1' only by the real part of, and k is odd.

Is equal to the conjugate of.

For the received signal r (k + d), an instantaneous correlation value of each sequence may be calculated as follows for r_i (k + d) + jr_q (k + d). r_i () means the real part of the received signal, r_q () means the imaginary part of the received signal.

First of all, for the convenience of description, the cross correlation value (the cross correlation value between the received signal and the sequence known to the receiver) for the received signal is defined as follows.

)을 실수 부분과 허수 부분으로 구분하면, 하기 수학식 17c와 같다. First, cross-correlation values for the sequence known to the receiver and the even sequence of the received signal (

) Is divided into a real part and an imaginary part, as shown in Equation 17c.

The result of Equation 17c may be divided into a real part (hereinafter referred to as 'Reven0') and an imaginary part (hereinafter referred to as 'Ieven0').

)을 실수 부분과 허수 부분으로 구분하면, 하기 수학식 17d와 같다. In addition, cross-correlation values for odd-numbered sequences of received signals (

) Is divided into a real part and an imaginary part, as shown in Equation 17d below.

The result of Equation 17d may be divided into a real part (hereinafter referred to as 'Rodd0') and an imaginary part (hereinafter referred to as 'Iodd0').

)을 실수 부분과 허수 부분으로 구분하면, 하기 수학식 17e와 같다. In addition, the cross-correlation value for the even-numbered sequence for the conjugate of the received signal (

) Is divided into a real part and an imaginary part, as shown in Equation 17e.

The result of Equation 17e may be divided into a real part (hereinafter referred to as 'Reven1') and an imaginary part (hereinafter referred to as 'Ieven1').

)을 실수 부분과 허수 부분으로 구분하면, 하기 수학식 17f와 같다. In addition, cross-correlation values for odd-numbered sequences of conjugates of the received signal (

) Is divided into a real part and an imaginary part, as shown in Equation 17f.

The result of Equation 17f may be divided into a real part (hereinafter referred to as 'Rodd1') and an imaginary part (hereinafter referred to as 'Iodd1').

In this case, the calculation of Reven0, Ieven0, Rodd0, Iodd0, Reven1, Ieven1, Rodd1, and Iodd1 described above is performed by Reven_i_i, Revenq_q, Ieven_i_q, Ieven_q_i, Rodd_i_i, Rodd_q_i_i_i_i_i_i_i_i_i_i_i_i_i_i_i_i_i_i_i_od_od_o_i_i_i_i_i_od_od_o_i_i_i_i_i_i_i_i_i_i_i_i_i_i_i_i_i_i_i_o_i_i_i_i_i_i_i_i_i_i_i_i_i_i_i_i_ during until during until It is the same as asking.

Hereinafter, a method of obtaining Reven_i_i, Revenq_q, Ieven_i_q, Ieven_q_i, Rodd_i_i, Rodd_q_q, Iodd_i_q, and Iodd_q_i described above will be described with Equation 17g.

The process of Equation 17g can be approximated and calculated. That is, the calculation of Equation 17g can be easily performed using quantization.

For example, 0.93969-> 1, 0.17365-> 0.125 (= 1/8), 0.76604-> 0.75 (= 1/2 + 1/4), 0.34202-> 0.375 (= 1/4 + 1/8), 0.98481-> 1, 0.64279-> 0.625 (= 1/2 + 1/8), 0.99619-> 1, 0.70711-> 0.75 (= 1/2 + 1/4), 0.57358-> 0.625 (= 1/2 + 1/8), 0.42262-> 0.375 (= 1/4 + 1/8), 0.087156-> 0.125 (= 1/8), 0.81915-> 0.875 (= 1-1 / 8), 0.90631-> 0.875 (= It is preferable to approximate to 1-1 / 8).

When the contents of Equation 17g are approximated, Equation 17h is obtained.

It should be noted that the result of Equation 17h is generated by one received sequence (preferably a sequence corresponding to the parent sequence index) and the received signal. Even if four PSCs are used in a cell, the receiver obtains values of Equation 17h using only one sequence corresponding to the parent sequence index. In addition, cross-correlation values for all four PSCs may be obtained using the values of Equation 17h.

A method of obtaining the cross correlation values for all four PSCs using the result of Equation 17h is as follows.

Equation 17i is a cross-correlation value between the sequence corresponding to the parent sequence index m0 and the received signal, Equation 17j is a cross-correlation value between the sequence corresponding to the remaining sequence index m1 and the received signal, and Equation 17k is A cross-correlation value between the sequence corresponding to the remaining sequence index m2 and the received signal, Equation 17l is a cross-correlation value between the sequence corresponding to the remaining sequence index m3 and the received signal,

In other words, when generating a plurality of sequences according to the method proposed in the third embodiment, cross-correlation of a plurality of sequences corresponding to a plurality of sequence indexes is performed using a sequence corresponding to one sequence index and a received signal. You can get the value.

18 is a diagram showing the structure of a receiving end according to the fourth embodiment.

As shown in FIG. 18, a signal received by the receiver and a sequence already known by the receiver are input to the index demapper 1900. The 1950 unit of the receiving end of FIG. 18 may obtain Reven_i_i, Revenq_q, Ieven_i_q, Ieven_q_i, Rodd_i_i, Rodd_q_q, Iodd_i_q, and Iodd_q_i through Equation 17g or 17h.

Reven_i_i, Revenq_q, Ieven_i_q, Ieven_q_i, Rodd_i_i, Rodd_q_q, Iodd_i_q, Iodd_q_i are calculated as Reven0, Ieven0, Rodd0, Iodd0, Reven1, Ieven1, Ieven1, Idod, Idod1, Rodd1 by Equation 17c to Equation 17f. For example, Reven_i_i + Reve_q_q by Equation 17c is calculated by Reven ^0, -Ieven_i_q + Ieven_q_i is calculated to be ⁰ Ieven. The calculation according to Equations 17c to 17f is performed in 1960 units.

Reven0, Ieven0, Rodd0, Iodd0, Reven1, Ieven1, Rodd1, and Iodd1, which are the result of the 1960 unit, are added and subtracted according to Equations 17i to 17l to each sequence index (m ₀ , m ₁ , m ₂ , m ₃ ). Four correlation values are obtained. For example, the correlation value for m ₀ is calculated according to Equation 17i. Specifically, the sum of Reven ⁰ and Rodd ⁰ is the real part of the correlation value for m0, and Ieven ⁰ and Iodd ⁰ are summed. The result is the imaginary part of the correlation value for m0.

18 and 17c to 17l, it can be seen that even though the 1960 unit does not exist independently, the final result can be obtained through the result of the 1850 unit, and the final result can be obtained only by the 1960 unit without the 1950 unit. have.

18 is described in another manner as follows.

And a reception signal, to obtain the cross-correlation between a sequence corresponding to the m _o, if the real part of the cross-correlation value for the even-numbered m ₀ sequences as the first result, the first result according to equation 17c Reven ^It becomes ⁰ . In FIG. 18, the first result corresponds to 1901.

In addition, when the imaginary part of the cross correlation values for the even m0 sequence is referred to as the second result, the second result becomes Ieven ⁰ according to Equation 17c. In FIG. 18, the second result corresponds to 1902.

In addition, when the real part of the cross-correlation value for the odd-numbered m0 sequence is called a third result, the third result is Rodd0 according to Equation 17d. In FIG. 18, the third result corresponds to 1903.

In addition, when the imaginary part of the cross-correlation values for the odd m0 sequence is referred to as a fourth result, the fourth result becomes Iodd0 according to Equation 17d. In FIG. 18, the fourth result corresponds to 1904.

In addition, when the real part of the cross-correlation value for the conjugate of the even m0 sequence is a fifth result, the fifth result is Reven1 according to Equation 17e. In FIG. 18, the fifth result corresponds to 1905.

In addition, when the imaginary part of the cross-correlation values for the conjugate of the even m0 sequence is called the sixth result, the sixth result is Ieven1 according to Equation 17e. In FIG. 18, the sixth result corresponds to 1906.

Further, when the real part of the cross-correlation value for the conjugate of the odd m0 sequence is the seventh result, the seventh result is Rodd1 according to Equation 17f. In FIG. 18, the seventh result corresponds to 1907.

In addition, when the gjtn portion of the cross-correlation value for the conjugate of the odd-numbered m0 sequence is an eighth result, the eighth result is Iodd1 according to Equation 17f. In FIG. 18, the eighth result corresponds to 1908.

According to the method described above, the first to eighth results are determined. When the two of the first to eighth results are added to or subtracted from each other, a calculated value of 1970 units is calculated.

For example, the real part of the correlation value for the m0 sequence is the same as the sum of 1901 and 1903. In addition, the imaginary part of the correlation value for the m1 sequence has the same sum result of 1906 and 1906.

In summary, the receiving end may calculate the first to eighth results described above, and add or subtract two of the calculated first to eighth results to obtain a cross correlation value for the m0 to m3 sequences.

18 is an example of the case where the length of the sequence is even. Obviously, the above description can be applied to both even and odd numbers. Hereinafter, a receiver for an odd length sequence will be described with reference to FIG. 19 and equations.

First, when the length is 35, two sequence indices may be selected. For example, the parent sequence index may be 1 in length and the remaining sequence index may be 34 in length.

In this case, the equation corresponding to Equation 17b is shown in Equation 17m.

In this case, the correlation value may be expressed as in Equation 17n.

To simplify the result of Equation 17n, a variable such as Equation 17o is defined.

The result of Equation 17n is displayed based on Equation 17o.

An example of a receiver that performs the operation of Equation 17p is shown in FIG. 19. In FIG. 19, four variables defined in Equation 17o are calculated to calculate correlation values for an odd length sequence at one time.

As described above, a receiver for various length sequences can be designed.

Method 2 of the fourth embodiment

Method 2 of the fourth embodiment describes a method of obtaining a cross-correlation value for sequences having a length of 32 and selected as mo = 1, m1 = 15, m2 = 17, m3 = 32.

Since the specific method has been described in the method 1 of the fourth embodiment, the method 2 displays a specific equation and only the equation of the method 1 corresponds to the equation. Those skilled in the art can implement not only the method 2 of the fourth embodiment but also the reception methods for various types of sequence indexes based on the method 1 of the fourth embodiment.

Equation 18a is the same as that of Equation 17a.

Equation 18b is an equation corresponding to Equation 17b.

Equation 18c is an equation corresponding to Equation 17c.

Equation 18d is an equation corresponding to Equation 17d.

Equation 18e is an equation corresponding to Equation 17e.

Equation 18f is an equation corresponding to Equation 17f.

Equation 18g is an equation corresponding to Equation 17g.

Equation 18h is an equation corresponding to Equation 17h.

Equation 18i is an equation corresponding to Equation 17i.

Equation 18j is an equation corresponding to Equation 17j.

Equation 18k is an equation corresponding to Equation 17k.

Equation 18l is an equation corresponding to Equation 17l.

According to the fourth embodiment, the amount of calculation is reduced as follows. According to the first method of the fourth embodiment, in order to calculate the d-th correlation value for four PSC sequences of length L = 36 and types (if the amount of calculation by the sign converter is ignored), the conventional method is 576. We need real multiplication and 568 real additions.

However, according to the fourth embodiment, 28 real multiplications and 140 real additions are required. In addition, in the case of quantization, the proposed method requires 0 real multiplication, 156 real addition, and 54 bit shift operation.

Since sign inversion and bit shift are not included in the calculation amount in hardware implementation, the amount of calculation for each technique is shown in Table 20. With the proposed method, we can get the cross-correlation values for four PSC sequences with only 156 real time additions.

Calculation # of real multiplications # of real additions Conventional method 576 568 Fourth embodiment 28 140 Fourth Embodiment Approximated Through Quantization 0 156

In addition, when the length (L) is 32, there is a performance difference as shown in Table 21 below.

Calculation # of real multiplications # of real additions Conventional method 512 504 Fourth embodiment 20 120 Fourth Embodiment Approximated Through Quantization 0 132

Those skilled in the art through the above description can be seen that various changes and modifications can be made without departing from the spirit of the present invention. Therefore, the technical scope of the present invention should not be limited to the contents described in the detailed description of the specification but should be defined by the claims.

The sequence generated by the present invention has an excellent effect of maintaining a certain level or more of correlation characteristics in the time domain and showing low PAPR characteristics.

When the sequence proposed in the present invention is applied to a communication standard such as an LTE system, a channel having excellent performance can be configured.

Claims (33)

A method of selecting sequences for a specific channel in a multi-carrier system, Determining any sequence index of a plurality of sequence indexes identifying a sequence of length L as a mother sequence index; Determining at least one remaining sequence index using an integer multiple of half of the value corresponding to one period in the generation formula for the sequence and the parent sequence index; And Selecting sequences corresponding to at least two sequence indexes of the parent sequence index and the remaining sequence indexes as sequences for the particular channel; Containing A method of selecting sequences for a particular channel in a multicarrier system. The method of claim 1, Sequences for the particular channel, Used for Primary Synchronous Channel A method of selecting sequences for a particular channel in a multicarrier system. The method of claim 1, The length L sequence is Sequence with ideal correlation in the time and frequency domains A method of selecting sequences for a particular channel in a multicarrier system. The method of claim 1, The length L sequence is Zadoff-Chu sequence A method of selecting sequences for a particular channel in a multicarrier system. The method of claim 1, When the sequence length L is even, the parent sequence index is m ₀ , the remaining sequence index is m _i , and any positive integer is n, The parent sequence index and the remaining sequence index

or

According to A method of selecting sequences for a particular channel in a multicarrier system. The method of claim 5, The length L sequence is Zadoff-Chu sequence A method of selecting sequences for a particular channel in a multicarrier system. The method of claim 1, When the sequence length L is odd, the parent sequence index is m ₀ , and the remaining sequence index is m _i , The parent sequence index and the remaining sequence index

According to A method of selecting sequences for a particular channel in a multicarrier system. The method of claim 7, wherein The length L sequence is Zadoff-Chu sequence A method of selecting sequences for a particular channel in a multicarrier system. The method of claim 1, When the sequence length L is even, the parent sequence index is m ₀ , the remaining sequence index is m _i , and any positive integer is n, The parent sequence index and the remaining sequence index

or

According to A method of selecting sequences for a particular channel in a multicarrier system. The method of claim 9, The length L sequence is Frank sequence A method of selecting sequences for a particular channel in a multicarrier system. The method of claim 1, When the sequence length L is odd, the parent sequence index is m ₀ , and the remaining index is m _i , The parent sequence index and the remaining sequence index

According to A method of selecting sequences for a particular channel in a multicarrier system. The method of claim 11, The length L sequence is Frank sequence A method of selecting sequences for a particular channel in a multicarrier system. The method of claim 1, The sequence of length L is A sequence of parts of a sequence of prime numbers greater than L A method of selecting sequences for a particular channel in a multicarrier system. The method of claim 1, When a value corresponding to one period in the generation formula is P _L , the mother sequence index is m0, the remaining sequence index is mi, and any positive integer is n, The parent sequence index and the remaining sequence index

or

Following A method of selecting sequences for a particular channel in a multicarrier system. In the method for generating sequences for a specific channel in a multi-carrier system, Determining an arbitrary sequence index among a plurality of sequence indexes that identify a sequence of length L as a mother sequence index, and determining at least one remaining sequence index using the mother sequence index and the sequence length step; Selecting at least two sequence indices of the parent sequence index and the remaining sequence indexes; Generating a sequence corresponding to the selected sequence indices in a frequency domain or a time domain; Performing data processing to remove a direct current (DC) component on the generated sequences; And Converting the sequence on which the data processing is performed into a time domain sequence Containing A method for generating sequences for a particular channel in a multicarrier system. The method of claim 15, When the sequence is a Zadoff-Chu sequence, the sequence length is even, the parent sequence index is m ₀ , the remaining sequence index is m _i , and any positive integer is n, The parent sequence index and the remaining sequence index

or

According to A method for generating sequences for a particular channel in a multicarrier system. The method of claim 15, When the sequence is a Zadoff-Chu sequence, the sequence length is odd, the parent sequence index is m ₀ , and the remaining sequence index is m _i , The parent sequence index and the remaining sequence index

According to A method for generating sequences for a particular channel in a multicarrier system. The method of claim 15, The sequence is a Frank sequence, the sequence length is even, the parent sequence index is m ₀ , the remaining sequence index is m _i , the square root of the sequence length is m, and any positive integer Is n, The parent sequence index and the remaining sequence index

or

According to A method for generating sequences for a particular channel in a multicarrier system. The method of claim 15, When the sequence is a Frank sequence, the sequence length is odd, the parent sequence index is m ₀ , the remaining sequence index is m _i , and the square root of the sequence length is m, The parent sequence index and the remaining sequence index

According to A method for generating sequences for a particular channel in a multicarrier system. The method of claim 15, Generating the sequence, Generating a sequence corresponding to the selected sequence indices in a time domain; Converting to a signal in the frequency domain using an FFT operation or a DFT operation Containing A method for generating sequences for a particular channel in a multicarrier system. The method of claim 20, Performing data processing of repeating the sequence generated in the time domain a plurality of times in the time domain; Containing more A method for generating sequences for a particular channel in a multicarrier system. The method of claim 15, Performing data processing to remove the DC (Direct Current) component, A step of puncturing a subcarrier corresponding to a DC component is performed. A method for generating sequences for a particular channel in a multicarrier system. The method of claim 15, Performing data processing to remove the DC (Direct Current) component, Shifting frequency components or sequence components such that a particular sequence component is not mapped to a subcarrier corresponding to the DC component; A method for generating sequences for a particular channel in a multicarrier system. The method of claim 15, The sequence of length L is A sequence of parts of a sequence of prime numbers greater than L A method for generating sequences for a particular channel in a multicarrier system. A method for calculating cross correlation values of a plurality of sequences with a received signal received from a transmitting side, Obtaining a cross-correlation value between the received signal and a first sequence among the plurality of sequences, obtaining a first result that is a real part and a second result that is an imaginary part of the cross-correlation values for an even first sequence, and an odd number Obtaining a third result that is a real part and a fourth result that is an imaginary part of the cross-correlation values for the first sequence; Obtaining a cross-correlation value between the received signal and the conjugate complex sequence of the first sequence, obtaining a fifth result that is a real part and a sixth result that is an imaginary part of the cross correlation values for the even-numbered conjugate complex sequence, and an odd number Obtaining a seventh result that is a real part and an eighth result that is an imaginary part of a cross correlation value for the first conjugate complex sequence; Summing or subtracting two of the first to eighth results to calculate a real part or an imaginary part of the cross-correlation value for at least one of the plurality of sequences Containing A method for calculating cross correlation values of a plurality of sequences with a received signal received from a transmitting side. The method of claim 25, The plurality of sequences comprises the first to fourth sequences, The plurality of sequences are Zadoff-Chu sequences, The first to fourth sequences correspond to first to fourth sequence indices, When the first sequence index is mo, any one of the second to fourth sequence indexes is mi, the length of the plurality of sequences is L, and any integer is n, The sequence indexes

or

Determined by A method for calculating cross correlation values of a plurality of sequences with a received signal received from a transmitting side. The method of claim 26, The real part of the cross correlation value for the received signal and the first sequence is the sum of the first result and the third result, An imaginary part of the cross-correlation value for the received signal and the first sequence is the sum of the second result and the fourth result A method for calculating cross correlation values of a plurality of sequences with a received signal received from a transmitting side. The method of claim 26, The real part of the cross correlation value for the received signal and the second sequence is obtained by subtracting the fifth result and the eighth result, The imaginary part of the cross correlation value for the received signal and the second sequence is the sum of the seventh result and the sixth result. A method for calculating cross correlation values of a plurality of sequences with a received signal received from a transmitting side. The method of claim 26, The real part of the cross correlation value for the received signal and the third sequence is obtained by subtracting the first result and the fourth result, An imaginary part of the cross-correlation value for the received signal and the third sequence is the sum of the third result and the second result A method for calculating cross correlation values of a plurality of sequences with a received signal received from a transmitting side. The method of claim 26, The real part of the cross-correlation value for the received signal and the fourth sequence is obtained by subtracting the fifth result and the seventh result, An imaginary part of the cross-correlation value for the received signal and the fourth sequence is subtracted from the sixth and eighth results A method for calculating cross correlation values of a plurality of sequences with a received signal received from a transmitting side. A method of calculating cross-correlation values between a plurality of sequences including a first to fourth sequence and a received signal, the method comprising: Obtaining a cross-correlation value between the received signal and a first sequence among the plurality of sequences, obtaining a first result that is a real part and a second result that is an imaginary part among cross-correlation values for an even first sequence, and an odd number Obtaining a third result that is a real part and a fourth result that is an imaginary part of the cross-correlation values for the first sequence; Obtaining a cross-correlation value between the received signal and the conjugate complex sequence of the first sequence, obtaining a fifth result that is a real part and a sixth result that is an imaginary part of the cross correlation values for the even-numbered conjugate complex sequence, and an odd number Obtaining a seventh result that is a real part and an eighth result that is an imaginary part of a cross correlation value for the first conjugate complex sequence; The real part of the cross correlation value for the received signal and the first sequence is calculated using the first result and the third result, and the mutual value for the received signal and the first sequence is calculated using the second result and the fourth result. Calculating a imaginary part of a correlation value; calculating a real part of a cross correlation value for the received signal and the second sequence by using the fifth and eighth results; Calculating a imaginary part of a cross correlation value for the received signal and the second sequence using a result; and using the first and fourth results to cross-correlate the received signal and the third sequence. Calculating a real part of a value and calculating an imaginary part of a cross-correlation value for the received signal and the third sequence using the third result and the second result, and the fifth result A seventh result is used to calculate a real part of the cross correlation value for the received signal and the fourth sequence, and the sixth and eighth results are used to calculate the cross correlation value for the fourth signal and the fourth sequence. Performing at least one of d steps of calculating an imaginary part Including but not limited to: The first sequence index corresponding to the first sequence is set to a predetermined index, Second to fourth sequence indexes corresponding to each of the second to fourth sequences are: Determined based on the first sequence index and the plurality of sequence lengths A method of calculating cross correlation values between a plurality of sequences and a received signal. The method of claim 31, wherein The first to fourth sequences are Zadoff-Chu sequences, When the first sequence index is mo, any one of the second to fourth sequence index is mi, the length of the plurality of sequences is L, and any integer is n, The sequence indexes

or

A method of calculating cross correlation values between a plurality of sequences and a received signal. The method of claim 31, wherein Steps a to d of the at least one step is performed at the same time A method of calculating cross correlation values between a plurality of sequences and a received signal.

Images