KR20090096986A - Apparatus for receiving to improve signal-to-noise ratio of ofdm systems by using oversampling - Google Patents

Abstract

A receiver for improving SNR as much as oversample rate by using over-sampling in an OFDM(Orthogonal Frequency Division Multiplexing) system is provided to improve SNR as much as the oversample rate by converting Nyquist frequency. An ADC(Analog Digital Converter)(10) samples the frequency multiplying the oversampling ratio. A protection remover(20) removes the signal corresponding to protection period corresponding to the guard interval. According to a serializer is the oversampling ratio, the series signal line is parallel converted. An oversampling signal FFT demodulator(40) modulates the signal according to the converted signal. A parallel to serial converter(50) serially converts the demodulated parallel signal.

Description

Apparatus for receiving to improve signal-to-noise ratio of OFDM systems by using oversampling}

The present invention relates to a method and apparatus for improving the signal-to-noise ratio of a demodulated signal by increasing the sampling rate of an AD converter in a receiver of an orthogonal frequency division multiplexing (OFDM) transmission system. The present invention relates to a receiver for over-sampling a received signal at a real time multiple of one or more times the Nyquist frequency in a receiver to improve the signal-to-noise ratio of the demodulated signal.

In general, the OFDM system is a digital communication method used as a modulation method in digital broadcasting. The number and frequency bands of subcarriers vary depending on the purpose of use, but the basic modulation method is the same. The OFDM receiver demodulates the signal by sampling at an analog-to-digital (AD) converter at a speed corresponding to the Nyquist frequency of the OFDM signal and then performing a fast Fourier transform (FFT). The OFDM signal introduces additive white Gaussian noise (AWGN) through the channel. In order to minimize the effect of the incoming AWGN, Korean Patent No. 605109 discloses a method of improving the signal-to-noise ratio of a demodulated signal by oversampling with an integer multiple of the Nyquist frequency.

In Korean Patent 605109, the input analog signal is oversampled into a digital data by oversampling at a frequency corresponding to the oversampling rate times the Nyquist frequency, and the converted digital data is converted to a fast Fourier at a predetermined operation speed, The phase of the frequency domain data is compensated for by referring to the oversampling rate of the digital data which has been converted. The data of the phase compensation are combined and output as a demodulated signal having an improved signal-to-noise ratio corresponding to the oversampling rate.

However, in order to improve the signal-to-noise ratio of the demodulated signal by increasing the sampling rate of the AD converter in the OFDM receiver, the method of determining the sampling rate of the AD converter only as an integer multiple of the Nyquist frequency has a high Nyquist frequency. In the case of oversampling, the frequency of the AD converter is excessively high, resulting in the inability to implement the AD converter or the sudden increase in the price of the AD converter.

In addition, in case of integer sampling, the maximum sampling frequency of the AD converter and the maximum sampling rate of the AD converter cannot be fully utilized.

For example, when the maximum sampling rate of an AD converter is 2.8 times the Nyquist frequency of the OFDM signal to be demodulated, the method proposed in Korean Patent No. 605109 shows that the maximum oversampling rate is 2 because the real multiples between integers are not possible. And thus the signal-to-noise ratio is 2.

However, according to the present invention, which oversamples a received signal of a receiver of an OFDM transmission system by more than 1 times the Nyquist frequency, the maximum sampling rate of the AD converter is 2.8 times the Nyquist frequency of the OFDM signal to be demodulated. In this case, 2.8 times oversampling is possible, which improves the signal-to-noise ratio by 2.8 times. Thus, the maximum sampling frequency of the AD converter and the maximum sampling rate of the AD converter can be fully utilized.

In the present invention, when the OFDM receiver is to increase the sampling rate of the AD converter to improve the signal-to-noise ratio of the demodulated signal, the AD conversion is possible by a real multiple of the Nyquist frequency. Signal-to-noise ratio using oversampling in an orthogonal frequency division multiplexing system that can freely determine the sampling rate within the maximum speed of the AD converter and improve the signal-to-noise ratio as much as the over-sampling ratio, compared to the conventional method that AD conversion was possible. An object of the present invention is to provide a receiver for improving.

The present invention for performing such an object,

An AD converter that samples the analog baseband signal at a frequency multiplied by the Nyquist frequency multiplied by a real multiple of the oversampling rate;

A guard interval remover connected to the output side of the AD converter to remove a signal corresponding to the guard interval according to the oversampling ratio;

A serial-to-parallel converter connected to an output side of the guard interval eliminator and converting serial signal strings in parallel according to the oversampling ratio;

An oversample signal FFT demodulator for demodulating a signal passing through a serial-to-parallel converter according to the oversample rate; And

And a parallel-to-serial converter for serially converting the parallel signals output from the FFT demodulator.

The present invention can freely determine the sampling rate within the maximum sampling rate of the AD converter larger than the Nyquist frequency in order to increase the sampling rate of the AD converter in the OFDM receiver to improve the signal-to-noise ratio of the demodulated signal. Maximum signal-to-noise ratio performance can be achieved. For example, when the maximum speed of the AD converter is 2.8 times the Nyquist frequency of the OFDM signal to be demodulated, the conventional integer multiple method only doubles the oversampling and the signal-to-noise ratio is twice as high. According to the present invention, the 2.8 times oversampling is possible and the signal-to-noise ratio is improved by 2.8 times.

1 is a diagram illustrating a structure of a general OFDM symbol, FIG. 2 is a block diagram illustrating a sampling and demodulation process of a general OFDM receiver, and FIG. 3 is a sampling and demodulation of an OFDM receiver oversampled according to the present invention. A block diagram showing the process.

길이의 보호구간(guard interval)과

길이의 유용구간(useful symbol period)으로 구성되어 있다.

개의 부반송파를 사용하는 OFDM 시스템인 경우, 나이퀴스트 주파수로 표본화하는 일반적인 OFDM 수신기에서는

구간동안

개의 표본을 취한다. OFDM symbols are continuously transmitted OFDM symbols, OFDM symbol as shown in Figure 1

The guard interval of the length

It consists of a useful symbol period of length.

In an OFDM system using two subcarriers, a typical OFDM receiver sampling at the Nyquist frequency

During the interval

Take a sample of dogs.

개를 모아 병렬 신호로 변환하는 직병렬 변환기(30)가 접속되고, 직병렬 변환기(30)의 출력측에는 동시에 입력된

개의 신호를 푸리에 변환하는 FFT처리기(40)가 접속되며, FFT처리기(40)의 출력측에는

개의 복조된 병렬 신호를 직렬로 변환하여 출력하는 병직렬 변환기(50)가 접속되어 구성된다. In the conventional OFDM receiver, as shown in FIG. 2, a guard interval remover 20 for removing a sample corresponding to a guard interval from an OFDM symbol is connected to an output side of the AD converter 10 for converting a baseband analog signal into a digital signal. . The output signal of the guard section remover 20 has a serial signal.

A serial / parallel converter 30 which collects dogs and converts them into parallel signals is connected, and simultaneously input to the output side of the serial / parallel converter 30.

An FFT processor 40 for Fourier transforming the four signals is connected, and on the output side of the FFT processor 40,

The parallel-to-serial converter 50 which converts two demodulated parallel signals in series and outputs them is connected and comprised.

In the conventional OFDM receiver, if the input signal of the FFT processor 40 is expressed as <Equation 1>,

<식 1>

<Equation 1>

The output of the FFT processor 40 can be expressed as Equation 2.

<식 2>

<Equation 2>

은 FFT 크기가

임을 의미한다. In <Equation 2>

Is the FFT size

Means.

이라 할 때, 나이퀴스트 주파수의

배이고, 과표본화율

은 <식 3>을 만족하도록 결정된다.In contrast, in the structure of the OFDM receiver according to the present invention, as shown in FIG. 3, the baseband analog signal is converted into a digital signal by the AD converter 210. At this time, the sampling rate is determined by the oversampling rate.

Is called the Nyquist frequency

Double, oversample rate

Is determined to satisfy <Equation 3>.

<식 3>

<Equation 3>

은

인 어떤 정수이고,

은

보다 작거나 같은 최대 정수이다.In <Equation 3>

silver

Is any integer that is

silver

Maximum integer less than or equal to

개의 표본마다 병렬로 변환된다. 직병렬 변환기(230)를 통과한 신호는 과표본신호 FFT 복조기(240)에서

이 1 인 경우와

인 경우로 나누어 복조하고 신호처리하여

개의 디지털 신호를 출력한다. 과표본신호 FFT 복조기(240)에서 출력된

개의 디지털 신호는 병직렬 변환기(250)를 통과하여

개의 병렬 신호가 직렬로 변환되어 출력된다.The digital signal converted by the AD converter 210 is input to the guard interval eliminator 220 to remove a sample corresponding to the guard interval, and is input to the serial-to-parallel converter 230.

Samples are converted in parallel. The signal passing through the serial-to-parallel converter 230 is transferred from the oversampled signal FFT demodulator 240.

If it is 1 and

To demodulate and signal

Output digital signals. Oversampled signal output from FFT demodulator 240

Digital signals pass through parallel-to-serial converter 250

Parallel signals are converted and outputted in series.

이 1 인 경우와

인 경우(과표본화율이

인 경우 만 가능)에 과표본신호 FFT복조기(240)의 구조가 달라진다. 먼저,

이 1 일 때 과표본신호 FFT 복조기(240)의 구조는 도 4에서 보는 바와 같이 FFT처리기(241)의 출력단에 과표본 제거기(242)가 접속된 구성을 갖는다.In the present invention

If it is 1 and

If (the oversample rate is

Only if the structure of the over-sampled signal FFT demodulator 240 is different. first,

In this case, the oversampled signal FFT demodulator 240 has a structure in which the oversampler 242 is connected to the output terminal of the FFT processor 241 as shown in FIG.

이 1인 경우에

개의 입력 신호를 받아 FFT처리기(241)에서

점(

-point) FFT를 수행한다. 이때 FFT처리기(241)는

점 FFT를 하는 것이 아니라

점 FFT를 수행한다. FFT처리기(241)의 입력은 <식 4>로 표현되며, FFT처리기(241)의 출력은 <식 5>로 표현된다.

If this is 1

FFT processor 241 receives two input signals

point(

-point) Perform FFT. At this time, the FFT processor 241

Instead of doing a point FFT

Perform a point FFT. The input of the FFT processor 241 is represented by <Equation 4>, and the output of the FFT processor 241 is represented by <Equation 5>.

<식 4>

<Equation 4>

<식 5>

<Equation 5>

개의 신호는 도 5와 같이 데이터 전송에 사용한

개의 부반송파에 해당한다. 따라서 과표본 제거기(242)에서 마지막

개의 신호를 제거하고 출력한다. 과표본 제거기(242)의 출력

는 <식 6>과 같이 나타낼 수 있다.First in the result of FFT in FFT processor 241

Signals are used for data transmission as shown in FIG.

Corresponds to two subcarriers. Therefore, in oversample eliminator 242

Remove and output the two signals. Output of Oversample Eliminator 242

Can be expressed as shown in Equation 6.

<식 6>

<Equation 6>

개의 입력에는 데이터 성분과 잡음성분이 모두 포함되어 있는데, FFT처리기(241)의 출력에서 마지막

개의 신호에는 잡음성분만 포함되어 있으므로, 과표본 제거기(242)에서 마지막

개의 신호를 제거함으 로써 과표본신호 FFT복조기(240)의 출력은 과표본신호 FFT복조기(240)의 입력에 비해 잡음이

로 감소한다. 따라서 SNR은 r 배 향상된다.Of the FFT processor 241

Inputs contain both data and noise components, the last of the output of FFT processor 241

Signals contain only noise components, so the oversampler 242

By removing the two signals, the output of the oversampled signal FFT demodulator 240 is noisy compared to the input of the oversampled signal FFT demodulator 240.

Decreases. Therefore, the SNR is improved r times.

일 때 과표본 신호 FFT 복조기(340)의 구조는 도 6에서 보는 바와 같이 직병렬 변환기(230)를 통과한 신호를 입력받는 역다중화기(341)의 출력단에

개의 FFT 처리기로 구성된 FFT처리기 블록(342)이 접속되며, FFT 처리기 블록(342)의 출력단에는 FFT처리기 블록(342)을 구성하는

개의 FFT처리기와 각각 접속되는

개의 과표본제거기로 구성된 과표본 제거기 블록(343)이 접속된다. 과표본 제거기 블록(343)의 출력단에는 과표본 제거기 블록(343)을 구성하는

개의 과표본 제거기중 하나의 과표본 제거기를 제외한 나머지 과표본 제거기와 각각 접속되는

개의 위상 보상기로 구성된 위상 보상기 블록(344)이 접속된다. 위상보상기 블록(344)의 출력단에는 벡터 덧셈기(345)가 접속되어 구성된다. Also,

When the oversampled signal FFT demodulator 340 has a structure at the output terminal of the demultiplexer 341 that receives the signal passing through the serial-to-parallel converter 230, as shown in FIG.

An FFT processor block 342 composed of two FFT processors is connected, and an output terminal of the FFT processor block 342 configures an FFT processor block 342.

Connected to each of the four FFT processors

An oversampler block 343 consisting of two oversamplers is connected. At the output of the oversample remover block 343, an oversample remover block 343 is formed.

Are connected to each of the oversampler except one of the oversampler

A phase compensator block 344 consisting of two phase compensators is connected. The vector adder 345 is connected to the output terminal of the phase compensator block 344.

개의 신호로 이루어진 직병렬 변환기(230)의 출력은 역다중화기(341)를 통해서 각각이

개의 신호로 이루어진

개의 병렬신호를 출력한다 즉

을 만족한다. 여기서

은 AD 변환기의 과표본화율이며,

은 병렬신호로 변환된 후 각 경로신호(

개의 경로에 입력되는 경로신호)에서의 과표본화율에 해당한다. 정리하면 <식 3>으로부터

와 병렬 경로수

과의 관계는 <식 7>과 같이 표현되며 <식 8>을 만족해야 한다.

The output of the serial-to-parallel converter 230, which is composed of four signals, is respectively passed through the demultiplexer 341.

Signal

Output parallel signals

To satisfy. here

Is the oversample rate of the AD converter,

Is converted to a parallel signal and then each path signal (

Corresponding to the oversampling rate). In summary, from <Equation 3>

And parallel paths

The relationship with is expressed as <Equation 7> and must satisfy <Equation 8>.

<식 7>

<Equation 7>

<식 8>

<Equation 8>

일 때,

면

이고, 가능한

쌍은 <식 7>과 <식 8>을 만족해야 하므로,

,

,

,

가 된다. 역다중화기(341)의 입력을 <식 9>와 같이 표현할 때, E.g

when,

if

Is possible

The pair must satisfy <Equation 7> and <Equation 8>,

,

,

,

Becomes When the input of the demultiplexer 341 is expressed as <Equation 9>,

<식 9>

<Equation 9>

번째 FFT처리기(342)로 입력되는 역다중화기(341)의 출력

은 다음 <식 10>과 같이 나타낼 수 있다.

Of demultiplexer 341 input to the first FFT processor 342

Can be expressed as follows.

<식 10>

<Equation 10>

개의 FFT로 이루어진 FFT처리기 블록(342)에서 각 FFT처리기는

개의 입력을 받아

점 FFT를 수행한 후 출력한다. 이 때 FFT처리기 블록(342)의

번째 FFT의 출력을

이라 할 때 다음 <식 11>과 같이 나타낼 수 있다.

In an FFT processor block 342 consisting of four FFTs, each FFT processor

Take input

Output after performing point FFT. At this time, the FFT processor block 342

Output of the first FFT

This can be expressed as in the following <Equation 11>.

<식 11>

<Equation 11>

번째 FFT처리기의 출력벡터에서 k 번째 원소를

라 놓으면 다음 <식 12>와 같은 관계식이 성립된다.In <Equation 11>

The k th element in the output vector of the first FFT

If you put it, the relation like the following <Equation 12> is established.

,

;

<식 12>

,

;

<Equation 12>

개의 원소로 이루어진 입력벡터에서 처음 마지막

개의 원소를 제거하고

개의 원소로 이루어진 벡터

을 출력한다. In each oversample remover of oversample remover block 343, as shown in Equation 13,

First last in input vector of elements

Removed elements

Vector of elements

Outputs

                                                                        <Equation 13>

번째 과표본 제거기의 출력벡터의 k 번째 원소

의 위상을 <식 14>에서와 같은 위상만큼 회전시키면

와 동일한 값을 얻게 된다. 즉

번째 위상보상기의 출력벡터

은

가 된다. From Equation 11, in phase compensator block 344

K th element of the output vector of the first oversample eliminator

If we rotate the phase of by the same phase as in <Equation 14>

You get the same value as. In other words

Output vector of the second phase compensator

silver

Becomes

<식 14>

<Equation 14>

번째 위상보상기의 구조가 성립된다. Thus, as shown in FIG.

The structure of the first phase compensator is established.

개의 입력 벡터를 산술적으로 더하여 출력하여 벡터 덧셈기(345)의 출력

는 다음 <식 15>와 같다.The vector adder 345 simply

Output of vector adder 345 by arithmetically adding the two input vectors

Is shown in Equation 15 below.

<식 15>

<Equation 15>

배로 커지고, 불규칙한 잡음은 더해져 서로 상쇄된다. 그리므로, 과표본신호 FFT 복조기(340)은

개의 입력을 받아 최종적으로

개를 출력하므로 SNR은

배 향상된다. In the operation of the vector adder 345, the data components are added with the same value.

It is doubled and irregular noise is added to cancel each other out. Thus, the oversampled signal FFT demodulator 340 is

Take inputs and finally

Outputs, so the SNR is

Times are improved.

Although the preferred embodiments of the present invention have been described in detail with reference to the accompanying drawings, the present invention is not limited thereto and may be improved or modified by those skilled in the art within the scope of the technical idea of the present invention.

1 is a diagram illustrating a structure of a general OFDM symbol.

2 is a block diagram illustrating a sampling and demodulation process of a typical OFDM receiver.

3 is a block diagram illustrating a sampling and demodulation process of an OFDM receiver oversampled according to the present invention.

일 때 본 발명에 따른 과표본 신호 FFT 복조기의 처리과정을 도시한 블록도이다. 4 is the number M of parallel paths in the present invention.

Is a block diagram showing the processing of the oversampled signal FFT demodulator according to the present invention.

일 때 본 발명에 따른 과표본 신호 FFT 복조기에서 FFT 출력의 개수 및 형태를 나타낸 그래프이다. 5 is the number M of parallel paths in the present invention.

Is a graph showing the number and shape of the FFT output in the oversampled signal FFT demodulator according to the present invention.

일 때 본 발명에 따른 과표본 신호 FFT 복조기의 처리과정을 도시한 블록도이다. Figure 6 is the number M of parallel paths in the present invention

Is a block diagram showing the processing of the oversampled signal FFT demodulator according to the present invention.

7 is a block diagram illustrating the operation of a phase compensator inside an oversampled signal FFT demodulator of the present invention.

Claims (8)

An AD converter that samples the analog baseband signal at a frequency multiplied by the Nyquist frequency multiplied by a real multiple of the oversampling rate; A guard interval remover connected to an output side of the AD converter to remove a signal corresponding to the guard interval according to the oversampling ratio; A serial-to-parallel converter connected to an output side of the guard interval eliminator and converting serial signal strings in parallel according to the oversampling ratio; An oversample signal FFT demodulator for demodulating the signal passing through the serial-to-parallel converter according to the oversample rate; And And a parallel-to-serial converter for serially converting parallel signals output from the oversampled signal FFT demodulator. 2. The sampling rate of the AD converter according to claim 1, wherein

Is called the Nyquist frequency Doubled and oversampled

Is the number of parallel paths

The number of digital signals output from the FFT demodulator

When I say And the receiver is determined to be this constant. The FFT demodulator of claim 1 or 2, wherein the oversampled FFT demodulator

this

And a supersampler is connected to the output of the FFT processor when = 1. The FFT demodulator of claim 1 or 2, wherein the oversampled FFT demodulator

this

When the FFT processor block composed of a plurality of FFT processor is connected to the output terminal of the demultiplexer receiving the output signal of the serial-to-parallel converter, and the output terminal of the FFT processor block and the plurality of FFT processor constituting the FFT processor block; An oversample remover block composed of a plurality of oversample removers connected to each other is connected, and at the output end of the oversample remover block, the rest of the oversamplers except one oversample remover of the plurality of oversample removers constituting the oversample remover block are connected. And a phase compensator block comprising a plurality of phase compensators respectively connected to the canceller, and a vector adder connected to an output terminal of the phase compensator block. The method of claim 5, wherein

Of the demultiplexer input to the first FFT processor

Oversampling rate

And the number of parallel paths

In this case, the receiving device characterized in that represented by the following <Equation 1>.

<Equation 1> 6. The method according to claim 1 or 5, wherein the oversampled signal FFT demodulator has a supersample rate.

Depending on the

A receiver characterized in that it is an FFT for point discrete Fourier transform. 6. The oversample remover of claim 3 or 5, wherein the oversample remover constituting the oversample signal FFT demodulator

A receiver characterized in that it removes the remaining samples leaving only two samples. The method of claim 5, wherein the phase compensator block

Of the output vector of the first oversample eliminator

Th element

Rotate the phase by the phase expressed in <Equation 2>

A receiving device, characterized in that to obtain the same value as.

<Equation 2>

Images