JP2000041019A - Multi-carrier transmission system and receiving device - Google Patents

Abstract

(57) [Problem] In a multicarrier transmission system, a demodulation side estimates a frequency difference caused by a local oscillator output frequency difference between transmission and reception and a frequency error caused by Doppler fluctuation, and
Correct the error. SOLUTION: A received pilot symbol 54a is supplied to a multiplier 20 via a switch 12, and is multiplied by a demodulation-side pilot symbol 56c output from a parallel-serial converter 23. Thereby, the convolution operation in the frequency domain can be calculated as multiplication in the time domain. At this time, the reception pilot symbol 54a and the demodulation-side pilot symbol 56c are set based on a code sequence having excellent autocorrelation characteristics such as an M sequence. In this case, the output signal of the multiplier 20 has a sharp peak corresponding to the frequency error in the frequency domain, and the peak detector 26 detects this peak position. Thereafter, the NCO 29 generates a complex signal 54g corresponding to the frequency error, and the signal is multiplied by the data symbol in the multiplier 14. Thus, the frequency error is suitably corrected on the demodulation side.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a multicarrier transmission system and a receiving apparatus, for example, an orthogonal frequency division multiplexing (OFDM).
The present invention relates to a technique for estimating a frequency error caused by a frequency difference of a local oscillator output between transmission and reception, Doppler fluctuation, and the like in a lexing system.

[0002]

2. Description of the Related Art In recent years, an OFDM system is a communication system whose application to a terrestrial digital broadcasting system or the like is being considered, and is known as a TDMA (Time Division Multiple Acces
This is a type of multi-carrier transmission system that performs carrier wave transmission using two or more waves, unlike a system that performs carrier wave transmission using only one wave, that is, single carrier transmission as in s).

[0003] According to the single carrier transmission, a large signal loss occurs in a situation such as so-called selective fading in which fading characteristics generated in a transmission path differ for each frequency within a transmission band. On the other hand, according to multicarrier transmission, due to the property of using a plurality of carriers, even if some carriers are affected by level drop due to selective fading, carriers that are not affected or carriers that are less affected are still present. Remains. Therefore, OFDM adopting such multi-carrier transmission
The system can be said to be a method advantageous for improving the error rate characteristics with respect to multipath fading.

Further, in an OFDM system, a time width Tg (hereinafter, referred to as a "guard interval") taking into account the delay time is given to intersymbol interference caused by a delayed wave generated on a transmission path being combined with a preceding wave and received. .) Is added to the beginning of each symbol. The guard interval Tg is set by copying the latter half time signal in the symbol. Note that a portion of one symbol time excluding a portion called a guard interval Tg is called an effective symbol length.

In this method, each carrier (hereinafter, referred to as carrier)
It is called "subcarrier". ) Is equal to the reciprocal of the effective symbol length Ts as shown in FIG. 17 (that is, ΔF = 1 / Ts). By doing so, the receiving side can suppress the influence from other subcarriers by the integration operation for a time equal to the effective symbol length Ts for each symbol.

[0006] In such an OFDM system,
First, it is common to apply IFFT (fast inverse Fourier transform) on the modulation side and FFT (fast Fourier transform) on the demodulation side. In particular, in the FFT processing on the demodulation side,
The subcarriers are separated from each other using the frequency information of each subcarrier and the information of the sampling frequency, and each subcarrier is demodulated by an integration operation over a time of one effective symbol length Ts for each symbol.

However, considering an actual transceiver, the local oscillators on the transmitting side and the receiving side are independent, and there is a frequency difference between the two. In addition, if the transmitter or the receiver moves, a frequency difference due to Doppler fluctuation may occur. Therefore, even if the frequency interval between subcarriers (Δ
Even if F) is correctly maintained, the frequency conversion output of the receiver, such as a baseband signal, is frequency-shifted by a certain amount, which adversely affects demodulation characteristics.

[0008] In order to solve such a problem, conventionally, two methods are mainly proposed, which are roughly classified depending on whether or not a known pattern called a pilot symbol is used.

First, as a method not using a pilot symbol, there is a method utilizing the fact that the same signal as that of the guard interval Tg also exists in the latter half of the OFDM symbol. In this method, a frequency error is estimated using a phase rotation amount obtained from a complex correlation between a guard interval portion Tg and a copy source portion thereof.

Hereinafter, this method will be described with reference to FIG. FIG. 1 shows only a part related to frequency error estimation in a receiving apparatus used in an OFDM system. In this device, the complex reception signal 6 a that has been quadrature-detected and input to the terminal 7 is input to the multiplier 14. Then, after the symbol timing is established for the complex signal 6b that is the output, the guard interval extracting unit 32 extracts and holds, for example, all the sample data of the guard interval Tg portion. The conjugate processing unit 33 generates a conjugate signal by inverting the sign of the orthogonal component of the complex signal 6c. Such a conjugate signal is subjected to complex multiplication by the multiplier 34 with respect to the portion of the complex signal 6b after the guard interval Tg in the same symbol.

Since the guard interval Tg and its copy source component are originally the same signal, the amount of frequency error is expressed as a phase change amount according to the delay time between the guard interval and its copy source. This phase change amount is calculated by the phase change amount calculation unit 35. And the Δf converter 3
In step 7, when the timing of the guard interval Tg and the timing of the copy source coincide, the output of the phase change amount calculation unit 35 is averaged with all the sample data in the guard interval period, thereby estimating the frequency error.

Thereafter, NCO (Numerical Controlled O)
The scillator 29 supplies the complex signal 6d corresponding to the estimated value to the multiplier 14. Thus, the complex signal 6d reflecting the frequency error is multiplied by the complex received signal 6a, and as a result, the influence of the frequency error on the receiver is suppressed to a low level. Then, the complex signal 6b output from the multiplier 14 is divided into each subcarrier component by the serial-parallel conversion unit 16, and the output signal 6f is Fourier transformed by the Fourier transform unit 17. Further, the output signal 6g is serial-converted by the parallel-serial conversion unit 18 and is provided for subsequent processing.

On the other hand, the method using pilot symbols is performed, for example, with a configuration as shown in FIG. In the figure, components corresponding to those in the device shown in FIG. 18 are denoted by the same reference numerals, and description thereof is omitted here. The apparatus shown in the figure roughly estimates in a first estimator 380 a large frequency error larger than the subcarrier interval ΔF, and uses a second estimator for a small frequency error smaller than the subcarrier interval.
The estimation unit 410 estimates. Normally, in an OFDM system, subcarriers are arranged for each frequency interval ΔF, which is the reciprocal of the effective symbol length Ts. If the number of subcarriers is N and the power of 2 larger than N is M (= 2 ＾ L; L is a natural number), the sampling frequency for the OFDM signal is expressed as M / Ts (= M · ΔF). You. However,
If the subcarriers are not arranged consecutively, M that includes all the subcarriers is selected. That is, the sampling frequency is set so as to cover a wider range than the band in which the transmitted subcarrier exists, and FFT is performed using the sampling frequency information f and the information of the subcarrier interval ΔF. And the first
The estimating unit 380 uses the output of the FFT 17 and the power calculating unit 38 calculates the power component of the frequency component 6i higher by one subcarrier than the highest frequency subcarrier in the subcarrier group. Among them, the power component of the frequency component 6j lower by one subcarrier than the lowest frequency subcarrier is calculated.

In this method, since these two frequencies are out of the band in which the subcarriers should be arranged, rough estimation of the frequency error is performed by calculating the power components of the frequency components 6i and 6j. That's what I'm trying to do. That is, the comparing section 40 included in the first estimating section 380 compares the outputs of the power calculating sections 38 and 39 and estimates the frequency error based on the magnitude relation. The estimation result is supplied to the multiplier 14 as a complex signal 6k via the combining unit 42 and the NCO 29, and the complex reception signal 6k
multiplied by a.

The second estimating section 410 includes a first estimating section 3
It performs frequency error estimation with higher accuracy than 80,
In the second estimating unit 410, the phase change amount calculating unit 41
Complex signal 6 output from parallel-serial converter 18
Using h, the phase change amount of pilot symbols that are temporally separated in the same subcarrier is calculated, and the calculated phase change amount is averaged to some extent. After that, the Δf converter 43 sets the phase change amount
1 is converted to a frequency error. In addition, the phase change amount calculation unit 41 may calculate the phase change amount using the phase change amounts of the pilot symbols in all the subcarriers.

The output of the Δf conversion unit 43 is
Are combined with the frequency error calculated by the first estimating unit 380, and the complex signal 6k is converted into the complex received signal 6 using the NCO 29.
Returned to a. Thus, the first estimating unit 380 and the second
Using the estimating unit 410, it is possible to preferably suppress the frequency error in the receiving device.

[0017]

When the former method (FIG. 18) is used among the above conventional methods, the original transmission information (hereinafter referred to as "data symbol") is called a pilot symbol.
That. ) Becomes unnecessary, and transmission efficiency can be improved.

However, in such a system, the time difference between the guard interval and the copy source in the same symbol determines the frequency range in which the frequency error can be estimated. For example, as shown in FIG. Then, the range of the frequency error that can be estimated is limited from -1 / (2ΔT) to + 1 / (2ΔT). And this Δ
Since T is equal to Ts (= 1 / ΔF) in the case of OFDM, it is impossible to estimate a frequency error exceeding the subcarrier interval ΔF, for example.

Further, suppose that -1 / (2ΔT) is changed to + 1 /
Even if a frequency error exists within the range of (2ΔT), a large number of pilot symbols to be observed are required to obtain a stable error estimation result. For example, the C / N ratio of a complex reception signal is low. If it is low, it is necessary to observe several tens of symbols or more before stable reception of the carrier. For this reason, in the above-mentioned method, there is a problem that it takes time until stable reception of the carrier. And this conventional method is OFDM
Similarly, in multi-carrier transmission using a guard interval which is not limited to this, there is still a problem that the estimation range is narrow and the stabilization time is long.

On the other hand, when the latter method (FIG. 19) of the above-mentioned conventional method is used, it is possible to cope with a frequency error equal to or longer than the subcarrier interval ΔF, but the first and second estimation methods are also used in this method. In order to increase the estimation accuracy of units 380 and 410, a large number of observation symbols is required. Particularly, in the second estimating unit 410, it is necessary to reduce the number of pilot symbols as much as possible from the viewpoint of transmission efficiency. As a result, there is a problem that it takes time to pull in the received signal.

In this regard, the second estimating section 410 may calculate the phase change amount for each subcarrier in addition to the time direction, and average it. However, in that case, there is a problem that the processing amount increases as the number of subcarriers increases.

Further, not only the pilot symbols but also the data symbols which change at random are also stored in the second estimator 41.
A method used for estimating the frequency error at zero is also conceivable. However, this method has a problem that it takes a very long time to average random data components under the condition that the residual frequency error by the first estimator 380 is sufficiently small and the C / N ratio is low. In the conventional system shown in FIG. 19, the same problem remains for multicarrier transmission using pilot symbols.

In short, in the conventional method, 1) when the pilot symbol is not used, there is a problem that the stabilization time is increased and the detection frequency range is limited. 2) The pilot symbol is not used. Even when the method is used, there is a problem that the stabilization time increases and the processing amount also increases in order to increase the estimation accuracy.

SUMMARY OF THE INVENTION The present invention has been made in view of the above problems, and has as its object to provide a multicarrier transmission system and a receiver capable of performing frequency error estimation with a simple configuration on a receiver side. Is to do. Another object of the present invention is to provide a multicarrier transmission system capable of estimating a frequency error in a relatively short period of time and with high accuracy, and a receiver thereof.

[0025]

(1) In order to solve the above problems, the present invention includes a transmitting device and a receiving device, and transmits a transmission data sequence from the transmitting device to the receiving device by multicarrier transmission. In the multi-carrier transmission system, the transmitting device sequentially performs serial-parallel conversion and multi-carrier modulation on a code sequence including a predetermined modulation-side pilot code sequence for a pilot in the transmission data sequence, and generates a transmission code sequence. Including transmission code string generation means to generate, the receiving device receives the transmission code string, for a portion corresponding to the modulation-side pilot code string,
Multiplying means for multiplying a predetermined demodulation-side pilot code string corresponding to the modulation-side pilot code string; Fourier transform means for performing a Fourier transform on the multiplication result; and the transmitting device and the reception section based on the result of the Fourier transform. Frequency error estimating means for estimating a frequency error occurring between the apparatus and the apparatus.

The present invention also relates to a receiving apparatus used for a multicarrier transmission system, wherein the receiving apparatus receives a transmitting code string including a predetermined pilot pilot code string for a pilot from the transmitting apparatus, and includes: Multiplication means for multiplying a portion corresponding to the modulation-side pilot code string by a predetermined demodulation-side pilot code string corresponding to the modulation-side pilot code string; Fourier transform means for performing a Fourier transform on the multiplication result; Frequency error estimating means for estimating a frequency error occurring between the transmitting device and the receiving device based on a result of the conversion.

According to the present invention, in a multicarrier transmission system such as an OFDM system, a transmission apparatus includes a predetermined modulation-side pilot code string in a transmission data sequence. The modulation-side pilot code sequence may be a code having a high autocorrelation characteristic, such as a code sequence generally called an M sequence, or a sequence obtained by multiplying the M sequence and a code sequence called a Barker sequence. It is desirable to use columns. Then, the demodulation side multiplies the modulation-side pilot code sequence in the transmission code sequence by the demodulation-side pilot code sequence. As the demodulation-side pilot code sequence, for example, a code sequence obtained by time-inverting the modulation-side pilot code sequence and multi-carrier modulation, or a code sequence obtained by multiplying the code sequence by a code sequence representing a window function. Can be adopted.

Thus, it is possible to obtain a code sequence having a spectrum distribution corresponding to a frequency error generated between the transmitting device and the receiving device. Therefore, based on the result obtained by performing the Fourier transform on the result of the multiplication, it is possible to appropriately estimate the frequency error generated between the transmitting device and the receiving device.

Further, since multiplication processing is performed on serial data before performing serial-parallel conversion or Fourier transformation on the transmission code string, the same processing is performed on the frequency domain after performing serial-parallel conversion or Fourier transformation. It is possible to estimate the frequency error with a simple configuration as compared with the case where the above is performed. Further, according to the present invention, highly accurate frequency error estimation can be performed using only a relatively small number of modulation-side pilot code strings. Also, the range of the frequency error that can be estimated can be made wider than in the conventional scheme using pilot symbols.

Here, the term “multi-carrier modulation” refers to a modulation method in which a code string that has been subjected to serial-parallel conversion is assigned to a plurality of carriers having different frequencies and transmitted to the receiving side. In this multi-carrier modulation, a case where the frequencies of a plurality of carriers are orthogonal to each other corresponds to the OFDM modulation.

In the case where OFDM is applied as one aspect of the present invention, the Fourier transform means includes a subcarrier for the transmission data sequence other than a portion corresponding to the modulation-side pilot code sequence in addition to the multiplication result. It is characterized in that it is also used for Fourier transform generally used for separation. In this case, when OFDM is applied, there is no need to provide two types of Fourier transform means for estimating the frequency error and Fourier transform means for separating the subcarriers, thereby simplifying the configuration of the receiver. Can be. In the case other than OFDM, a time-frequency conversion unit for separating subcarriers may be prepared separately from a Fourier transform for estimating a frequency error, and operated on signals other than the pilot code string. In this case, subcarriers can be separated by applying a phase change amount determined by a subcarrier frequency and a sampling frequency set for each subcarrier.

(2) In one aspect of the present invention, the transmission data sequence is a complex code sequence, and the transmitting device has 0 in one of in-phase and quadrature components of the modulation-side pilot code sequence. In addition to being assigned, each bit of a predetermined code string is sequentially assigned to the other. By doing so, it is possible to estimate the frequency error using only one of the in-phase and quadrature components of the signals handled by the system, and the configuration of the system can be simplified.

(3) In one aspect of the present invention, the transmission code sequence is a sequence of complex numbers, and the receiving device converts the complex number based on the frequency error estimated by the frequency error estimating means into the transmission code sequence. A frequency error correcting means for reducing the influence of a frequency error generated between the transmitting device and the receiving device by complex multiplying the column is further included. This makes it possible to control the phase of the transmission code string and easily reduce the influence of the frequency error.

(4) In one aspect of the present invention, the receiving device controls an oscillation frequency of a local oscillator included in the receiving device based on the frequency error estimated by the frequency error estimating means, It is characterized by further including a frequency error correcting means for reducing an influence of a frequency error generated between the transmitting device and the receiving device. Even in this case, the influence of the frequency error can be reduced on the receiver side.

(5) In one aspect of the present invention, the receiving device further includes a storage unit for storing the demodulation-side pilot code string. That is, as described above, the demodulation-side pilot code sequence is multi-carrier modulated on a code sequence obtained by time-inverting the modulation-side pilot code sequence, and the converted data is further multiplied by a window function to perform real-time processing. However, in this aspect, the demodulation-side pilot code string is stored in the storage means in advance, so that the configuration of the receiving apparatus can be further simplified.

(6) In one aspect of the present invention, the frequency error estimating means included in the receiving device, based on the result of the Fourier transform and predetermined correction data, determines whether the transmitting device and the receiving device have the same frequency. A frequency error is estimated, and the predetermined correction data corresponds to at least the demodulation-side pilot code sequence and is calculated in advance.

That is, the result of the Fourier transform has a frequency characteristic corresponding to the demodulation-side pilot code string. This characteristic reflects, for example, the frequency characteristic inherent in the modulation-side pilot code sequence and the frequency characteristic of the window function used when setting the demodulation-side pilot code sequence. In this embodiment, correction data reflecting such characteristics is prepared in advance, and the frequency error is estimated using the correction data. This way,
For example, even if an error due to the frequency characteristic of the modulation-side pilot code string or the window function occurs in the output of the Fourier transform, the frequency error can be estimated with high accuracy.

[0038]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Preferred embodiments of the present invention will be described below in detail. Hereinafter, an example in which the present invention is applied to an OFDM system among multicarrier transmission systems will be described.

A. Principle In the OFDM system according to the present embodiment, for example, when a continuous wave is assumed, the loss of transmission efficiency is made very small by using only one OFDM symbol as a pilot symbol per frame. A code sequence having a high autocorrelation characteristic is employed.

At the time of demodulation, also on the demodulation side, one OFDM symbol obtained from a code sequence obtained by time-inverting the same code sequence as the transmitted pilot symbol is prepared, and the pilot symbol is extracted on the demodulation side. After that, the prepared pilot symbol and the received pilot symbol are multiplied on the time axis. Then, a windowing process is performed on the multiplication result, and an FFT process is performed on the windowing process result.

Next, two frequency components indicating the maximum peak and the second highest peak are extracted from the result of the FFT processing. At this time, correction data is prepared in advance for a side lobe generated by windowing a code sequence used as a pilot symbol for the two pieces of extracted information. Then, a correction based on the correction data is given to the frequency components of the maximum peak and the second peak of the FFT result. As a result, it is possible to expand the frequency error detection range, reduce the processing amount, and improve the frequency error estimation accuracy.

The FFT processing used here is OFDM processing.
It is not necessary to prepare two types of FFT processing units for frequency error detection and demodulation, since the FFT generally used for demodulation can be shared.

If the above configuration is adopted, first,
Since the autocorrelation of pilot symbols on the frequency axis can be realized by multiplication on the time axis, one OFDM for a received signal before subcarrier separation by FFT is performed.
Multiplication processing only by the number of symbol samples is sufficient, and a correlation peak can be obtained at a frequency corresponding to the frequency number error on the frequency axis while the processing amount is relatively suppressed.

Second, according to the above configuration, the gain of the correlation peak can be maximized by employing a code sequence having excellent autocorrelation characteristics for the pilot symbols.

Third, a so-called rectangular window is applied to a sample sequence including only a pilot symbol portion and a side lobe is generated more than necessary. In the above configuration, for example, a windowing called a Hanning window is used. , Unnecessary side lobes in the FFT output can be suppressed. As a result, the accuracy of error detection can be improved.

Fourth, in the above configuration, in the FFT output for the sample sequence subjected to the windowing process, the maximum peak appears at a frequency corresponding to the frequency error, and ± 1 of the FFT is obtained.
The fact that the second peak appears at one of the frequencies separated by the resolution is used to perform the interpolation processing on both, so that even if a frequency error exists between the frequency resolutions, Can be estimated.

Fifth, if a frequency error component is present at a frequency that matches the resolution of the FFT, the above interpolation process increases the error due to the influence of side lobe characteristics possessed by the data used for the FFT. In consideration of the fact that the side lobe characteristics are caused by the side lobe characteristics inherently included in the code sequence used for the pilot symbol and the side lobe characteristics that occur in the windowing process called the Hanning window, both of them are considered. The effect of F
Calculated in advance in the range of the resolution of the FT and prepared in a memory or the like. For this reason, it is possible to appropriately correct the side lobe characteristics generated by the pilot symbol and the windowing process on the FFT output, and it is possible to suppress the occurrence of an estimation error.

According to the OFDM system, the sampling range of the OFDM system is set wider than the range covered by the FFT, and the sampling frequency of the OFDM system is set wider than the range covering all subcarriers. Even if a frequency error exists in a very wide range, it can be estimated with a small amount of processing and with high accuracy. Further, according to the OFDM system, only one pilot symbol is required for one frame, so that a loss in transmission efficiency can be reduced.

The above principle of the OFDM system can be applied to a multicarrier transmission system. Multicarrier modulation in a multicarrier transmission system is
This is means for generating a plurality of subcarrier modulated waves. In this system, the number of subcarriers is K, the subcarrier frequency set for each subcarrier is fk (k is the subcarrier number, 2 ≦ k ≦ K), and all subcarriers are included. When the set sampling frequency is f ′, the amount of phase change per sample (Δθ ′ = 2πfk / f ′) can be determined for each subcarrier. The multicarrier transmission system can be realized by applying the phase change amount Δθ ′ to each subcarrier signal in consideration of the sampling time.

In particular, when applied to OFDM, the subcarrier frequency interval is ΔF = 1 / Ts (Ts is an effective symbol length), and the sampling frequency that can cover all subcarriers is f = M / Ts (M is the subcarrier If the number is N, a power of 2 larger than N), the phase change amount per sample (Δθ = 2πfn / f, where n is the number of the subcarrier and 2 ≦ n ≦ N) is determined. it can.
By using the phase amount change amount Δθ, a method called IFFT (inverse Fourier transform) can be applied as a means for generating subcarrier modulation.

On the other hand, also in the FFT (Fourier transform) in the receiving apparatus described in the present invention, the sampling frequency f 'which can cover all the subcarriers and the modulation or demodulation pilot obtained by the sampling frequency f' This is implemented by using as many point numbers N ′ as possible for all of the code strings. For example, when applying to OFDM,
O is used for generating the subcarrier modulation signal in the above-described transmitting apparatus.
Considering the same as when FDM is applied, and noting that the phase control direction is reversed, the sampling frequency f and the number of points M can be applied to the FFT.

B. Specific Example (1) Transmitting Apparatus FIG. 1 is a diagram showing a configuration of a main part of a transmitting apparatus used in an OFDM system according to an embodiment of the present invention.
The transmitting device shown in FIG. 3 is a configuration example of the transmitting device for the case where the continuous wave signal format shown in FIG. 2 is used. As shown in the figure, in this format, one frame includes a null symbol 50a, a pilot symbol 50b, data symbols 50c-1, 50c-2,
, 50c-n are included in this order, and a predetermined modulation method is applied to each of the data symbol 50c and the pilot symbol 50b. In the transmitting apparatus shown in FIG. 1, a control unit (not shown) considers the modulation scheme of data symbols 50c and pilot symbols 50b and the number of subcarriers, and controls the frame control signal to realize the continuous wave signal format shown in FIG. Is generated and supplied to the switch 1. The switch 1 supplies one of a null signal, a transmission signal, and a pilot signal to the serial-parallel converter 2 based on the frame control signal. It should be noted that the null symbol 50a does not transmit particularly significant data, but is a symbol having only a carrier.

The serial-parallel converter 2 allocates the number of bits per symbol to the in-phase and quadrature components for the data symbol 50c according to a predetermined modulation scheme. Then, the parallel data 52a including the in-phase and quadrature components of each subcarrier to be transmitted is converted to IFF data.
Output to the T unit 3 respectively.

The serial-to-parallel converter 2 supplies 0 (zero) to the orthogonal component of each subcarrier for the pilot symbol 50b, and a code sequence for the pilot symbol 50b prepared in advance to the in-phase component. Are supplied in order from the first code. At this time, if the code sequence is “1” and “0”, “1” is output as it is,
"0" is converted to "-1" and output. In this way, when the pilot symbols 50b are compared, there are two types, 0 (rad) and π (rad), on the complex plane.

The pilot symbols 50b transmitted in the OFDM system include, as known symbols to be used, a code sequence called an M-sequence having excellent autocorrelation characteristics, or a Barker sequence and an M-sequence also having excellent autocorrelation characteristics. Among the code sequences generally called a PN (Pseudo Noise) pattern, such as a code sequence obtained by multiplying the product, a code sequence having excellent autocorrelation characteristics is employed.

The cycle of the code sequence set for pilot symbol 50b is set to be equal to or less than the number of subcarriers to be transmitted, and it is advantageous in terms of the characteristics of the autocorrelation that a code sequence output of at least one cycle is arranged. It is. However, even if an M-sequence shorter than one cycle is used, the gain of the autocorrelation of the pilot symbol 50b only decreases by a certain amount, and if the gain reduction is within an allowable range, This condition does not necessarily have to be satisfied.

When the number of subcarriers to be transmitted is N, it is not always necessary to set pilot symbols 50b for all N subcarriers, and as described above, the period of the code sequence used for pilot symbols 50b is one. What is necessary is just to allocate more than a period. Alternatively, if the reduction of the correlation gain is within an allowable range, a code sequence of less than one cycle may be assigned. In these cases, a number for identifying the subcarrier on which the pilot symbol 50b is arranged may be determined in advance on the receiving side. Further, in the pilot symbol 50b, the order of the code arrangement for each subcarrier is as follows: 1) start from the subcarrier with the lowest carrier frequency, or 2) start from the subcarrier with the highest carrier frequency. And the demodulation side in advance, there is no restriction on how to allocate subcarriers.

As described above, the null symbol 50a,
If the data symbol 50c and the pilot symbol 50b are subjected to serial-parallel conversion, the converted signal is supplied to the IFFT unit 3 as a complex signal 52a including in-phase and quadrature components for each subcarrier. And the IFFT unit 3
Hereinafter, for example, the following commonly used O
An FDM modulation process is performed.

That is, the IFFT section 3 considers the sampling frequency f covering a frequency range wider than the frequency range of all the subcarriers to be transmitted and the subcarrier frequency interval ΔF, and performs phase control according to each subcarrier. Perform processing. Thus, for each subcarrier, the IFFT
The output of the section 3 is divided into in-phase and quadrature components, which are supplied to the parallel-serial conversion section 4 as a complex signal 52b.

In the parallel-serial converter 4, the IFF
The in-phase components of the respective sub-carriers output from the T unit 3 are added for each sampling period while synchronizing all the sub-carriers. Similarly, the orthogonal components of the respective subcarriers output from the IFFT unit 3 are added in synchronization with all the subcarriers for each sampling period.

The two systems of the in-phase and quadrature components to which all the subcarrier components are added, that is, the complex signal 52c, are supplied to the guard signal adding section 5 at the subsequent stage. As shown in FIG. 3, the guard signal adding unit 5 copies a part of the time range Tg having the end of the effective symbol length to the beginning of the effective symbol for each of the in-phase and quadrature components for each effective symbol. Thus, a signal called a guard interval Tg is added. Thus, one OFD obtained by combining the guard interval Tg and the effective symbol length Ts for each in-phase and quadrature component
An M symbol length T is generated. As a result, the output of the guard signal adding unit 5 is 1 OF from the beginning of the guard interval.
Even if only the DM symbol T is extracted, information for one effective symbol can be obtained. As a result, the effect of multipath can be reduced and proper demodulation can be performed.

The complex signal 52d output from the guard signal adding section 5 is passed through a D / A converter (not shown) or a low-pass filter (LPF), and then converted to a predetermined frequency using a quadrature transmitter. Thereafter, necessary frequency conversion is appropriately performed, and the converted signal is transmitted to the receiving side.

(2) Receiver Next, a demodulation process for the OFDM signal modulated as described above will be described. FIG. 4 is a diagram showing a configuration example of a main part of a receiving device used in the OFDM system according to one embodiment of the present invention. In the figure, the received OF
The DM signal is subjected to quadrature detection, and a sampling frequency of M · ΔF (M is a power of 2 larger than the number N of subcarriers, Δ
F is a subcarrier interval, and M · ΔF is a sampling frequency capable of covering all subcarriers. ), The signal is sampled by an A / D (not shown), and is input from the terminal 7 as a reception complex signal 54a. Then, in consideration of the continuous signal format (see FIG. 2), a switching signal indicating a break of each symbol in the frame is generated and supplied to the terminal 8. This signal is, for example, a null symbol 50a.
A frame synchronization signal is generated using the signal power ratio between the signal and the pilot symbol 50b, and a symbol synchronization signal is generated using the guard interval Tg and the complex correlation with the copy source in the same symbol, which corresponds to the logical sum of both signals. It can be generated by generating a signal. Then, based on the switching signal input from the terminal 8, the switch 12,
15 and 19 are respectively switched to the xz side, and a processing operation for frequency error estimation is performed.

More specifically, the signal received by the switch 12 is supplied to the multiplier 20. On the other hand, a complex signal of a pilot symbol prepared in advance on the demodulation side (hereinafter, referred to as “demodulation-side pilot symbol”) is also input to the multiplier 20. This demodulation-side pilot symbol is generated as follows.

That is, first, a code sequence obtained by time-inverting the code sequence for pilot symbol 50b set on the modulation side is input from terminal 10, and pilot symbol 5
In accordance with the order in which 0b is allocated to the subcarriers, the code sequence is allocated with “1” and “0” of the code sequence (however, “0” is converted after being converted to “−1”). , 0 (zero) is assigned to the orthogonal component.
Thereafter, the serial-to-parallel converter 25 converts the parallel signal into the same number of parallel signals as the number of subcarriers, and
a is generated and supplied to the IFFT unit 24.

In the IFFT section 24, the IFFT of the transmitting device
Similarly to the unit 3, the sampling frequency f for the received signal
And phase control is performed according to the subcarrier frequency interval ΔF. Then, a complex signal 56b is supplied to the parallel-serial converter 23 for each subcarrier. Parallel-
In the serial conversion section 23, the reception pilot symbol 50
b while synchronizing with the sampling period f of
The demodulation-side pilot symbol 56c obtained by converting 6b to serial is supplied to the multiplier 20.

The serial-to-parallel converter 25 uses a code sequence obtained by time-inverting the code sequence for pilot symbol 50b. At this time, the demodulation-side pilot symbol 56c is arranged on the frequency axis on the modulation side. Is symmetric with respect to the frequency origin (DC). Then, pilot symbol 5 given to each subcarrier
The codes assigned to the in-phase and quadrature components of 0b are the same in the demodulated pilot symbol 56c in the subcarrier inverted with the DC symmetric.

The multiplier 20 sequentially receives data from the first sample data in the effective symbol section of the received pilot symbol 50b to the last sample data in the effective symbol section (a pair of in-phase and quadrature components). On the other hand, the multiplier 20 sequentially receives data from the first sample data to the last sample data of the demodulation-side pilot symbol 56c output from the parallel-serial converter 23. Then, with the sampling clock phase synchronized with both inputs, one-to-one multiplication is performed for each of the in-phase and quadrature components for all the samples in the effective symbol. Since the product in the time domain corresponds to convolution in the frequency domain, the multiplication in the multiplier 20 can perform processing equivalent to convolution processing in the frequency axis on the time axis.

Thereafter, the result of the multiplication is supplied to a multiplier 21 for multiplication with a window function such as a Hanning window shown in FIG. Multiplier 2
1 multiplies the in-phase and quadrature components output from the multiplier 20 by window function data over the effective symbol length, and supplies the switch 15 with a complex signal 54 b that is the result of the multiplication.

Here, the multiplication of the window function in the multiplier 21 is performed on the output of the multiplier 20, but the multiplication timing is performed anywhere in the output of the parallel-serial conversion unit 23 or the output of the switch 12. May be
Multiplication may be performed within a range synchronized with the effective symbol length of pilot symbol 50b. Further, the multiplier 21
Is multiplied by a constant window function,
The window function generator 22 calculates the constant in advance, stores the constant in a memory, and synchronizes the output of the multiplier 20 with the sampling clock phase to output the same, thereby performing multiplication over the effective symbol length of the pilot symbol 50b. You may.

The switch 15 is set on the xz side until all samples of the windowed effective symbol length output from the multiplier 21 pass. And a multiplier 2
The output of 1 is supplied to the serial-to-parallel converter 16 at the subsequent stage. Regarding the timing of the switching operation in the switch 15, it is calculated in advance how much time is required until all the sample data of the effective symbol length of the symbol output from the multiplier 21 passes. Based on this, a delay of switching from the xz side to the xy side may be set.

Further, in the receiving apparatus shown in the figure, the serial-parallel converter 16 and the serial-parallel converter 16 prepared to use the serial-parallel mutual conversion processing and the FFT processing accompanying the estimation of the frequency error for normal OFDM demodulation. FFT unit 1
7 and the parallel-serial converter 18. Therefore, the circuit configuration can be simplified.

That is, the serial-parallel converter 16
Receives the complex signal 54c from the switch 15, divides it into pairs of in-phase and quadrature components for each sampling frequency f, converts a total of M pairs into parallel signals,
7 The FFT unit 17 is set to execute FFT of M points (M is a power of 2 larger than the number N of subcarriers, and is set so that all subcarriers fall within a target range of FFT). The phase control according to the subcarrier interval ΔF and the sampling frequency f is performed on the M-point complex signal 54d. Then, within the frequency range covered by the FFT unit 17 (−f
/ 2 to + f / 2), the resolution (f / f) of the FFT depends on the sampling frequency and the number of parallel data.
M) Frequency components are calculated for each of the in-phase and quadrature components. Thereafter, the complex signal 54e output from the FFT unit 17 is also supplied to the subsequent parallel-serial conversion unit 18 as a complex signal composed of in-phase and quadrature components. Here, the resolution (f / M) of the FFT unit 17 is f = M
/ Ts, the resolution of the FFT unit 17 is the reciprocal of the effective symbol length Ts, and coincides with the subcarrier frequency interval ΔF.

The parallel-to-serial conversion unit 18
Complex signal 5 given for each frequency resolution ΔF of unit 17
4e are rearranged on the frequency axis in ascending or descending order of frequency, and the output is supplied to the switch 19 as a complex signal 54f. Then, if the signal input to the switch 19 is the timing of the pilot symbol 50b, the switch 19 supplies it to the peak detection unit 26,
If it is any other timing, it is supplied to the terminal 9 for determination.

In the switch 19, similarly to the switch 15, the time until all the sample data within the effective symbol length Ts of the pilot symbol 50b passes is calculated in advance, and all the sample data have passed. The delay may be set so that the switch 19 switches from the xz side to the xy side later. Also, terminal 9
Is supplied with the received data symbol 50c, phase offset correction is performed as necessary, and identification and determination are made in accordance with the modulation method for each subcarrier.

On the other hand, the output of the switch 19 supplied to the peak detector 26 is
This is the autocorrelation result on the frequency axis between b and the demodulation-side pilot symbol 56c. Since the FFT processing has been performed by the FFT unit 17, a peak that is the autocorrelation result exists at a frequency position corresponding to the frequency error. . Further, when an original peak exists at a frequency shifted from the frequency resolution of the FFT unit 17, the peak appears before, after, or at one of the resolution positions. The peak detector 26 uses the sum of the squares of the in-phase component and the quadrature component of the complex signal 6 for the entire output of the FFT unit 17 in accordance with the arrangement order on the frequency axis in the parallel-serial conversion unit 18, for example. In each of the 17 frequency ranges, the frequency is sequentially examined from the minimum frequency to the maximum frequency for each frequency resolution, and the maximum peak is detected. Further, the second peak following the maximum peak is detected again as described above.

At this time, usually, the second peak is adjacent to the maximum peak, and if these two components are used, for example, IEICE Transactions A Vol. J70-A
No. 5 pp. 798-805 ("High-precision frequency determination method using FFT", co-authored by Tabei and Ueda), it is possible to estimate even a frequency error smaller than the resolution of the FFT unit 17.

The frequency estimating method will be described briefly. The frequency showing the maximum peak is f _max , and the adjacent frequency is f _max −
ΔF, f _max + ΔF, and the power at each frequency is b
_{(F ma x), b (} f max -ΔF), b when (f _max + ΔF) to the frequency error Δf to the actual frequency error Δf
e can be estimated by the following equation.

[0079]

(Equation 1) However, in the estimation method shown in the above document,
The frequency error Δf has its resolution ΔF due to the influence of the window function generated by the window function generator 22 and the side lobe characteristic originally held by the code sequence of the pilot symbol 50b.
There is a problem that the estimation accuracy is deteriorated when it is located on the frequency axis for each resolution as compared with the case where it is located in the middle between the two. Therefore, in the present OFDM system, as a countermeasure against this problem, the side lobe correction unit 27 corrects the side lobe. Specifically, first, for example, the window function generated by the window function generation unit 22 includes:

(Equation 2) Is used. In this case, as shown in FIG. 5, the spectrum of the window function itself has side lobes having an amplitude of 1/2 before and after the center main lobe. As a result, by applying this window, the peak on the frequency axis due to the autocorrelation function generates side lobes before and after the center. Also, when a multiplication output of an M sequence and a Barker sequence is used as the code sequence used for the pilot symbol 50b, a side lobe occurs in the autocorrelation output. Therefore, according to the conventional method, the estimation accuracy of the frequency error is deteriorated due to the influence of these side lobes.

On the other hand, OFDM according to the present embodiment
In the system, the window function and the pilot symbol 5 described above are used.
Utilizing the fact that the code sequence of 0b is known, the influence of side lobe characteristics generated from them is calculated in advance to generate correction data. Then, the two outputs of the peak detector 26 are corrected based on the correction data, and the corrected result is supplied to a frequency error estimator 28 at a subsequent stage. More specifically, a term common to the above equations (1) and (2) is modified as follows to provide a correction term β.

[0081]

(Equation 3) The new correction term β in the equation (4) is a term calculated in advance prior to demodulation. For example, it is assumed that the frequency error Δf is half (ΔF / 2) of the resolution ΔF of the FFT unit 17. Then, for an arbitrary frequency error Δf within the range (0 to ΔF / 2), the estimation error (| Δf−Δfe)
|) Is calculated by simulation or the like in advance. Δfe is a frequency error estimated on the demodulation side. In this way, the correction term β is calculated in advance, and it is stored in a memory (not shown) of the side lobe correction unit 27. Each time the output from the peak detection unit 26 is input, it is read and provided for correction. Good.

This correction term β is used in the subsequent frequency error estimating section 2
8 and the frequency error estimator 28 uses the corrected maximum peak and the second peak to calculate the above equation (1) or (2) and the above equation (4) indicating the correction term β. By calculating, the frequency error Δfe can be calculated with high accuracy from the autocorrelation output of the received pilot symbol 50b. The estimation result Δfe of the frequency error obtained by the frequency error estimating unit 28 is thereafter supplied to the NCO 29.

In the NCO 29, the frequency error estimating section 28
Is calculated, and the ratio of the phase control amount per sample is calculated. Then, using a sine table prepared in a memory (not shown) in advance, cos corresponding to the phase control amount for each sample,
Each component of sin is read to generate a complex signal 54g,
It is supplied to the complex multiplier 14. In the complex multiplier 14, the NCO
By complex-multiplying the received data symbol 50c by the complex signal 54g output from 29, a data symbol with a frequency error canceled is generated.

The data symbol is supplied to the FFT unit 17 via the serial-parallel conversion unit 16. Then, the FFT unit 17 can perform subcarrier separation in which the influence of the frequency error is removed. After that, the output of the FFT unit 17 is output from the terminal 9 via the parallel-serial conversion unit 18 and the switch 19, and is provided for the subsequent determination processing.

As for the switching timing of the switches 12, 15, and 19, considering that the delay time is equal to the processing delay required for the frequency estimation described above, the switching timing from the xz side to the xy side is considered. The delay time can be calculated in advance. Therefore, a switching control signal is generated based on the processing delay time, and the switching control signal is supplied from the terminal 8 to the switches 12, 15, and 19, and the switches 12, 1 and 19 are supplied.
What is necessary is just to perform switching control of 5,19. At this time, the delay element 13 provided after the switch 12, which is configured by a shift register, a memory, or the like, may be set in advance so as to delay the processing delay time.

Further, in the configuration example shown in the figure, the output of the parallel-serial converter 18 is input to the peak detector 26 via the switch 19, but the peak detector 26 is
You may connect to each output of T part 17. In this case, a switch that performs the same operation as the switch 19 is provided for each in-phase and quadrature component of each resolution output of the FFT unit 17, and when a received symbol is a pilot symbol 50 b by a switching control signal input from the terminal 8, FFT unit 1
7 may be supplied to the peak detection unit 26, and when the data symbol 50 c is output from the FFT unit 17, the data symbol 50 c may be supplied to the terminal 9.

(3) Modifications Various modifications of the OFDM system described above are possible. For example, the configuration shown in FIG. 6 or FIG. 7 is also possible as another configuration example of the receiving device.

That is, in the receiving apparatus shown in FIG. 4, 1) the received symbol is serial-to-parallel converted to a code sequence obtained by time-inverting the code sequence for pilot symbol 50b used on the modulation side. A signal obtained by sequentially performing IFFT and parallel-serial conversion; 2)
The window function was multiplied by the window function. As shown in FIG. 6, the signal of 1) and the signal of 2) were previously calculated and multiplied by each other. 30 may be stored. As the complex signal 56d multiplied by the window function, for example, a code sequence obtained by time-inverting the code sequence for pilot symbol 50b used on the modulation side is input from terminal 10 in FIG. IFFT unit 24,
It is obtained by executing the parallel-serial converter 23, the window function generator 22, and the multiplier 21 in synchronization with the sampling frequency of the received signal.

In this way, a predetermined value is stored in the memory 30, and the logical sum of the switching control signal input from the terminal 8 and the sampling clock input from the terminal 11 is supplied as a read signal of the memory 30. If the received pilot symbol 50b is multiplied by the read window function multiplied complex signal 56d by the multiplier 20, the complex signal 54b obtained as the output of the multiplier 21 in the configuration shown in FIG. Can be obtained as For this reason, according to this configuration, compared to the configuration shown in FIG.
The processing amount on the demodulation side can be significantly reduced.

In the receiving apparatus shown in FIG. 4, the complex signal 54g output from the frequency error
Is fed back to the received signal using the NCO 29, and the delay element 1
Although the phase control is performed by multiplying the data symbol 50c through 3 by the multiplier 14, the configuration can be further simplified as shown in FIG.

That is, for example, if the local oscillator can control the frequency with the voltage, the frequency error Δf calculated by the frequency error estimating unit 28 can be obtained as shown in FIG.
e is supplied to the voltage setting unit 31, where a necessary correction amount for the current set voltage is calculated, and a corresponding control signal is generated. Then, the control signal is supplied to a local oscillator through a low-pass filter and a D / A converter (not shown),
The correction amount is added or subtracted from the current set voltage value. By doing so, the same phase control as that of the receiving apparatus shown in FIG.
It can be performed with a simpler configuration. According to this configuration example, since the delay element 13 and the multiplier 14 can be reduced, the circuit scale and processing amount on the demodulation side can be reduced.

Further, in the above description, a case where a continuous wave format is assumed has been described. However, the present OFDM system is not limited to a continuous wave but can be similarly applied to a burst wave. That is, burst detection and symbol synchronization can be performed, for example, by performing correlation processing within the same symbol time of the guard interval.
As a result, the timing of the pilot symbol 50b in which only one OFDM symbol is arranged in the burst is obtained. Therefore, the frequency error can be estimated for the pilot symbol 50b by using the configuration described above. It is also clear that the configuration examples shown in FIGS. 6 and 7 can be applied to burst waves.

(4) Estimation of Frequency Error Using Auto-Correlation of Pilot Symbol Here, the principle of estimating the frequency error using the auto-correlation of the pilot symbol 50b on the frequency axis will be further described. FIG. 8 shows, for example, a pilot symbol 50.
This is a description of a case where an M sequence having a code length of 15 bits is used as b. In the figure, the subcarrier frequency interval is Δ
F. When the code output of the M-sequence is “1” and “0”, all of the orthogonal components in the IFFT section 3 as “0 (zero:
No signal) "and the M-phase code output is reflected in the in-phase component. That is," 1 "of the M-sequence code
, The in-phase component is set to “1” and the M-sequence output is set to “0”.
, The in-phase component is “−1”.

In the figure, “1” of the M-sequence code output
"Is expressed in a positive direction (upward in the figure),
In the case of “0”, it is expressed in a negative direction (downward in the figure). When the pilot signal P (f) shown in FIG. 11A is used as the modulation-side pilot signal, the pilot signal Q (f) shown in FIG. 10B is used as the demodulation-side pilot signal. Can be As described above, the pilot signal Q (f) is obtained by transposing the code sequence used on the modulation side from DC to the left and right. In this case, the pilot symbol 5
0b is received with a frequency error Δf, the convolution obtained on the demodulation side is P (f−Δ
f) * Q (f) (however, "*" means convolution)
It is expressed as Then, by performing this operation,
A correlation peak can be obtained at a frequency corresponding to the frequency error Δf, and the frequency error can be preferably calculated by the frequency error estimator 28 and the like. The result of this convolution is shown in FIG.

In the OFDM system, the operation corresponding to the convolution operation is performed on the time axis by utilizing the fact that the convolution operation on the frequency axis corresponds to the multiplication on the time axis. As a result, the circuit configuration can be greatly simplified as compared with the case where the convolution operation is performed on the frequency axis. In addition, in the description of the figure, since the purpose is to explain that a peak corresponding to the frequency error due to the autocorrelation is obtained, the description of the influence by the window function generation unit 22 and its correction is omitted. I have.

FIG. 9 shows a case where a code sequence of a multiplication output of an M sequence and a Barker sequence is used as the pilot symbol 50b, for example. As shown in FIG. 3A, a pilot signal P (f) on the frequency axis is set by multiplying a 7-bit length M sequence by a 3-bit length Barker sequence, as shown in the middle left side of FIG. can do. Note that in FIG. 9 as well as in FIG.
The quadrature component of the T section 3 is "0" (no signal), and "1" or "-1" is input for the in-phase component.

On the other hand, on the demodulation side, similarly to the above embodiment and the case of FIG. 8, a code sequence obtained by time-inverting the pilot symbol 50b, that is, the pilot sequence shown in FIG. A pilot signal Q (f) having the signal P (f) symmetric with respect to DC is prepared. And FIG.
As described in relation to the above, in the OFDM system,
The convolution operation of P (f) and Q (f) is realized by multiplication on the time axis. Therefore, if such an operation is performed, a result as shown in FIG. In this example, side lobes occur due to the use of the Barker sequence. Therefore, in the OFDM system, a frequency error is calculated after correcting the side lobe.

FIG. 11C shows a pilot signal P (f) shown in FIG. 10A and a pilot signal Q shown in FIG.
The result of convolution with (f) is shown. In FIG. 3C, the side lobe frequency position (−14Δ
F-Δf and 14ΔF-Δf) and the size thereof are based on the used M sequence length, Barker sequence length, subcarrier interval ΔF,
And the frequency error Δf.

(5) Waveforms of Respective Parts of Receiver FIG. 10 to FIG. 14 show examples of signal waveforms on the demodulation side of the OFDM system. First,
When an input signal having the waveform shown in FIG. 10 as a waveform having a length of 1 OFDM symbol is input to terminal 7 of the receiving apparatus in FIG. 4, the signal is multiplied by demodulation-side pilot symbol 56c by multiplier 20 of the receiving apparatus. Thus, a signal having a waveform shown in FIG. 11 is obtained.

When the signal shown in FIG. 11 is further output by, for example, a Hanning window from the window function generator 22 and the two are multiplied by the multiplier 21, the output becomes a signal having a waveform shown in FIG. The waveform shown in FIG.
When the processing is performed by the FT unit 17, a peak component appears at the position of the frequency error on the frequency axis as shown in FIG.

FIG. 14 is an enlarged view showing the vicinity of the peak in FIG. According to the figure, it can be seen that, in this example, the frequency error is 2ΔF. By using the maximum peak and the second peak in the figure, the side lobe correction unit 27 and the frequency error estimating unit 28 can calculate the frequency error 2ΔF with high accuracy. In addition, in the figure, when the second peak exists on both sides of the maximum peak, only one of them may be used.

(6) Effect As described above, according to the OFDM system, in the case of a continuous wave, only one OFDM symbol is prepared for a pilot symbol per frame. By preparing only symbols, the processing amount can be reduced, the estimation accuracy can be improved, and the time until pull-in can be accelerated.

In particular, regarding the estimation accuracy, FIGS.
As shown in FIG. 7, the effect of the presence / absence of side lobe correction in the present OFDM system is remarkable. 15 and 16, two examples are applied as pilot symbols. That is, the case where only the M sequence having a sequence length of 511 bits is different from the case where the sequence lengths of the M sequence and the Barker sequence are each 127
Bit and 7 bits.

First, in FIG. 15, the horizontal axis represents the FFT unit 1
7 is plotted with the frequency of the output of No. 7, specifically, a residual Δg obtained by dividing the actual frequency offset Δf by the subcarrier interval ΔF. The vertical axis represents the absolute value of the frequency estimation error normalized by the subcarrier interval ΔF. In FIG. 3, the frequency is normalized by the frequency resolution of the FFT unit 17, that is, the subcarrier interval ΔF, and the display shows only half the resolution, that is, the range of 0 to ΔF / 2. The reason is that the second half of the resolution appears symmetrically with the first half at 0.5. It is apparent from the comparison between the case without correction (broken line) and the case with correction (solid line) in the characteristic example of FIG. 2 that the effect of the side lobe correction in the OFDM system is remarkable.

FIG. 16 shows an example of characteristics when noise is present. In the figure, the horizontal axis is E _b / N ₀ (dB), and the vertical axis is the same as FIG. Also, the used code sequence is the same as in the case of FIG.

In FIG. 16 as well, a comparison between the case without correction (dashed line) and the case with correction (solid line) shows that E _b / N _0.
The effect of correcting the frequency estimation error by the present OFDM system is evident except when the frequency is extremely low. Then, even for the lower part of E _b / N _0, or applied to the system is intended primarily for what range as E _b / N _0, by selecting the code sequence in consideration of the point that easy The problem can be avoided.

[Brief description of the drawings]

FIG. 1 is a diagram showing a configuration on a modulation side of an OFDM system according to an embodiment of the present invention.

FIG. 2 is a diagram illustrating a continuous wave format in the OFDM system according to the embodiment of the present invention.

FIG. 3 is a diagram showing a data structure of one OFDM symbol in the OFDM system according to the embodiment of the present invention.

FIG. 4 is a diagram showing a configuration on a demodulation side of the OFDM system according to the embodiment of the present invention.

FIG. 5 is a diagram showing a spectrum of a Hanning window.

FIG. 6 is a diagram showing a modification of the configuration on the demodulation side of the OFDM system according to the embodiment of the present invention.

FIG. 7 is a diagram showing still another modification of the configuration on the demodulation side of the OFDM system according to the embodiment of the present invention.

FIG. 8 is a diagram illustrating autocorrelation characteristics when only M sequences are applied as pilot symbols.

FIG. 9 is a diagram illustrating autocorrelation when an M sequence and a Barker sequence are applied as pilot symbols.

FIG. 10 is a diagram in which a signal waveform of one symbol length of a received pilot symbol is observed.

FIG. 11 is a diagram showing a signal waveform when autocorrelation processing is performed on the demodulation side in the OFDM system according to the embodiment of the present invention.

12 is a diagram showing a signal waveform obtained by multiplying the signal of FIG. 11 by a window function.

FIG. 13 is a diagram showing a signal waveform obtained by subjecting the signal of FIG. 12 to FFT processing.

14 is an enlarged view showing the vicinity of a peak in FIG.

FIG. 15 is a diagram illustrating an effect of the OFDM system according to the embodiment of the present invention.

FIG. 16 is a diagram illustrating an effect of the OFDM system according to the embodiment of the present invention under noise conditions.

FIG. 17 is a diagram illustrating an example of a subcarrier arrangement in an OFDM system.

FIG. 18 is a diagram illustrating a configuration example on the demodulation side in a conventional OFDM system.

FIG. 19 is a diagram illustrating another configuration example of the demodulation side in the conventional OFDM system.

[Explanation of symbols]

1, 12, 15, 19 switch, 2, 16, 25 serial-parallel converter, 3, 24 IFFT unit, 4,
18, 23 parallel-serial conversion section, 5 guard signal addition section, 7 (for receiving signal input) terminal, 8 (for switching control signal input) terminal, 9 (for output) terminal, 10 (for demodulation side pilot symbol generation) Terminal, 11 (for sampling clock input) terminal, 13 delay element, 14, 3
4 complex multiplier, 20, 21 multiplier, 22 window function generator, 26 peak detector, 27 side lobe corrector, 28 frequency error estimator, 29 NCO, 30 memory, 31 voltage setting unit.

 ────────────────────────────────────────────────── ─── Continuing on the front page (72) Muneo Konishi, Inventor 2-4-1-503, Nishi-Ochiai, Shinjuku-ku, Tokyo (72) Satoshi Miura 5-1-1 Shimorenjaku, Mitaka-shi, Tokyo Japan Radio Co., Ltd. F-term (reference) 5K022 DD01 DD18 DD23 DD33 DD43

Claims (7)

[Claims] 1. A multi-carrier transmission system including a transmitting device and a receiving device, wherein a multi-carrier transmission of a transmission data sequence from the transmitting device to the receiving device is performed. A transmission code sequence generating means for sequentially performing serial-parallel conversion and multi-carrier modulation on a code sequence including a predetermined modulation-side pilot code sequence, and generating a transmission code sequence, wherein the reception device includes the transmission code sequence. Multiplication means for multiplying a portion corresponding to the modulation-side pilot code sequence by a predetermined demodulation-side pilot code sequence corresponding to the modulation-side pilot code sequence, and a Fourier transform for performing a Fourier transform on the multiplication result. Converting means for estimating a frequency error generated between the transmitting device and the receiving device based on a result of the Fourier transform. Multi-carrier transmission system, characterized in that it comprises a frequency error estimation means, a to. 2. The multicarrier transmission system according to claim 1, wherein the transmission data sequence is a complex code sequence, and the transmitting device includes one of an in-phase component and a quadrature component of the modulation-side pilot code sequence. A multicarrier transmission system, wherein 0 is assigned and each bit of a predetermined code string is assigned to the other in order. 3. The multi-carrier transmission system according to claim 1, wherein the transmission code sequence is a sequence of complex numbers, and the receiving device generates a complex number based on a frequency error estimated by the frequency error estimating unit. A multi-carrier transmission system, further comprising frequency error correcting means for reducing the influence of a frequency error generated between the transmitting device and the receiving device by complexly multiplying the transmission code sequence. 4. The multi-carrier transmission system according to claim 1, wherein the receiving device changes an oscillation frequency of a local oscillator included in the receiving device based on a frequency error estimated by the frequency error estimating means. A multicarrier transmission system, further comprising frequency error correction means for controlling and reducing an influence of a frequency error generated between the transmitting device and the receiving device. 5. The multicarrier transmission system according to claim 1, wherein said receiving device further includes a storage unit for storing said demodulation-side pilot code string. . 6. The multi-carrier transmission system according to claim 1, wherein the frequency error estimating means included in the receiving device is a transmitting device based on a result of the Fourier transform and predetermined correction data. A multi-carrier transmission system comprising: estimating a frequency error between the multi-carrier transmission system and a receiving device; 7. A receiving device used in a multicarrier transmission system, comprising: a transmitting code sequence including a predetermined pilot pilot code sequence for a pilot received from a transmitting device; Multiplying means for multiplying a predetermined demodulation-side pilot code string corresponding to the modulation-side pilot code string; Fourier transformation means for performing a Fourier transform on the multiplication result; and And a frequency error estimating means for estimating a frequency error occurring between the receiving device and the receiving device.