WO2016067675A1

WO2016067675A1 - Phase noise compensation receiver

Info

Publication number: WO2016067675A1
Application number: PCT/JP2015/068614
Authority: WO
Inventors: ザカンタン; 純道荒木; 裕淵張; 光平松本
Original assignee: 国立大学法人東京工業大学
Priority date: 2014-10-29
Filing date: 2015-06-29
Publication date: 2016-05-06
Also published as: JP2016092454A

Abstract

[Problem] In order to solve a problem of large amount of signal processing calculation existing in the conventional nonlinear phase noise compensation techniques, provided is a phase noise compensation receiver wherein a linear signal processing is used in the frequency domain, thereby achieving a phase noise compensation technique that uses a smaller calculation amount than the conventional techniques and also achieving a highly reliable high-speed signal transfer. [Solution] A phase noise compensation receiver for receiving orthogonal frequency division multiplexed (OFDM) signals comprises: a phase noise compensation factor estimation circuit that estimates a phase noise compensation factor by use of a known pilot signal included in OFDM symbols and by use of a replica signal; and a phase noise compensation circuit that combines, by use of the phase noise compensation factor outputted from the phase noise compensation factor estimation circuit, received subcarrier signals having been subjected to three-adjacent-channel equalization, thereby compensating for local-oscillator phase noises included in the received signals.

Description

Phase noise compensation receiver

The present invention relates to a phase noise compensation receiver that compensates for phase noise caused by a local oscillator in a receiver in a millimeter wave band wireless communication system employing OFDM.

In recent years, with the spread of transmission and reception of large-capacity data such as high-definition moving pictures and the like and the tightness of the radio frequency of the microwave band, in order to realize a radio communication system using a higher frequency than the conventional microwave band Various efforts are actively being made. For example, IEEE802.11ad is established as a standard for a wireless LAN system using a radio frequency band of 60 GHz, and research on antennas, transmitter / receiver circuits, baseband signal processing, etc. for realizing the wireless LAN system has been conducted all over the world. It is done in

In IEEE802.11ad, the frequency band from 60 GHz to 68 GHz is divided into four channels, and it is considered to use a spatial multiplexing technique called multi-user MIMO (Multiple-Input-Multiple-Output) in each channel. Yes. This spatial multiplexing technology theoretically accommodates 16 users in the entire system and can provide a throughput of 6 Gbps (Gigabit per second) per user.

In recent years, research on wireless communication using terahertz electromagnetic waves having a frequency higher than that of millimeter waves (hereinafter referred to as “terahertz waves”) has been actively conducted. Terahertz wireless communication has not yet reached the stage of standard development, but a group for technical research called IEEE802.15WPANTM Terahertz Interest Group (IGthz) has already been organized within IEEE. There are discussions on terahertz wireless communication.

Thus, in recent years, frequencies used in wireless communication systems tend to move higher. It is the improvement of the high-frequency circuit manufacturing technology based on the Si CMOS integrated circuit technology that technically enables such a high-frequency shift of the wireless communication system. Since CMOS circuits that can be mass-produced can handle frequencies in the terahertz band and millimeter wave band, it is possible to manufacture devices at low cost.

However, a local oscillator in a millimeter wave / terahertz wave transmitter / receiver manufactured by Si CMOS integrated circuit technology has a problem that it has a large phase noise. This large phase noise is caused by the frequency fluctuation of the carrier wave (carrier) output from the local oscillator.

By the way, in recent years, orthogonal frequency division multiplexing (OFDM) is often used as a means for realizing high-speed signal transmission in wireless communication systems. OFDM is a scheme in which a wideband signal is converted into a superposition of narrowband signals called subcarriers and transmitted. OFDM is resistant to multipath effects and intersymbol interference, and can realize high frequency utilization efficiency. Therefore, it is used for LTE (Long Termination Evolution), WiMAX (Worldwide Interoperability for Microwave Access), wireless LAN, etc. It has been.

However, in a wireless communication system employing OFDM, there is a problem that the communication quality deteriorates because the phase noise as described above causes inter-subcarrier interference. For this reason, phase noise compensation technology that reduces the effects of phase noise (inter-subcarrier interference and subcarrier phase rotation caused by phase noise) is required, so studies of OFDM receivers having a phase noise compensation function are actively conducted. Has been done.

By the way, digital signal processing for performing phase noise compensation is roughly divided into those using linear processing and those using nonlinear processing. The former indicates that the input signal is subjected to linear conversion, while the latter indicates that the input signal is subjected to processing other than linear conversion (for example, processing including reproduction modulation and feedback loop).

Most of the conventional phase noise compensation techniques are based on nonlinear processing (hereinafter also simply referred to as “nonlinear phase noise compensation techniques”).

Here, FIG. 1 shows a configuration example (see Non-Patent Document 1) of an OFDM receiver using a conventional nonlinear phase noise compensation technique.

As shown in FIG. 1, an OFDM receiver using a conventional nonlinear phase noise compensation technique includes an antenna 10, an IQ demodulation circuit 11, an A / D conversion circuit 12, a CP removal circuit 13, and a decision pointing. Type phase noise estimator 14, phase noise compensation circuit 15a, channel estimation circuit 15, FFT circuit 16, decision-oriented channel estimator 17, channel equalizer 18, CP inserter 19, and CPE compensation 20, CPE compensation coefficient estimation circuit 21, QAM demodulator 22, OFDM modulation signal generator 23, deinterleaver 24, interleaver 25, error correction decoder 26, error correction encoder 27, , A switch 28 and a CRC decoder 29 are provided.

Among the blocks in FIG. 1, the portions surrounded by broken lines (that is, the decision-directed phase noise estimator 14, the phase noise compensation circuit 15 a, the decision-directed channel estimator 17, the CP inserter 19, the CPE compensator 20, The CPE compensation coefficient estimation circuit 21, the OFDM modulation signal generator 23, the interleaver 25, the error correction encoder 27, the switch 28, and the CRC decoder 29) are related to the conventional nonlinear phase noise compensation technique. Hereinafter, a conventional nonlinear phase noise compensation technique will be described with reference to FIG. The processing procedure in the OFDM receiver using the conventional nonlinear phase noise compensation technique shown in FIG. 1 can be divided into two stages: initial processing and iterative processing.

As shown in FIG. 1, first, a signal received from the antenna 10 is converted into a digital complex baseband signal through an IQ demodulation circuit 11 and an A / D conversion circuit 12. Of the received signal converted into the digital complex baseband signal, channel estimation is performed by the channel estimation circuit 15 using only the preamble. Next, of the received signal converted into the digital complex baseband signal, a cyclic prefix (CP: Cyclic Prefix) removal is performed by the CP removal circuit 13 on a signal that is not a preamble, that is, a data signal. Initial processing is first performed on the digital complex baseband signal from which the cyclic prefix has been removed.

In the initial processing, channel equalization by the channel equalizer 18, CPE compensation by the CPE compensator 20, QAM demodulation by the QAM demodulator 22, deinterleaving by the deinterleaver 24, and error correction decoding by the error correction decoder 26 are performed. Next, the output signal of the error correction decoder 26 is input to a CRC decoder (cyclic redundancy check decoder) 29. Here, when no error is detected by the CRC decoder 29 with respect to the output signal of the error correction decoder 26 input to the CRC decoder 29, the reception processing in the OFDM receiver using the conventional nonlinear phase noise compensation technique Ends. On the other hand, when an error is detected by the CRC decoder 29, the OFDM receiver using the conventional nonlinear phase noise compensation technique shifts to an iterative process.

In the iterative processing, the error correction encoder 27 performs error correction coding on the output signal of the error correction decoder 26 when an error is detected by the CRC decoder 29, interleaving by the interleaver 25, and OFDM modulation signal generator 23. A plurality of processes are performed, such as OFDM modulation by, and CP insertion by the CP inserter 19, and as a result, a transmission signal replica is generated. Next, the decision-directed phase noise estimator 14 estimates phase noise using the generated transmission signal replica and the received signal converted into the digital complex baseband signal, and the estimation result (estimated phase noise). ) And the phase noise is compensated by the phase noise compensation circuit 15a based on the output estimation result. Further, the decision-oriented channel estimator 17 performs channel estimation using the received signal from which the phase noise has been removed and the transmitted signal replica. In the OFDM receiver using the conventional nonlinear phase noise compensation technique shown in FIG. 1, the phase noise can be removed by repeating the above series of processing.

As described above, in the OFDM receiver using the conventional nonlinear phase noise compensation technique in which the phase noise is compensated in the time domain, in order to remove the phase noise, the iterative process is performed. There is a problem that the amount of calculation is enormous, and in particular, in a transmission signal replica generation process using a signal after error correction decoding, fast Fourier transform (FFT: Inverse Fast transform) or inverse fast Fourier transform (IFFT: Inverse Fast) (Fourier Transform) is performed, and there is a problem that the amount of calculation further increases.

As described above, the conventional nonlinear phase noise compensation technique has a drawback that the calculation amount is enormous as a trade-off while showing good phase noise compensation performance. In order to perform high-speed data signal processing of several Gbps class, it is necessary to perform complicated processes such as phase noise processing with low delay in addition to channel estimation and frame processing in the baseband processing in the receiver.

However, when the conventional nonlinear phase noise compensation technology is used, the calculation load of the baseband part of the receiver increases, resulting in an increase in power consumption and a delay in reception processing, resulting in a decrease in effective communication speed. Exists.

In order to efficiently realize millimeter-wave band broadband signal transmission, a phase noise compensation technique with a small amount of calculation and sufficient phase noise compensation performance is indispensable.

The present invention has been made under the circumstances described above, and an object of the present invention is to solve the problem that the amount of calculation associated with signal processing, which exists in the conventional nonlinear phase noise compensation technology, is enormous. Therefore, by using linear signal processing in the frequency domain, a phase noise compensation receiver is realized that realizes phase noise compensation technology with a smaller amount of calculation than before, and can realize high-reliability and high-speed signal transmission. There is to do.

The present invention relates to a phase noise compensation receiver for receiving an orthogonal frequency division multiplexing (OFDM) signal. The above object of the present invention is to estimate a phase noise compensation coefficient using a known pilot signal and replica signal included in an OFDM symbol. Using the phase noise compensation coefficient estimator circuit and the phase noise compensation coefficient output from the phase noise compensation coefficient estimator circuit to synthesize a reception subcarrier signal after equalization of three adjacent channels, And a phase noise compensation circuit for compensating for the local oscillator phase noise included in the circuit.

Further, the object of the present invention is that the replica signal includes the phase noise compensation coefficient, a pilot signal after channel equalization which is a frequency domain signal, and a frequency domain signal adjacent to the pilot signal after channel equalization. When the replica signal is generated based on two received signals after channel equalization, and the two channel equalized received signals that are the frequency domain signals are known or unknown. Alternatively, the phase noise compensation coefficient estimation circuit estimates the phase noise compensation coefficient so that a mean square error between the replica signal and the pilot signal is minimized, or the phase noise compensation coefficient estimation circuit Then, based on the replica signal and the pilot signal, the phase noise compensation coefficient using the MMSE method, the LMS method, or the RLS method By estimating is more effectively achieved.

The phase noise compensation receiver according to the present invention compensates for phase noise using linear signal processing in the frequency domain (ie, the phase of the subcarrier signal due to local oscillator phase noise in the transmitter and receiver). By performing phase noise compensation on the received signal after rotation and interference between subcarriers), phase noise compensation technology has been realized with a much smaller amount of computation than before, so calculation of the baseband part of the receiver In addition to reducing the load, it is possible to solve the reception processing delay problem that occurs when the conventional nonlinear phase noise compensation technique is used.

In the present invention, in order to compensate for the local oscillator phase noise, the frequency domain signal processing in the baseband part of the receiver uses only linear transformation without using iterative processing, thereby enabling phase noise compensation with a small amount of calculation. As a result, it is possible to realize an OFDM receiver with low power consumption and no reception processing delay.

According to the phase noise compensation receiver of the present invention, after the equalization of three adjacent channels using the pilot signal included in the OFDM symbol and the phase noise compensation coefficient estimated based on the “replica signal” of the present invention. Of the received subcarrier signals (ie, by summing up the received subcarrier signals after equalization of three adjacent channels after being weighted by the phase noise compensation coefficient). Since the local oscillator phase noise (phase rotation and inter-carrier interference caused by local oscillator phase noise in the transmitter and receiver) is compensated, transmission performed in the receiver by the conventional nonlinear phase noise compensation technique Since it does not perform repetitive signal processing with signal replica generation, the phase is much smaller than conventional nonlinear phase noise compensation technology. It is possible to realize a sound compensation technology.

It is a block diagram which shows the structural example of the OFDM receiver using the conventional phase noise compensation technique. It is a figure explaining the phase noise compensation circuit 140 and the phase noise compensation coefficient estimation circuit 145 which comprise the principal part of this invention. It is a block diagram which shows the structural example of the OFDM receiver using the phase noise compensation technique of this invention. In this invention, it is a figure which shows the structural example of a signal format. It is a graph which shows the comparison of the number of multiplications between real numbers of a conventional method and this invention. It is a graph which shows the comparison of the bit error rate of a conventional method and this invention. It is a graph which shows the comparison of the packet error rate of a conventional method and this invention.

In order to solve the problem of enormous amount of calculation existing in the conventional nonlinear phase noise compensation technique, the present invention reduces phase noise using linear signal processing (frequency domain signal processing in the receiver baseband unit) in the frequency domain. The present invention relates to a phase noise compensation receiver that realizes the phase noise compensation technique with a much smaller calculation amount than the conventional nonlinear phase noise compensation technique.

In the present invention, the phase noise compensation coefficient of the present invention for compensating for phase noise is learned using a known signal sequence called pilot (hereinafter also simply referred to as “pilot signal”), and the learned phase noise compensation coefficient is learned. Is used to compensate for phase noise caused by the local oscillator included in the received signal (hereinafter also simply referred to as “local oscillator phase noise compensation” or “phase noise compensation”).

In the present invention, in learning of a phase noise compensation coefficient of the present invention (hereinafter also referred to simply as “phase noise compensation coefficient”), which will be described later, the square error between the “replica signal” referred to in the present invention and a known pilot signal. Is used as a norm.

Here, the “replica signal” in the present invention refers to the phase noise compensation coefficient of the present invention, the “pilot signal after channel equalization” which is a frequency domain signal, and the “pilot signal after channel equalization”. It means a signal generated based on two adjacent “received signals after channel equalization” which are frequency domain signals. In the present invention, when generating “replica signals”, two “received signals after channel equalization” adjacent to “pilot signals after channel equalization” may be either known or unknown. . Hereinafter, the “replica signal” referred to in the present invention is simply referred to as “replica signal”.

In the present invention, in the millimeter wave band wireless communication system employing OFDM, attention is paid to the feature of the millimeter wave band channel in which the channel transfer functions of adjacent subcarriers are substantially equal, and it is caused by the local oscillator in the millimeter wave band transceiver. We also paid attention to the fact that the statistic regarding phase noise (hereinafter, also simply referred to as “millimeter-wave transmitter / receiver phase noise”) is almost constant during the delay time of the channel. In addition, the “statistic about phase noise of the millimeter wave band transceiver” referred to in the present invention is expressed by Equation 8 described later.

Means.

In the present invention, a phase noise compensation technique based on linear processing (hereinafter also simply referred to as “linear phase noise compensation technique”) with a small amount of calculation is realized by utilizing such a characteristic unique to the millimeter wave band.

As described above, the focus of the present invention is that “the channel transfer function of adjacent subcarriers is almost equal in a millimeter wave band channel” and “the statistic regarding the phase noise of the millimeter wave band transceiver is the delay time of the channel. It is almost constant between degrees. "

Hereinafter, the focus of the present invention and the main part of the present invention will be described with reference to mathematical expressions. Before explaining the focus of the present invention and the main part of the present invention, first, mathematical expressions relating to signal and digital signal processing in the OFDM system will be described. However, the following mathematical expressions are all expressed in an equivalent low-frequency system.

(1) Relationship between phase noise and OFDM system Transmission subcarrier signal in OFDM transmitter

far. However,

Represents the number of OFDM subcarriers. Further, a signal in the time domain in which the transmission subcarrier signal is OFDM-modulated (hereinafter also simply referred to as “OFDM modulated signal” or “OFDM modulated signal (time domain signal)”).

It expresses.

Also, the phase noise at the transmitter (ie the phase noise due to the local oscillator in the transmitter)

far. However,

Represents the cyclic prefix length.

Similarly, the phase noise at the receiver (ie the phase noise due to the local oscillator in the receiver)

far. In addition, the channel impulse response

It expresses. However,

Represents the number of paths of the channel (propagation path).

The relationship between the transmission subcarrier signal and the OFDM modulation signal can be expressed by the following equation (1).

However,

Represents an OFDM modulated signal (time domain signal) at the k-th sampling time.

Represents a transmission subcarrier signal (frequency domain signal) of the mth subcarrier. Note that j represents an imaginary unit, that is, the square of j is equal to -1.

Next, in the transmitter baseband, a cyclic subcarrier signal (frequency domain signal) for each subcarrier is converted into an OFDM modulated signal (time domain signal) obtained by transforming by inverse fast Fourier transform (IFFT). A prefix (CP) is added.

The OFDM modulated signal (time domain signal) after CP addition can be expressed by the following formula 2.

However,

Represents an OFDM modulated signal (time domain signal) after CP addition at the k-th sampling time.

The OFDM-modulated signal (complex baseband signal) after CP addition represented by Equation 2 is D / A converted by the D / A converter circuit, then modulated by the IQ modulator circuit, and further, a local oscillator in the transmitter Is converted into a millimeter wave signal.

At that time, the millimeter wave signal is affected by the phase noise caused by the local oscillator in the transmitter and causes phase rotation. The millimeter wave signal affected by the phase noise caused by the local oscillator in the transmitter can be expressed by the following equation (3).

However,

Represents the millimeter wave signal at the kth sampling time, which is affected by the phase noise caused by the local oscillator in the transmitter. Also,

Represents the phase noise due to the local oscillator in the transmitter at the k th sampling time.

The millimeter wave signal transmitted from the transmitter propagates through space and reaches the receiving antenna. At that time, like the transmitter, the received signal is affected by phase noise caused by a local oscillator in the receiver. The received signal affected by the phase noise caused by the local oscillator in the receiver can be expressed by the following equation (4).

However, y (k) represents the received signal (time domain signal) at the k-th sampling time, which is affected by the phase noise caused by the local oscillator in the receiver. Also,

Represents the phase noise due to the local oscillator in the receiver at the k th sampling time. Further, n (k) represents noise at the kth sampling time.

Next, the received signal (time domain signal) affected by the phase noise caused by the local oscillator in the receiver expressed by the above equation 4 is demodulated by the IQ demodulator circuit, and then the A / D converter circuit. A / D conversion is performed to convert the signal into a digital complex baseband signal, and the cyclic prefix (CP) is deleted. The received signal after CP deletion (digital complex baseband signal after CP deletion) is expressed by the following formula 5.

However,

Represents a received signal (digital complex baseband signal) after CP deletion of the k-th subcarrier, and is a time domain signal.

The received signal after CP deletion represented by Equation 5 (hereinafter also simply referred to as “OFDM signal after CP deletion”) is converted into a received subcarrier signal (frequency domain signal) by fast Fourier transform (FFT). . The conversion from the OFDM signal after CP deletion to the received subcarrier signal can be expressed by the following equation (6).

However, Y (m) represents the received subcarrier signal of the mth subcarrier and is a frequency domain signal.

Substituting the above formula 1 to the above formula 5 into the above formula 6 gives the relational expression between the received subcarrier signal and the transmitted subcarrier signal as shown in the following formula 7.

However,

Represents the channel transfer function of the m-th subcarrier.

Represents the channel transfer function of the subcarriers. Also,

Represents a transmission subcarrier signal of the subcarriers. N (m) represents the Fourier transform of additive white Gaussian noise (AWGN). Also,

Is expressed by Equation 8 below.

In

Is obtained by replacing 0 with 0, and represents the amount of phase rotation caused by the local oscillator phase noise of the transmitter and receiver of the transmission subcarrier signal of the mth subcarrier. In the absence of phase noise due to the transmitter and receiver local oscillators, no phase rotation occurs,

Is established. This is

It can be confirmed by substituting

The approximate expression expressed by Equation 7 represents the influence of phase noise caused by the local oscillators of the transmitter and the receiver on the OFDM system. That is, the first term on the right side of Equation 7 is the local in the transmitter and the receiver. This represents the phase rotation of the received signal due to the oscillator phase noise, and the second term on the right side of Equation 7 represents the interference from other subcarrier signals. Incidentally, the first term on the right side of Equation 7 is called CPE (Common Phase Error), and the second term on the right side of Equation 7 is called ICI (Inter carrier Interference).

Also, the following formula 8 holds.

However,

Received subcarrier signal before channel equalization of

This is a complex number representing the degree of interference received from the received subcarrier signal before channel equalization of the subcarrier.

The larger the absolute value of,

Received subcarrier signal before channel equalization of

This indicates that the amount of interference given to the received subcarrier signal before channel equalization of the subcarrier is large.

Is 0

Received subcarrier signal before channel equalization of

This indicates that no interference is given to the received subcarrier signal before channel equalization of the subcarrier.

Furthermore, it represents with the said Formula 8.

Is used, an approximate expression represented by the following formula 9 is used.

Here, the approximate expression represented by Equation 9 is a statistic regarding the phase noise of the millimeter wave band transceiver.

Is substantially constant for about the delay time of the channel. However, the term on the right side of Equation 9 is

Received subcarrier signal before channel equalization of

Is a complex number representing the degree of interference received from the received subcarrier signal before channel equalization of the subcarrier (hereinafter, “complex number representing the degree of interference” is also simply referred to as “interference amount”). D represents the delay time of the radio channel. That is, d = 0 corresponds to the first arrival wave, and d = D−1 corresponds to the final arrival wave.

As can be seen from equation (9) above, the term on the left side of equation (9) includes the radio channel delay time d, whereas the term on the right side of equation (9) does not include the radio channel delay time d.

Therefore, based on Equation 9,

Received subcarrier signal before channel equalization of

It can be said that the amount of interference received from the received subcarrier signal before channel equalization of the subcarriers (that is, the statistic regarding the phase noise of the millimeter wave band transceiver) is independent of the channel delay time d. Furthermore, based on Equation 9,

Received subcarrier signal before channel equalization of

It can be said that the amount of interference received from the received subcarrier signal before channel equalization of the subcarriers is independent of the parameters of the radio channel.

(2) Explanation of the point of interest The received subcarrier signal after channel equalization in the receiver baseband can be expressed by the following equation (10).

However,

Represents a received subcarrier signal after channel equalization of the mth subcarrier.

The millimeter wave band channel has a relatively small delay wave. Therefore, the frequency dependence of the channel is reduced, and the channel transfer functions of adjacent subcarriers can be regarded as being substantially equal. That is, the following formula 11 is established.

Where H (m−1) represents the channel transfer function of the (m−1) th subcarrier, H (m) represents the channel transfer function of the mth subcarrier, and H (m + 1) is ( It represents the channel transfer function of the (m + 1) th subcarrier.

Also, ICI is dominant between adjacent subcarriers.

From these facts, the “received subcarrier signal after channel equalization of the m-th subcarrier expressed by the above equation 10”.

Can be expressed by Equation 12 below.

However,

Represents the transmission subcarrier signal (frequency domain signal) of the mth subcarrier,

Represents the transmission subcarrier signal (frequency domain signal) of the (m−1) th subcarrier,

Represents a transmission subcarrier signal (frequency domain signal) of the (m + 1) th subcarrier. Also,

Is expressed by Equation 8 above.

In

Are respectively replaced by 0, 1, and −1, and the amount of phase rotation due to phase noise in the transmitter and receiver of the transmission subcarrier signal of the mth subcarrier, (m−1) th An amount representing the degree of interference that the transmission subcarrier signal of the subcarrier gives to the transmission subcarrier signal of the mth subcarrier, and the transmission subcarrier signal of the (m + 1) th subcarrier is the transmission subcarrier signal of the mth subcarrier Represents the amount of interference given to.

Using the relational expression expressed by the above formula 12, the following becomes clear.

That is, a complex coefficient that does not depend on a certain subcarrier number m

And the following Equation 13 is established.

However,

Represents the received subcarrier signal after channel equalization of the (m−1) th subcarrier,

Represents a received subcarrier signal after channel equalization of the (m + 1) th subcarrier.

In the present invention, complex coefficients

Does not depend on the subcarrier number m, the above complex coefficient is learned as the phase noise compensation coefficient of the present invention by utilizing the pilot signal, and the learned phase noise compensation coefficient and the received subcarrier after channel equalization are used. The focus is on the ability to compensate for local oscillator phase noise contained in the received signal based on the signal.

In other words, the focus of the present invention is
(A1) In the millimeter waveband channel, the channel transfer functions of adjacent subcarriers (that is, three adjacent subcarriers) can be regarded as equal.
(A2) The phase noise can be compensated by synthesizing three adjacent received subcarrier signals after channel equalization in the receiver baseband based on the above Equation 13 (that is, the phase noise in the transmitter and the receiver). The phase rotation and inter-subcarrier interference caused by
(A3) Coefficient for combining received subcarrier signals after equalization of three adjacent channels (that is, phase noise compensation coefficient in Equation 13 above)

Since this does not depend on the subcarrier number, the phase noise compensation coefficient can be learned using the subcarrier.

(3) Description of Essential Part of the Present Invention FIG. 2 is a diagram for explaining the phase noise compensation circuit 140 and the phase noise compensation coefficient estimation circuit 145 that constitute the principal part of the present invention. The main part of the present invention includes a phase noise compensation circuit 140 and a phase noise compensation coefficient estimation circuit 145, which are indicated by a two-dot chain line in FIG.

In the present invention, the phase noise compensation coefficient estimation circuit 145 estimates (learns) the phase noise compensation coefficient using the pilot signal and the replica signal, and outputs the estimated (learned) phase noise compensation coefficient to the phase noise compensation circuit 140. Like to do.

Also, the phase noise compensation coefficient estimation circuit 145 estimates the phase noise compensation coefficient so that the mean square error between the replica signal and the pilot signal is minimized.

Incidentally, in the present invention, when the phase noise compensation coefficient estimation circuit 145 estimates the phase noise compensation coefficient, for example, the MMSE (Minimum Mean Squared Error) method, the LMS (Least Mean Square) method or the RLS (Recursive Least-Squares) method. Known methods such as the method can be used.

Then, the phase noise compensation circuit 140 uses the phase noise compensation coefficient output from the phase noise compensation coefficient estimation circuit 145 and the received subcarrier signal after channel equalization, based on Equation 13 above, By synthesizing the received subcarrier signal after channel equalization, the local oscillator phase noise included in the received signal is compensated, and a compensation result (that is, a received signal subjected to phase noise compensation) is output.

That is, the compensation result (received signal subjected to phase noise compensation) output from the phase noise compensation circuit 140 includes three adjacent received subcarrier signals after channel equalization in the receiver baseband, and a phase noise compensation coefficient estimation circuit. This is a signal obtained by synthesizing the phase noise compensation coefficient estimated at 145 based on the above equation (13).

Hereinafter, in the present invention, a specific example of the phase noise compensation coefficient estimation process for estimating the phase noise compensation coefficient performed by the phase noise compensation coefficient estimation circuit 145 will be described.

First, the procedure of phase noise compensation coefficient estimation processing when the MMSE method is used when the phase noise compensation coefficient estimation circuit 145 estimates the phase noise compensation coefficient will be described.

When the MMSE method is used for the phase noise compensation coefficient estimation process, the phase noise compensation coefficient is calculated based on the following equation (14).

Is estimated.

However, the phase noise compensation coefficient is

It is. Also,

Is the complex coefficient

Is a transpose of the three-dimensional row vector. Also,

Is a vector whose elements are transmission pilot signals, and is expressed by the following equation (15). And below

Is a transpose symbol.

However,

Is a pilot signal included per OFDM symbol. Also,

Represents the subcarrier index of the pilot subcarrier.

Furthermore,

Is a matrix whose elements are a pilot signal after channel equalization and a received signal (frequency domain signal) after channel equalization adjacent to the pilot signal after channel equalization.

Represents the complex conjugate transpose of.

In the present invention, when the phase noise compensation coefficient estimation circuit 145 uses the MMSE method to estimate the phase noise compensation coefficient, a replica signal described later is used.

Is not explicitly generated. However, the phase noise compensation coefficient estimated by the above-described MMSE method (that is, based on Equation 14 above)

And expressed by Equation 17 below

Use the replica signal

And replica signal

It is mathematically possible that the mean square error of the pilot signal is minimized.

Next, the procedure of the phase noise compensation coefficient estimation process when the LMS method is used when the phase noise compensation coefficient estimation circuit 145 estimates the phase noise compensation coefficient will be described.

When the LMS method is used for the phase noise compensation coefficient estimation process, the phase noise compensation coefficient is based on the following LMS algorithm.

Is estimated.
LMS algorithm

However, in the above LMS algorithm,

Represents a replica signal. Also,

Is a vector whose elements are a pilot signal after channel equalization and two channel equalized reception signals (frequency domain signals) adjacent to the pilot signal after channel equalization.

Here, p _i represents the subcarrier number of the pilot signal.

Is a pilot signal after channel equalization. Also,

Are received signals after equalization of two channels adjacent to the pilot signal after channel equalization.

In the LMS algorithm, μ is a real positive constant called a step size parameter, and can be set arbitrarily. Since the convergence speed of the estimated phase noise compensation coefficient changes according to μ, it is necessary to set an appropriate μ.

(4) Examples of the Present Invention Here, specific embodiments of the present invention will be described with reference to the drawings. FIG. 3 is a block diagram showing a configuration example of an OFDM receiver using the phase noise compensation technique of the present invention. The reception operation of the OFDM receiver using the phase noise compensation technique of the present invention shown in FIG. 3 (also simply referred to as “phase noise compensation receiver of the present invention”) will be described below.

As shown in FIG. 3, the OFDM receiver using the phase noise compensation technique of the present invention includes an antenna 100, an IQ demodulation circuit 110, a local oscillator 111, an A / D conversion circuit 115, and a synchronization circuit 117. A CP elimination circuit 120, an FFT circuit 125, a channel equalizer 130, a channel estimation circuit 135, a phase noise compensation circuit 140, a phase noise compensation coefficient estimation circuit 145, a QAM demodulator 150, A leaver 155 and an error correction decoder 160 are provided. Further, the part indicated by the two-dot chain line in FIG. 3, that is, the phase noise compensation circuit 140 and the phase noise compensation coefficient estimation circuit 145 constitute the main part of the present invention.

As shown in FIG. 3, the antenna 100 is adapted to receive a burst OFDM modulated signal that is OFDM modulated on the transmission side. The IQ demodulation circuit 110 refers to the carrier signal input from the local oscillator 111 and analog-complexes a burst OFDM modulated signal (hereinafter also simply referred to as “received OFDM modulated signal”) received via the antenna 100. Convert to baseband signal and output.

Here, since the analog complex baseband signal output from the IQ demodulation circuit 110 is affected by the phase noise caused by the local oscillator 111 in the OFDM receiver, it is a received signal including the phase noise. Hereinafter, the analog complex baseband signal output from the IQ demodulation circuit 110 is also simply referred to as “received signal including phase noise”.

Next, the A / D conversion circuit 115 as sample quantization means performs sample quantization on the analog complex baseband signal output from the IQ demodulation circuit 110.

The synchronization circuit 117 performs timing synchronization processing and carrier frequency synchronization processing of the baseband signal sampled and quantized by the A / D conversion circuit 115 (hereinafter also simply referred to as “digital complex baseband signal”). The digital complex baseband signal is output.

That is, when receiving the “timing / carrier frequency synchronization preamble”, the synchronization circuit 117 receives the digital complex baseband signal after the sample quantization output from the A / D conversion circuit 115, and receives the carrier frequency synchronization and symbol. Timing synchronization is established.

Next, the CP removal circuit 120 deletes the CP from the digital complex baseband signal after the synchronization process output from the synchronization circuit 117, and deletes the CP. Also referred to as “digital complex baseband signal after CP removal”).

The FFT circuit 125 performs fast Fourier transform (FFT) on the digital complex baseband signal after CP deletion output from the CP removal circuit 120, thereby converting the received OFDM modulation signal into a signal for each subcarrier (hereinafter simply referred to as "reception"). Also referred to as “subcarrier signal”.

Next, the channel estimation circuit 135 estimates the channel transfer function using the reception subcarrier signal separated for each subcarrier output from the FFT circuit 125 when the “propagation path preamble” is received. The estimated channel transfer function is output.

Further, the channel equalizer 130 equalizes the received subcarrier signal output from the FFT circuit 125 using the estimated channel transfer function, and outputs the received subcarrier signal after channel equalization.

Next, the phase noise compensation coefficient estimation circuit 145 uses the “received subcarrier signal after channel equalization” and “pilot signal after channel equalization” output from the channel equalizer 130 to generate a phase noise compensation coefficient. And the learned phase noise compensation coefficient is output. The phase noise compensation circuit 140 uses the “phase noise compensation coefficient” output from the phase noise compensation coefficient estimation circuit 145 and the “received subcarrier signal after channel equalization” output from the channel equalizer 130, Compensate the phase noise included in the received signal, and the phase noise compensation result, that is, the received signal with phase noise compensation (hereinafter referred to as “received signal with phase noise compensation”) is simply “received signal after phase noise compensation”. Is also output.).

Next, the QAM demodulator 150 demodulates the phase noise compensation result (received signal after phase noise compensation) output from the phase noise compensation circuit 140 for each subcarrier, and the demodulated signal (demodulated phase noise compensation). The later received signal) is output. The deinterleaver 155 deinterleaves the demodulated signal output from the QAM demodulator 150, and outputs a deinterleaved signal (deinterleaved demodulated signal).

Finally, the error correction decoder 160 performs error correction decoding on the deinterleaved signal output from the deinterleaver 155, and the signal after error correction decoding (the signal after error correction decoded deinterleaving) Is output. This completes the reception operation of the phase noise compensation receiver of the present invention shown in FIG.

FIG. 4 is a diagram showing a configuration example of a format of a transmission OFDM modulation signal (time domain signal) in the present invention. As shown in FIG. 4, the transmission OFDM modulation signal (time domain signal) includes “timing / carrier frequency synchronization preamble”, “propagation preamble (also referred to as channel estimation preamble)” and “data”. It is roughly divided into three parts.

“Timing / carrier frequency synchronization preamble” and “propagation path preamble” are information shared by the OFDM transmitter and the OFDM receiver, and are called known signal sequences.

Also, the “timing / carrier frequency synchronization preamble” has a function for detecting the start timing of the received OFDM modulated signal (time domain signal) in the OFDM receiver. The “propagation path preamble” provides information for estimating a channel (also referred to as “radio channel” or “propagation path”) in the channel estimation circuit 135 of the OFDM receiver.

Furthermore, “data” is composed of a plurality of OFDM symbols. Each OFDM symbol is composed of a number of signals per subcarrier equal to the number of FFT / IFFT points. Among the signals for each subcarrier in the OFDM symbol, some subcarrier signals are known signals in transmission and reception, and are called pilot signals to distinguish them from preambles. As described above, this pilot signal provides information for estimating the phase noise compensation coefficient in the present invention.

(5) Comparison between the present invention and the conventional method In order to demonstrate the superior effect of the present invention, the phase noise compensation technique of the present invention and the conventional phase noise compensation technique disclosed in Non-Patent Document 1 (hereinafter simply referred to as “conventional”). The results are also shown in FIGS. 5 to 7. FIG.

Incidentally, FIG. 5 is a graph showing a comparison of the calculation amount between the conventional method and the present invention. In FIG. 5, the number of multiplications between real numbers is used as the calculation amount. FIG. 6 is a graph showing the result of comparing the bit error rate of the conventional method and the present invention by numerical simulation. Further, FIG. 7 is a graph showing the result of comparing the packet error rate of the conventional method and the present invention by numerical calculation simulation. The conditions for numerical calculation simulation are shown in Table 1 below.

By using the phase noise compensation technique of the present invention, in the baseband part of the phase noise compensation receiver according to the present invention, the amount of calculation spent for phase noise compensation (in FIG. 5, the number of multiplications between real numbers) is 30 minutes of the conventional method. It can be significantly reduced to about 1.

Specifically, for the present invention and the conventional method, the number of multiplications between real numbers related to phase noise compensation was obtained, and the obtained number of multiplications between real numbers was as shown in FIG. In calculating the amount of calculation related to the phase noise compensation technique of the present invention (the number of multiplications between real numbers in FIG. 5), as a specific example, the case where the MMSE method or the LMS method is used in the phase noise compensation coefficient estimation circuit 145 was verified. . When calculating the number of multiplications between real numbers, one multiplication between complex numbers is calculated as four multiplications between real numbers. The larger the value on the vertical axis of the graph of FIG. 5, the greater the amount of calculation and the longer the calculation time. As shown in FIG. 5, in the conventional method, the number of multiplications between real numbers related to phase noise compensation is 211968, whereas in the present invention, the phase noise compensation coefficient estimation circuit 145 uses the MMSE method. In addition, the number of multiplications between real numbers for phase noise compensation is 7488. Although not shown, in the present invention, when the LMS method is used in the phase noise compensation coefficient estimation circuit 145, the number of multiplications between real numbers related to phase noise compensation is 6624 times.

As described above, in the present invention, even when either the LMS method or the MMSE method is used, the calculation amount related to the phase noise compensation technique of the present invention (the calculation amount related to the phase noise compensation coefficient estimation circuit 145) is the same as the conventional method. Is about 1/30 of the calculation amount. From this comparison result, it is clear that the amount of calculation spent for phase noise compensation can be greatly reduced by using the phase noise compensation technique of the present invention.

In FIG. 6, the bit error rate characteristics in the conventional method and the present invention are compared. The horizontal axis in FIG. 6 is the signal-to-noise ratio (SNR), and the vertical axis in FIG. 6 is the bit error rate (BER). The smaller the value of BER, the better. From the simulation results in FIG. 6, it can be seen that the present invention requires a received power that is about 2 dB (@ BER = 10 ⁻⁴ ) larger than the conventional method using nonlinear processing in order to obtain the same bit error rate.

Further, “no compensation” in FIG. 6 represents a bit error rate characteristic when phase noise compensation is not performed in the receiver. The present invention and the conventional approach show a lower bit error rate at any signal to noise ratio (SNR) than without compensation. In other words, the present invention and the conventional method have succeeded in improving the communication quality deterioration caused by the phase noise.

FIG. 7 compares the packet error rate (PER) characteristics of the conventional method and the present invention. The horizontal axis in FIG. 7 is the signal-to-noise ratio (SNR), and the vertical axis in FIG. 7 is the packet error rate. The smaller the value of PER, the better.

Further, “no compensation” in FIG. 7 represents a packet error rate characteristic when the phase noise compensation is not performed in the receiver. The present invention and the prior art show a lower packet error rate at any signal-to-noise ratio (SNR) than without compensation. In other words, the present invention and the conventional method have succeeded in improving the communication quality deterioration caused by the phase noise.

From the simulation results of FIG. 7, it can be seen that the present invention requires a received power that is about ² dB (@ PER = 10 ⁻² ) larger than the conventional method using nonlinear processing in order to obtain the same packet error rate.

The embodiment of the present invention has been described above, but the specific configuration of the phase noise compensation receiver of the present invention is not limited to the above-described embodiment, and the gist of the present invention (the focus of the present invention) Various modifications can be made without departing from the scope of the present invention.

10,100 Antenna 11,110 IQ demodulation circuit 12,115 A / D conversion circuit 13,120 CP removal circuit 14 Decision-directed phase noise estimator 15,135 Channel estimation circuit 15a, 140 Phase noise compensation circuit 16,125 FFT circuit 17 Decision-oriented

channel estimator

18, 130 Channel equalizer 19 CP inserter 20 CPE compensator 21 CPE compensation

coefficient estimation circuit

22, 150 QAM demodulator 23 OFDM

modulation signal generator

24, 155 Deinterleaver 25

Interleaver

26 , 160 error correction decoder 27 error correction encoder 28 switch 29 CRC decoder 111 local oscillator 117 synchronization circuit 145 phase noise compensation coefficient estimation circuit

Claims

A phase noise compensation receiver for receiving an orthogonal frequency division multiplexing (OFDM) signal, comprising:
A phase noise compensation coefficient estimation circuit that estimates a phase noise compensation coefficient using a known pilot signal and replica signal included in the OFDM symbol;
By using the phase noise compensation coefficient output from the phase noise compensation coefficient estimation circuit and synthesizing adjacent reception subcarrier signals after three channel equalization, local oscillator phase noise included in the reception signal is obtained. A phase noise compensation circuit for compensating
A phase noise compensation receiver comprising:
The replica signal includes the phase noise compensation coefficient, a pilot signal after channel equalization which is a frequency domain signal, and two channel equalization signals which are adjacent to the pilot signal after channel equalization and which are frequency domain signals. Generated based on the received signal,
2. The phase noise compensation receiver according to claim 1, wherein, when generating the replica signal, the two channel equalized reception signals which are the frequency domain signals are known or unknown.
3. The phase noise compensation receiver according to claim 1, wherein the phase noise compensation coefficient estimation circuit estimates the phase noise compensation coefficient so that a mean square error between the replica signal and the pilot signal is minimized.
3. The phase noise compensation coefficient according to claim 1, wherein the phase noise compensation coefficient estimation circuit estimates the phase noise compensation coefficient using an MMSE method, an LMS method, or an RLS method based on the replica signal and the pilot signal. Receiving machine.