JP5143533B2 - Reception device and signal processing method - Google Patents

Description

  The present invention relates to a receiving apparatus and a signal processing method. In particular, the present invention relates to a MIMO (Multiple-Input and Multiple-Output) receiving apparatus and a signal processing method in a mobile communication system of a single carrier transmission system.

  As one technique for increasing the communication speed between wireless devices, a multi-input / multi-output transmission system is known. This method is literally based on signal input / output using a plurality of antennas. The feature of this method is that a plurality of transmission data can be transmitted at the same time and at the same frequency using a plurality of different antennas. Therefore, as the number of channels that can be transmitted simultaneously increases, the amount of information that can be transmitted per unit time can be increased by the increased number of channels. Further, this method has an advantage that the occupied frequency band does not increase when the communication speed is improved.

  However, since a plurality of modulated signals having carrier components of the same frequency are transmitted at the same time, a means for separating the modulated signals (multiplexed signals) that interfere with each other on the receiving side is required. Therefore, on the receiving side, a channel matrix representing the transmission characteristics of the wireless transmission path is estimated, and a transmission signal corresponding to each substream is separated from the received signal based on the channel matrix. The channel matrix is estimated using pilot symbols or the like.

  However, a special contrivance is required to sufficiently remove the influence of noise added in the transmission path and interference generated between substreams and accurately reproduce the transmission signal for each substream. In this regard, in recent years, various techniques related to MIMO signal detection have been developed. For example, a MMSE (Minimum Mean Squared Error) detection method and an MLD (Maximum Likelihood Detection) detection method capable of improving transmission characteristics as compared with this MMSE detection method are known. There is also known an improved method called QR decomposition MLD detection method (hereinafter referred to as QRD-MLD method) that can reduce the amount of calculation required for signal detection for each substream as compared with the MLD detection method. Further, the following Non-Patent Document 1 discloses an example of a QRD-MLD system in which multipath interference countermeasures are taken.

  Recently, as a technique related to single carrier transmission, a technique called a DFT-SOFDM (Discrete Fourier Transform Orthogonal Frequency Division Multiple Access) method has attracted attention. Also, standardization of this method is currently being studied by 3GPP (3rd Generation Partnership Project).

  This method uses a transmission-side technique in which a transmission signal is modulated and mapped in the frequency domain, and the modulated signal in the frequency domain is inversely transformed into the time domain and then transmitted, and two-dimensional (spatial domain + frequency domain) MMSE detection This is a combination with the receiving side technology for restoring the transmission signal. When this technique is applied, the transmission signal is modulated and mapped in the frequency domain, so that the power consumption (instantaneous maximum power) of the transmission apparatus can be suppressed. In the case of such a single carrier transmission system, the peak power can be suppressed lower than that of a multicarrier transmission system such as OFDM (Orthogonal Frequency Division Multiple Access), so that it can be applied to a portable terminal or the like that demands power consumption is severe. Is expected.

Maeda et al. , “QRM-MLD Combined with MMSE-Based Multipath Interference Canceller for MIMO Multiplexing in Broadband DS-CDMA”, IEEE PIMRC, 2004

  However, when a technique that applies the MMSE detection method such as the DFT-SOFDM method is applied to the MIMO system, there is a problem that diversity gain between antennas obtained when the MLD detection method or the like is applied cannot be obtained. This problem means that relatively high transmission power is required to achieve the signal power to noise power ratio (SNR) required to obtain the desired signal quality. That is, when trying to obtain high signal quality, even if the technology is applied, a large power burden is imposed on the portable terminal or the like.

  On the other hand, when a technique such as the MLD detection method is applied, transmission power required to obtain high signal quality can be suppressed by the amount that diversity gain between antennas is obtained. However, since the MLD detection method requires a large amount of calculation for signal detection for each substream and is not realistic, a QRD-MLD method with a relatively small amount of calculation is generally used. However, although diversity gain can be obtained by using the QRD-MLD method or the like, it is difficult to sufficiently equalize the influence of multipath interference (hereinafter referred to as MPI) generated independently for each element of the channel matrix. There is a problem that the signal quality is greatly deteriorated in an environment that is easily affected by path interference.

  Further, in Non-Patent Document 1 described above, a DS-CDMA (Direct Sequence Code Division Multiple Access) method is applied and a decrease in communication speed due to DS (Direct Sequence) processing is compensated by using a multiplexing method called multicode transmission. A transmitting-side technique and a receiving-side technique for performing auxiliary signal separation processing using an interference canceller. This technique suppresses the influence of multipath interference by handling multipaths as independent paths using the sharp autocorrelation characteristic of the DS-CDMA system. However, since this method is premised on the use of multi-code transmission, there is a problem in that an independent signal addition process in the time domain is required and a large peak power is generated. Therefore, it is not preferable to apply this technique to a device such as a portable terminal in which the maximum transmission power is limited. In addition, this technique has a problem that it cannot be applied to an environment such as a multiuser MIMO system in which arrival times of waves arriving via each path are not constant.

  To solve this problem, frequency domain equalization (FDE) is applied to the received signals corresponding to each row in order from the bottom row of the upper triangular matrix obtained by QR decomposition of the channel matrix based on the QRD-MLD scheme. A system for equalizing frequency selectivity (hereinafter, QRDE-MLD system) has been devised by applying a domain equalizer. However, if an equalization coefficient is applied to an arbitrary transfer function included in the upper triangular matrix for frequency domain equalization, the frequency selectivity of other transfer functions is emphasized due to the characteristics of QR decomposition. is there. Therefore, even when the QRDE-MLD method is applied, there is a problem that sufficient frequency domain equalization cannot be realized in the case of broadband communication in an environment where multipath interference is relatively large.

  Therefore, the present invention has been made in view of the above problems, and an object of the present invention is to perform a multi-signal separation process based on maximum likelihood detection even in an environment where multipath interference is relatively large. It is an object of the present invention to provide a new and improved receiving apparatus and signal processing method capable of sufficiently equalizing frequency selectivity.

  In order to solve the above problems, according to an aspect of the present invention, a receiving apparatus is provided that separates multiple signals received by a plurality of antennas by maximum likelihood detection. The reception apparatus decomposes an interference component subtraction unit that subtracts a frequency-dependent component of a channel matrix using a replica signal, and the channel matrix from which the frequency-dependent component has been subtracted into a product of a unitary matrix and an upper or lower triangular matrix. A matrix decomposition unit, and a signal equalization unit that equalizes frequency selectivity for each received substream by performing frequency domain equalization in units of rows of the upper or lower triangular matrix based on the upper or lower triangular matrix It is characterized by providing.

  In addition, a replica generation unit that generates the replica signal based on each substream component of the multiplexed signal separated using the MMSE equalization coefficient may be further provided.

  The replica generation unit selects whether to use the MMSE equalization coefficient or the unitary matrix and the upper or lower triangular matrix according to a transmission path condition when separating the multiplexed signal. May be.

  In order to solve the above problem, according to another aspect of the present invention, it is possible to suppress the influence of multipath interference when multiple signals received by a plurality of antennas are separated by maximum likelihood detection. A simple signal processing method is provided. The signal processing method includes an interference component subtraction step in which a frequency-dependent component of a channel matrix is subtracted using a replica signal, and the channel matrix from which the frequency-dependent component is subtracted is a product of a unitary matrix and an upper or lower triangular matrix. A matrix decomposition step to be decomposed and a signal whose frequency selectivity is equalized for each received substream by performing frequency domain equalization in units of rows of the upper or lower triangular matrix based on the upper or lower triangular matrix And a step of converting.

  By applying the above apparatus or method, a relatively large multipath interference component that cannot be completely removed by the frequency domain equalizer for each substream is subtracted in advance, so that the transmission path is highly susceptible to multipath interference. Even in this state, it is possible to effectively equalize the frequency selectivity. In addition, the replica signal is generated based on the signal separation method using the MMSE equalization coefficient at the first time of the multipath interference canceling, so that it is possible to prevent the deterioration of the decoding characteristic due to the generation error of the replica signal. it can. However, since the signal separation method based on the maximum likelihood detection is performed after the second time, a diversity gain between the antennas can be obtained.

  As described above, according to the present invention, it is possible to sufficiently equalize the frequency selectivity even in an environment where the multipath interference is relatively large in the separation process of the multiplexed signal based on the maximum likelihood detection.

  Exemplary embodiments of the present invention will be described below in detail with reference to the accompanying drawings. In addition, in this specification and drawing, about the component which has the substantially same function structure, duplication description is abbreviate | omitted by attaching | subjecting the same code | symbol.

[About QRDE-MLD]
First, prior to describing a preferred embodiment of the present invention, the QRDE-MLD system that is the basis of the technology according to the present invention will be briefly described. This method is a maximum likelihood detection technique capable of equalizing frequency selectivity independently included in each element of a channel matrix in a multiple signal demultiplexing means in a single carrier transmission MIMO system.

(Functional configuration of receiving apparatus 100)
First, the functional configuration of the receiving apparatus 100 according to the QRDE-MLD scheme will be described with reference to FIG. FIG. 1 is an explanatory diagram illustrating a functional configuration of a receiving apparatus 100 according to the QRDE-MLD scheme. Although the transmission side device is not specified, for example, for each substream, means for error correction coding / error detection coding, means for discrete Fourier transform, means for modulation mapping in the frequency domain, and modulation And a means for performing inverse Fourier transform on the signal in the time domain, and having a functional configuration that transmits the signal via a plurality of antennas.

  As illustrated in FIG. 1, the receiving apparatus 100 mainly includes an FFT unit 102, a multiplication processing unit 104, a signal equalization unit 106, a channel estimation unit 108, and a QR decomposition unit 110, in addition to a plurality of antennas. , Equalization coefficient calculation sections 112 and 114, multiplication coefficient calculation sections 116 and 118, IDFT section 120, EDC section 122, LLR section 124, S / P conversion section 126, and FEC section 128. The

(FFT unit 102, multiplication processing unit 104)
The FFT unit 102 is means for converting a time domain received signal y (t) input via each antenna into a frequency domain received signal y (ω). At this time, the FFT unit 102 converts the received signal into a frequency domain signal based on an algorithm of Fast Fourier Transform (Fast Fourier Transform). The received signal converted into the frequency domain by the FFT unit 102 is input to the multiplication processing unit 104. The received signal vector y is expressed by the following equation (1) using the channel matrix H. Where x is a transmission signal vector. N is a noise term.

The multiplication processing unit 104 is a means for calculating the product of the Hermite conjugate Q ^H of the unitary matrix Q input from the QR decomposition unit 110 described later and the received signal vector y input from the FFT unit 102. The product Q ^H y calculated by the multiplication processing unit 104 is input to the signal equalization unit 106. In the following description, each element of the product Q ^H y is expressed as {Q ^H y} (k) (k = 0 to 3; corresponding to the (k + 1) th row from the bottom of the upper triangular matrix R). Further, the product Q ^H y is expressed as in the following formula (2).

(Signal equalization unit 106)
The signal equalization unit 106 converts the output signal of the multiplication processing unit 104, equalization coefficients calculated by equalization coefficient calculation units 112 and 114, which will be described later, and multiplication coefficients calculated by the multiplication coefficient calculation units 116 and 118. This is means for equalizing the frequency selectivity of each element of the upper triangular matrix R calculated by the QR decomposition unit 110 described later, and further flattening the frequency selectivity. An output signal Z of the signal equalization unit 106 is input to the IDFT unit 120. The details of the functional configuration of the signal equalizer 106 will be described later.

(Channel estimation unit 108, QR decomposition unit 110)
Channel estimation section 108 is means for estimating channel matrix H indicating the state of the signal transmission path based on pilot symbols and the like. The channel matrix H estimated by the channel estimation unit 108 is input to the QR decomposition unit 110. The QR decomposition unit 110 is a unit that decomposes the channel matrix H estimated by the channel estimation unit 108 into a product of the unitary matrix Q and the upper triangular matrix R (so-called QR decomposition).

  The unitary matrix Q calculated by the QR decomposition unit 110 is input to the multiplication processing unit 104. In addition, the upper triangular matrix R calculated by the QR decomposition unit 110 is input to equalization coefficient calculation units 112 and 114, multiplication coefficient calculation units 116 and 118, and an EDC unit 122. Throughout this paper, the explanation will be made on the premise of QR decomposition. However, a modification using QR decomposition that decomposes into a product of a unitary matrix Q and a lower triangular matrix L instead of QR decomposition can be applied to the technical aspects of this embodiment. It goes without saying that it belongs to a range.

(Equalization coefficient calculation units 112 and 114)
The equalization coefficient calculators 112 and 114 are means for calculating equalization coefficients (Fa, Fs) for realizing frequency domain equalization for each reception substream. In particular, the equalization coefficient calculation unit 112 mainly calculates an equalization coefficient (hereinafter referred to as an addition equalization coefficient Fa) for frequency domain equalization with respect to a signal related to an addition multiplication coefficient Ca described later. To do. On the other hand, the equalization coefficient calculation unit 114 mainly calculates an equalization coefficient (hereinafter, subtraction equalization coefficient Fs) for equalizing a frequency domain with respect to a signal related to a subtraction multiplication coefficient Cs described later. To do. However, the output signal of the multiplication processing unit 104 corresponding to the lowermost row of the upper triangular matrix R does not need to consider the relationship with the multiplication coefficients Ca and Cs as will be described later.

In the following description, the equalization coefficient Fa for addition and the equalization coefficient Fs for subtraction corresponding to {Q ^H y} (k) are expressed as Fa (k) and Fs (k), respectively. Further, the equalization coefficient Fa (k) for addition and the equalization coefficient Fs (k) for subtraction are expressed as in the following expressions (3) and (4). However, N is the number of receiving antennas. The superscript bar represents the ensemble average. S represents signal power.

(Multiplication coefficient calculation units 116 and 118)
Multiplication coefficient calculation sections 116 and 118 are components corresponding to non-diagonal elements of upper triangular matrix R for each output signal of multiplication processing section 104 in the previous stage of performing frequency domain equalization processing for each received substream. Is a means for calculating multiplication coefficients (Ca, Cs) for removing or injecting. In particular, the multiplication coefficient calculation unit 116 adds a multiplication coefficient (hereinafter referred to as an addition multiplication coefficient Ca) for injecting a component corresponding to a non-diagonal element of the upper triangular matrix R to each output signal of the multiplication processing unit 104. calculate. On the other hand, the multiplication coefficient calculation unit 118 calculates a multiplication coefficient for removing components corresponding to the off-diagonal elements of the upper triangular matrix R (hereinafter referred to as a multiplication coefficient for subtraction Cs) for each output signal of the multiplication processing unit 104. calculate. Note that, as will be described later, the addition multiplication coefficient Ca calculated by the multiplication coefficient calculation unit 116 indicates the frequency selectivity of the signal corresponding to the off-diagonal element of the upper triangular matrix R and the frequency of the signal corresponding to the diagonal element. This has the effect of approximating the selectivity.

In the following description, the multiplication coefficient Ca for addition and the multiplication coefficient Cs for subtraction corresponding to the non-diagonal elements R _ij (i ≠ j; i, j = 0 to 3) of the upper triangular matrix R are respectively represented by Ca _ij and Cs _ij. Is written. Further, the multiplication coefficient for addition Ca _ij and the multiplication coefficient for subtraction Cs _ij are expressed as the following expressions (5) and (6).

(IDFT part 120)
The IDFT unit 120 performs an inverse discrete Fourier transform (IDFT) on each output signal Z of the signal equalization unit 106 to perform a time domain signal Z (t) from the frequency domain signal Z (ω). It is a means to convert into. The output signal of each IDFT unit 120 is input to the EDC unit 122.

(EDC unit 122)
The EDC unit 122 calculates an Euclidean distance between the output signal of the signal equalization unit 106 converted into a time domain signal by the IDFT unit 120 and a transmission symbol candidate corresponding to a predetermined modulation scheme. Means. However, a transmission symbol candidate is each signal point on a signal point arrangement diagram (constellation) corresponding to a predetermined modulation method. The Euclidean distance is calculated by an evaluation function expressed by the following equation (7).

但し、ｘｓは送信シンボル候補を示す。

Here, xs indicates a transmission symbol candidate.

  Note that the EDC unit 122 in the drawing has each stage described so as to correspond to each row of the upper triangular matrix R for convenience of explanation. For example, the EDC unit 122 and the upper triangular matrix R described in the lowermost stage Corresponds to the bottom line. Therefore, the EDC unit 122 calculates the Euclidean distance for each substream in order from the lowest level, and selects a combination of transmission symbol candidates (hereinafter, estimated transmission signal vector) that minimizes the likelihood corresponding to all substreams. The estimated transmission signal vector selected by the EDC unit 122 is input to the LLR unit 124.

(LLR unit 124, S / P conversion unit 126, FEC unit 128)
The LLR unit 124 is a means for calculating a log likelihood ratio (LLR) of the estimated transmission signal vector input from the EDC unit 122. The S / P converter 126 is means for separating the estimated transmission signal vector for each substream. The estimated transmission signal for each substream separated by the S / P conversion unit 126 is input to the FEC unit 128. The FEC unit 128 performs error correction (Forward Error Correction) on the estimated transmission signal for each substream based on the log likelihood ratio calculated by the LLR unit 124, and outputs reproduced data.

  The overview of the overall functional configuration of the QRDE-MLD receiver 100 has been described above. However, the feature of the QRDE-MLD method is mainly in the functional configuration of the signal equalization unit 106 whose details are not described above. Therefore, a more detailed description of the functional configuration of the signal equalization unit 106 is added.

(Detailed functional configuration of signal equalization unit 106)
The functional configuration of the signal equalization unit 106 will be described in more detail with reference to FIG. FIG. 2 is an explanatory diagram illustrating a functional configuration of the signal equalization unit 106 according to the QRDE-MLD scheme.

  As shown in FIG. 2, the signal equalization unit 106 according to the QRDE-MLD system mainly includes FDE units 172, 176, 180, and 184 and addition / subtraction units 174, 178, and 182.

(First stage processing)
First, the signal equalization unit 106 branches the signal {Q ^H y} (3) input from the multiplication processing unit 104 into two and inputs the signal to the FDE unit 172. The FDE unit 172 multiplies one of the input signals by the addition equalization coefficient Fa (3), and outputs the multiplication result (Z (3)) to the IDFT unit 120. Further, the FDE unit 172 multiplies the other of the input signals by a subtraction equalization coefficient Fs (3). The signal multiplied by the subtraction equalization coefficient Fs (3) is branched and input to the addition / subtraction units 174, 178, and 182.

(Second stage processing)
Next, the signal equalization unit 106 branches the signal {Q ^H y} (2) input from the multiplication processing unit 104 into two and inputs the signal to the addition / subtraction unit 174. The addition / subtraction unit 174 multiplies the signal input from the FDE unit 172 by an addition multiplication coefficient Ca ₂₃ . Then, the addition / subtraction unit 174 adds the signal obtained by the multiplication to the signal {Q ^H y} (2) to generate an addition signal. Similarly, the addition / subtraction unit 174 multiplies the signal input from the FDE unit 172 by a subtraction multiplication coefficient Cs ₂₃ . Then, the addition / subtraction unit 174 subtracts the signal obtained by the multiplication from the signal {Q ^H y} (2) to generate a subtraction signal. The addition signal and the subtraction signal are input to the FDE unit 176.

  The FDE unit 176 multiplies the addition signal input from the addition / subtraction unit 174 by the equalization coefficient Fa (2) for addition, and outputs the multiplication result (Z (2)) to the IDFT unit 120. Further, the FDE unit 176 multiplies the subtraction signal input from the addition / subtraction unit 174 by a subtraction equalization coefficient Fs (2). The signal multiplied by the subtraction equalization coefficient Fs (2) is branched and input to the adder / subtractors 178 and 182.

(3rd step)
Next, the signal equalization unit 106 branches the signal {Q ^H y} (1) input from the multiplication processing unit 104 into two and inputs the signal to the addition / subtraction unit 178. Subtraction unit 178 multiplies the addition for the multiplication coefficient _{Ca 13} to the signal input from FDE section 172. Then, the addition / subtraction unit 178 adds the signal obtained by the multiplication to the signal {Q ^H y} (1) to generate a first addition signal. Similarly, the addition / subtraction unit 178 multiplies the signal input from the FDE unit 172 by a subtraction multiplication coefficient Cs ₁₃ . Then, the addition / subtraction unit 178 generates a first subtraction signal by subtracting the signal obtained by the multiplication from the signal {Q ^H y} (1).

Furthermore, subtraction unit 178 multiplies the addition for the multiplication coefficient _{Ca 12} to the signal input from FDE section 176. Then, the addition / subtraction unit 178 adds the signal obtained by the multiplication to the first addition signal to generate a second addition signal. Similarly, the addition / subtraction unit 178 multiplies the signal input from the FDE unit 176 by a subtraction multiplication coefficient Cs ₁₂ . Then, the addition / subtraction unit 178 generates a second subtraction signal by subtracting the signal obtained by the multiplication from the first subtraction signal. The second addition signal and the second subtraction signal are input to the FDE unit 180.

  The FDE unit 180 multiplies the second addition signal input from the addition / subtraction unit 178 by the equalization coefficient Fa (1) for addition, and outputs the multiplication result (Z (1)) to the IDFT unit 120. To do. Further, the FDE unit 180 multiplies the second subtraction signal input from the addition / subtraction unit 178 by the subtraction equalization coefficient Fs (1). The signal multiplied by the subtraction equalization coefficient Fs (1) is branched and input to the adder / subtractor 182.

(4th stage processing)
Next, the signal equalization unit 106 branches the signal {Q ^H y} (0) input from the multiplication processing unit 104 into two and inputs the signal to the addition / subtraction unit 182. Subtraction unit 182 multiplies the addition for the multiplication coefficient _{Ca 03} to the signal input from FDE section 172. Then, the adder / subtractor 182 adds the signal obtained by the multiplication to the signal {Q ^H y} (0) to generate a first addition signal. Similarly, subtraction unit 182 multiplies the addition for the multiplication coefficient _{Ca 02} to the signal input from FDE section 176. Then, the adder / subtractor 182 adds the signal obtained by the multiplication to the first addition signal to generate a second addition signal. Furthermore, subtraction unit 182 multiplies the addition for the multiplication coefficient _{Ca 01} to the signal input from FDE section 180. Then, the adder / subtractor 182 adds the signal obtained by the multiplication to the second addition signal to generate a third addition signal. The third addition signal is input to the FDE unit 184.

  The FDE unit 184 multiplies the third addition signal input from the addition / subtraction unit 182 by the addition equalization coefficient Fa (0), and outputs the multiplication result (Z (0)) to the IDFT unit 120. To do. By performing the above-described processing by the signal equalization unit 106, it is possible to equalize the frequency selectivity that each element of the upper triangular matrix R has independently. This is shown in FIG.

(Regarding the effect of the signal equalization unit 106 processing)
Here, the effect of the processing of the signal equalization unit 106 will be described briefly with reference to FIG. FIG. 3 is an explanatory diagram showing the effect of equalization processing executed in the third stage of the signal equalization unit 106. The third stage process will be described as an example, but the same applies to the other stage processes.

First, reference is made to (IN) in FIG. (IN) schematically shows the frequency characteristics of the elements R ₁₁ , R ₂₁ , R ₃₁ corresponding to the third row from the bottom of the upper triangular matrix R. As shown in (IN), the elements R ₁₁ , R ₂₁ , and R ₃₁ have independent frequency characteristics. Therefore, the signal {Q ^H y} (3) input from the multiplication processing unit 104 has independent frequency selectivity for each substream.

Next, reference is made to (A1) in FIG. (A1) schematically shows the frequency characteristics of the elements R ₁₁ , R ₂₁ , and R ₃₁ of the upper triangular matrix R in the first addition signal. As shown in (A1), the signal obtained by multiplying the output signal of the first stage by the multiplication coefficient Ca ₁₃ for addition is added to the signal {Q ^H y} (3), whereby the element R ₃₁ it is possible to obtain the effect of frequency characteristic is corrected to the same frequency characteristics as the element R _11.

Referring now to (A2) in FIG. 3, by a signal obtained by multiplying the sum for multiplication coefficient Ca ₁₂ to the second stage of the output signal is further added, the frequency characteristic of the element R ₂₁ is it is possible to obtain an effect that is corrected to the same frequency characteristics as the element R _11. Finally, by multiplying the output signal by the equalization coefficient Fa (1), the frequency characteristics of the elements R ₁₁ , R ₂₁ , R ₃₁ are flattened as shown in (A3). Can be obtained.

Next, reference is made to (S1), (S2), and (S3) in FIG. (S1) is the frequency characteristic of the signal obtained by subtracting the signal obtained by multiplying the output signal of the first stage by the subtraction multiplication coefficient Cs ₁₃ from the signal {Q ^H y} (3). Similarly, (S2) from the signal, the frequency characteristic of the second-stage output signal to the signal obtained by subtracting a signal obtained by multiplying the subtraction for multiplication coefficient Cs _12. That is, in (S1) (S2), each component of the signal {Q ^H y} (3) corresponding to the off-diagonal elements of the upper triangular matrix R is removed. Finally, the subtraction equalization coefficient Fs (1) is multiplied to flatten the signal component corresponding to the diagonal element of the upper triangular matrix R.

  As described above, the signal equalization unit 106 can equalize the frequency selectivity included independently in each element of the upper triangular matrix R. As a result, in a single carrier transmission MIMO system, it is possible to effectively suppress the influence of multipath interference. In addition, since the signal separation method is based on maximum likelihood detection, a high S / N ratio is realized by obtaining diversity gain, and as a result, the maximum power consumption can be reduced. However, since the frequency selectivity of the substreams processed in other stages is emphasized by the equalization process executed in each stage, it may not be sufficiently removed when the influence of multipath interference is large.

<First Embodiment>
Next, a first embodiment of the present invention will be described. The present embodiment is based on the above QRDE-MLD method, and is characterized in that multipath interference removal processing by iterative processing is executed before signal equalization processing. This makes it possible to take advantage of the QRDE-MLD method even in the case of broadband transmission or the like that is greatly affected by multipath interference.

[Functional configuration of receiving apparatus 200]
First, the functional configuration of the receiving apparatus 200 according to the present embodiment will be described with reference to FIG. FIG. 4 is an explanatory diagram illustrating a functional configuration of the receiving device 200 according to the present embodiment. Note that components having substantially the same functions as those of the reception apparatus 100 according to the QRDE-MLD method are denoted by the same reference numerals, and detailed description thereof is omitted.

  As illustrated in FIG. 4, the reception apparatus 200 mainly includes an FFT unit 102, a multipath interference removal unit 202, a multiplication processing unit 104, a signal equalization unit 106, a channel estimation unit 108, and a QR decomposition unit. 110, equalization coefficient calculation units 112 and 114, multiplication coefficient calculation units 116 and 118, IDFT unit 120, EDC unit 122, LLR unit 124, S / P conversion unit 126, FEC unit 128, The control unit 204, the modulation unit 206, and the DFT unit 208 are configured. The main structural differences from the above receiving apparatus 100 are a multipath interference canceller 202, a controller 204, a modulator 206, and a DFT section 208. Therefore, these differences will be mainly described.

(Control unit 204, modulation unit 206, DFT unit 208)
The control unit 204 is a unit that performs control so as to generate a transmission replica signal x ′ using reproduction data reproduced based on substantially the same processing as that of the QRDE-MLD receiver 100. First, the reproduction data output from the FEC unit 128 is input to the modulation unit 206 to generate a time-domain transmission replica signal x ′ (t). Next, the time-domain signal replica signal x ′ (t) is input to the DFT unit 208 and converted into a frequency-domain transmission replica signal x ′ (ω). The frequency domain transmission replica signal x ′ (ω) is input to the multipath interference canceling unit 202 and the multiplication processing unit 104.

(Multipath interference removal unit 202)
The multipath interference removing unit 202 is means for removing the influence of multipath interference included in the frequency domain received signal input from the FFT unit 102 by iterative processing. The multipath interference canceling unit 202 subtracts the frequency-dependent component ΔH (ω) of the channel matrix H using the transmission replica signal x ′ output from the DFT unit 208 as shown in the following equation (8). Remove frequency selectivity. However, the transmission replica signal x ′ may contain an error, and there is a possibility that the frequency selectivity remains without the transmission signal x and the transmission replica signal x ′ being in agreement. However, this residual frequency selectivity (or residual multipath interference) is removed by the signal equalizer 106 at the subsequent stage. H _av represents the average of the channel matrix H in the frequency domain.

  FIG. 5 shows a functional configuration of the multipath interference removal unit 202. As illustrated in FIG. 5, the multipath interference cancellation unit 202 includes subtraction units 232, 234, 236, and 238 for each reception substream. Each subtraction unit 232, 234, 236, 238 receives the subtraction signal dry (k) = {ΔH (ω) x ′} (k) (k = 0-3) generated based on the transmission replica signal x ′. A process of subtracting from the signal y (k) (y (k) -dry (k)) is performed.

  Processing such as the output of reproduction data, generation of a transmission replica signal, and removal of multipath interference as described above is repeated until it is determined that there is no error for all substreams. The presence / absence of an error is determined by performing a cyclic redundancy check (CRC) on each substream. In addition, in this iterative process, the repetitive process may be terminated when a predetermined processing delay time is reached by urging the transmitting side to retransmit.

(Detailed functional configuration of the control unit 204)
Here, a detailed functional configuration of the control unit 204 will be described. The control unit 204 controls the operation of the multipath interference removal unit 202 while monitoring the transmission path condition and the decoded data decoding quality. The average value H _av of the channel matrix H may be zero or a very low level according to changes in the phase, amplitude, etc. within the frequency band, for example. In such a case, the control unit 204 changes the averaging method for calculating the average value H _av of the channel matrix H. For example, the control unit 204 uses power weighted average in the frequency domain as an averaging method, or applies a transfer function of a path having the maximum level in the time domain (impulse response) instead of the average value H _av. May be. The control unit 204 can change such a method according to the state of the transmission path.

When determining the averaging method, the control unit 204 calculates the average value H _av of the channel matrix H and the frequency-dependent component ΔH (ω). Next, the control unit 204 causes the multipath interference removal unit 202 to perform the subtraction process expressed by the above equation (8) using the transmission replica signal x ′. Further, the control unit 204 causes the QR decomposition unit 110 to perform QR decomposition on the average value H _av of the channel matrix H. Further, the control unit 204 causes the signal equalization unit 302 to perform frequency domain equalization processing on the residual component of the frequency dependent component ΔH (ω).

  Furthermore, the control unit 204 may weight each substream assuming that the transmission replica signal x ′ is not correctly generated. Actually, since the signal power to noise / interference power ratio (SINR) required for obtaining a desired quality differs for each modulation method, the signal quality after decoding may be different for each modulation method. However, the decoding quality can be estimated from the absolute value average (for example, Σ | tanh (LLR) |) of the soft value bits. Thus, by using this estimated value to assign a weight to the transmission replica signal x ′ for each substream, it is possible to mitigate the degradation of the decoding quality caused by the low-quality transmission replica signal.

  For example, consider a case where QPSK is assigned to stream number 1 and 64QAM is assigned to stream number 2. In this case, by assigning a weight of 1.0 to QPSK (stream number 1) and giving a weight of 0.5 (<1.0) to 64QAM (stream number 2), the level of a stream that is prone to error (64QAM) is increased. It is possible to reduce the degradation of the decoding quality. Note that the control unit 204 may change the weight for each substream in the DFT cycle, or may change it in one section of the error correction code.

(Summary of the first embodiment)
Heretofore, the functional configuration and signal processing method of the receiving apparatus 200 according to the present embodiment have been described. The above method is premised on iterative decoding, and is characterized in that multipath interference is removed in the frequency domain in the second and subsequent iterative decoding processes. Also, QRDE-MLD signal equalization processing is performed on the signal output after multipath interference is removed. The control related to this processing is adaptively executed according to the condition of the transmission path and the decoding quality.

  When the above apparatus and method are used, multipath interference can be suppressed at maximum likelihood detection without imposing a high power burden on the transmission side. As a result, the communication quality can be improved without sacrificing the high-speed performance during high-speed packet transmission. In addition, since diversity gain between receiving antennas by maximum likelihood detection can be obtained, a significant improvement in signal characteristics is expected compared to the MMSE method. As a result, the transmission power required to obtain the desired communication quality is reduced, leading to an increase in the battery duration of the mobile terminal. Also, considering the downlink, power saving of infrastructure facilities can be realized. Of course, the effect of extending the radio wave reach is also obtained.

(simulation result)
FIG. 6 shows a simulation result of the reception characteristics according to this embodiment. The simulation conditions are based on Evolved UTRA and UTRAN (commonly called LTE: Long Term Evolution), which is being standardized by 3GPP. Also, a turbo interleaver configuration using a turbo code as a channel code is adopted. In this configuration, an interleaver is applied before the DFT on the transmission side, and a deinterleaver is applied before turbo decoding on the reception side. By adopting this configuration, it is possible to reduce the correlation between the substreams, so that improvement in reception characteristics can be expected. In this simulation, the turbo code has a constraint length of 4 and a maximum number of iterations of 7 decoding.

  For comparison, FIG. 6 shows characteristics of SC-MMSE (Soft-Cansel MMSE) and QRDE-MLD, in addition to the system according to the present embodiment (indicated as + MPIC). As shown in FIG. 6, in the system according to the present embodiment, the SNR is improved by about 1.8 dB at a packet error rate (PER) 0.1 compared to the QRDE-MLD system. In addition, when the PER is 0.1 or less, a significant characteristic improvement is realized as compared with the SC-MMSE system.

Second Embodiment
Next, a second embodiment according to the present invention will be described. In the present embodiment, in order to avoid the frequency selectivity enhancement effect resulting from QR decomposition, the MMSE-based equalization coefficient is used at the first iteration, and the processing related to the QRDE-MLD scheme is executed after the next time. There is a feature in the point to do. Therefore, in the first embodiment and the present embodiment, the coefficients given to the multiplication processing unit 104 and the signal equalization unit 302 shown in FIG. 7 and the substantial processing flow by the signal equalization unit 302 are mainly used. Is different. Therefore, these differences will be mainly described with reference to FIG.

(Problem presentation)
As already described, there is a problem that frequency selectivity is emphasized during frequency domain equalization by using QR decomposition. This problem suggests that the frequency selectivity may be increased by applying the technology to broadband communication. For example, when frequency domain equalization is performed after applying a signal separation method based on QR decomposition, there is a high possibility that frequency selectivity is not corrected in order to suppress noise enhancement. Therefore, residual multipath interference increases, resulting in a decrease in decoding performance at the first iteration. Therefore, in the present embodiment, a method of performing equalization processing in the frequency domain based on the MMSE-based equalization coefficient only for the first time of iterative decoding and continuing the subsequent processing using the replica signal obtained as a result. suggest.

[Functional configuration of receiving apparatus 300]
First, the functional configuration of the receiving apparatus 300 according to the present embodiment will be described with reference to FIG. FIG. 7 is an explanatory diagram illustrating a functional configuration of the receiving device 300 according to the present embodiment.

  As illustrated in FIG. 7, the reception apparatus 300 mainly includes an FFT unit 102, a multipath interference removal unit 202, a multiplication processing unit 104, a signal equalization unit 302, a channel estimation unit 108, and a QR decomposition unit. 110, equalization coefficient calculation units 112 and 114, multiplication coefficient calculation units 116 and 118, IDFT unit 120, EDC unit 122, LLR unit 124, S / P conversion unit 126, FEC unit 128, The control unit 304, the modulation unit 206, and the DFT unit 208 are configured. The main structural differences from the above receiving apparatus 200 are the signal equalization unit 302 and the control unit 304.

(Signal equalization unit 302, control unit 304)
Similar to the signal equalization unit 106 in the first embodiment, the signal equalization unit 302 performs frequency domain equalization on the received substream corresponding to each row of the upper triangular matrix in units of rows in order from the lowest row of the upper triangular matrix. Is a means for equalizing frequency selectivity generated independently for each element of the upper triangular matrix. However, the signal equalization unit 302 can perform signal equalization using an MMSE-based equalization coefficient described later.

Note that the MMSE-based equalization coefficient is calculated by the control unit 304 and input to the multiplication processing unit 104 and the signal equalization unit 302. Multiplication processing unit 104, instead of the unitary matrix Q ^H, multiplies the equalization coefficient W of MMSE criterion expressed by the following equation (9) to the received signal. The signal equalization unit 302 performs equalization processing using a diagonal matrix formed by the diagonal component of the product WH of the MMSE-based equalization coefficient W and the channel matrix H, instead of the upper triangular matrix R. Run.

  When the transmission replica signal used in the multipath interference removal unit 202 is first generated, the control unit 304 causes the equalization processing based on the MMSE-based equalization coefficient expressed by the above equation (9) to be executed. In addition, after the processing by the multipath interference removal unit 202 is performed, the control unit 304 is configured to switch from equalization processing based on the MMSE-based equalization coefficient to equalization processing based on QR decomposition. 104 and the signal equalizer 302 are controlled. Note that this control is realized by changing the equalization coefficient.

  With the above control, it is possible to improve characteristics when the maximum likelihood decoding based on QR decomposition is poor due to a significant decrease in the level of elements located at the lower right of the upper triangular matrix R. However, if the equalization process is executed based on the MMSE standard, the reception diversity gain is reduced. However, since the substreams are theoretically orthogonalized and the multipath interference between the substreams is suppressed to a noise level at most, in a poor situation where the multipath interference exceeds the diversity effect, The performance of the entire receiving apparatus can be improved by applying the MMSE standard.

  Further, the reception diversity effect can be obtained by switching to the equalization process of the QR decomposition standard in the second and subsequent stages of the iterative process. Therefore, further improvement of characteristics is realized. As a result, the SNR necessary for obtaining the desired reception quality is reduced, and effects such as power saving of the terminal device or expansion of the service area provided by the base station can be obtained. Furthermore, by applying the technique according to the present embodiment, it is possible to reduce the number of iterations in the decision feedback type demodulation method based only on the MMSE criterion. As a result, it is possible to obtain a significant reduction effect of the circuit scale or the processing delay time, and it is possible to expect an economic effect that the infrastructure cost is reduced and the provision cost of the mobile communication service is reduced.

  As mentioned above, although preferred embodiment of this invention was described referring an accompanying drawing, it cannot be overemphasized that this invention is not limited to the example which concerns. It will be apparent to those skilled in the art that various changes and modifications can be made within the scope of the claims, and these are naturally within the technical scope of the present invention. Understood.

  For example, in the description of each embodiment, the number of reception antennas N is set to four, but the present invention is not limited to this, and any number of N ≧ 2 may be used.

  By the way, the characteristics of frequency domain equalization greatly change according to a fixed value included in the denominator of the coefficient. Therefore, instead of considering only the SNR, a coefficient corresponding to the wireless communication bandwidth may be set to a fixed value. Further, a multipath interference term estimated from the component of the upper triangular matrix R may be added to a fixed value.

  Further, although a method called a pseudo-inverse MMSE standard (Pseudo-inverse MMSE) in which the concept of MMSE is added to QR decomposition has been reported, the technique according to the present invention can be applied to the pseudo-inverse MMSE standard. . The technology according to the present invention can be widely applied to a single carrier communication system in which multipath interference is a problem. For example, a radio communication method according to IFDMA (Interleaved Frequency Domain Multiple Access), VSCRF-CDMA (Variable Spreading Chip Repetition Factor-CDMA), or a cell with a very short time is allocated to each user and other sectors. The present invention is applied to a wireless communication system or the like according to a high-speed packet transmission system that reduces interference between the same channels.

It is explanatory drawing which shows the function structure of the receiver which concerns on a QRDE-MLD system. It is explanatory drawing which shows the function structure of the signal equalization part which concerns on the same system. It is explanatory drawing which shows the mode of the frequency selectivity removal by the signal equalization part of the same system. It is explanatory drawing which shows the function structure of the receiver which concerns on 1st Embodiment of this invention. It is explanatory drawing which shows the function structure of the multipath interference removal part which concerns on the same embodiment. It is explanatory drawing which shows the effect by the receiver which concerns on embodiment of this invention. It is explanatory drawing which shows the function structure of the receiver which concerns on 2nd Embodiment of this invention.

Explanation of symbols

100, 200, 300 Receiver 102 FFT unit 104 Multiply processing unit 106 Signal equalization unit 108 Channel estimation unit 110 QR decomposition unit 112, 114 Equalization coefficient calculation unit 116, 118 Multiplication coefficient calculation unit 120 IDFT unit 122 EDC unit 124 LLR Unit 126 S / P conversion unit 128 FEC unit 172, 176, 180, 184 FDE unit 174, 178, 182 addition / subtraction unit 202 multipath interference removal unit 204 control unit 206 modulation unit 208 DFT unit 302 signal equalization unit 304 control unit

Claims (4)

A receiving device for separating multiple signals received by a plurality of antennas by maximum likelihood detection,
An interference component subtraction unit that subtracts the frequency-dependent component of the channel matrix using the replica signal;
A matrix decomposition unit that decomposes the channel matrix from which the frequency-dependent component has been subtracted into a product of a unitary matrix and an upper or lower triangular matrix;
Based on the upper or lower triangular matrix, a signal equalization unit that equalizes frequency selectivity for each reception substream by performing frequency domain equalization in units of rows of the upper or lower triangular matrix;
A receiving apparatus comprising:   The replica generation unit according to claim 1, further comprising a replica generation unit configured to generate the replica signal based on each substream component of the multiplexed signal separated using a MMSE (Minimum Mean Squared Error) equalization coefficient. Receiver.   The replica generation unit selects whether to use the MMSE equalization coefficient or the unitary matrix and the upper or lower triangular matrix according to a transmission path condition when separating the multiplexed signal. The receiving device according to claim 2, wherein: A signal processing method capable of suppressing the influence of multipath interference when multiple signals received by a plurality of antennas are separated by maximum likelihood detection,
An interference component subtraction step in which the frequency dependent component of the channel matrix is subtracted using the replica signal;
A matrix decomposition step in which the channel matrix from which the frequency dependent component has been subtracted is decomposed into a product of a unitary matrix and an upper or lower triangular matrix;
Based on the upper or lower triangular matrix, a signal equalization step in which frequency domain equalization is performed in units of rows of the upper or lower triangular matrix and frequency selectivity is equalized for each reception substream;
A signal processing method.

Images