Receiving device, wireless communication system, and receiving method

Abstract

To suppress inter-stream interference and improve wireless transmission quality, even when propagation path fluctuations are rapid. [Solution] A receiving device comprising: multiple Fourier transform units that convert the wireless frames received by multiple antennas that receive multiple wireless frames of the same frequency transmitted from one or more other wireless communication devices into the frequency domain at predetermined intervals; and multiple filters that estimate the path response in each of the multiple frequency domains using multiple known signals present in the wireless frame and perform equalization processing based on the estimated multiple path responses.

Description

[Technical Field]

[0001] The present invention relates to a receiving device, a wireless communication system, and a receiving method. [Background technology]

[0002] Conventionally, in order to cope with the increasing capacity of wireless communication, the use of MIMO (Multiple-Input Multiple-Output) technology, which multiplexes and transmits multiple streams of signals on the same space, time, and frequency, has been explored (see, for example, Non-Patent Document 1). In MIMO communication, the wireless propagation path between multiple antennas is estimated in advance by transmitting a known pilot signal, etc. (for example, called propagation path estimation, channel estimation, etc.), and the separation process of multiple signal streams is performed based on the estimated wireless propagation path between the multiple antennas.

[0003] On the other hand, the wireless propagation path changes moment by moment, mainly due to the movement of the wireless communication device on the terminal side and changes in the surrounding environment. Therefore, if a discrepancy occurs between the wireless propagation path estimated in advance and the wireless propagation path through which the signal stream is actually transmitted, inter-stream interference occurs, degrading the transmission characteristics. Thus, methods to avoid such problems are being considered (see, for example, Non-Patent Document 2). In the technology described in Non-Patent Document 2, the transmitting side transmits data using encoding with LDPC (Low Density Parity Check) codes in the transmission sequence, and the receiving side uses a replica of the interference signal generated based on the high-reliability symbol after decoding the data. Then, by applying frequency-domain repeated equalization, which is repeatedly subtracted from the received signal at the receiving side, inter-stream interference is suppressed and diversity gain is improved. [Prior art documents] [Non-patent literature]

[0004] [Non-Patent Document 1] Tomoaki Otsuki, "Overview of MIMO Technology," Journal of the Institute of Image Information and Television Engineers, 2016, Vol. 70, No. 1, pp. 2-5. [Non-Patent Document 2] Yuki Nagano, Eiichi Murata, "Transmission Experiment of Frequency Domain Repetition Equalization in a MIMO System Based on Terminal Shared Reception," Institute of Electronics, Information and Communication Engineers, IEICE Technical Report RCS2014-366(2015-3), pp. 387-391. [Overview of the project] [Problems that the invention aims to solve]

[0005] However, even when frequency-domain iterative equalization is applied, inter-stream interference may not be sufficiently suppressed when propagation path fluctuations are rapid. This is not limited to iterative equalization; when performing MIMO equalization in the frequency domain, a data sequence (data block) of a specific time length is accumulated in advance, and a Fast Fourier Transform (FFT) is applied to the accumulated data sequence (data block) to convert it to the frequency domain. Because of this process, the estimation of the propagation path matrix used for inter-stream separation and the generation of weights based on it can only be done in units of data blocks of a specific time length. Furthermore, if the propagation path fluctuates significantly before the inter-stream separation process is actually performed using the weights, the interference suppression effect of the weights decreases, and inter-stream interference cannot be completely suppressed. Thus, conventionally, when propagation path fluctuations are rapid, inter-stream interference cannot be suppressed, resulting in a decrease in wireless transmission quality.

[0006] In view of the above circumstances, the present invention aims to provide a technology that can suppress inter-stream interference and improve wireless transmission quality even when propagation path fluctuations are rapid. [Means for solving the problem]

[0007] One aspect of the present invention is a receiving device comprising: a plurality of Fourier transform units that convert the wireless frames received by each of a plurality of antennas that receive a plurality of wireless frames of the same frequency transmitted from one or more other wireless communication devices into the frequency domain at predetermined time intervals; and a plurality of filters that estimate the path response in each of the plurality of frequency domains using a plurality of known signals present in the wireless frame and perform equalization processing based on the estimated plurality of path responses.

[0008] One aspect of the present invention is a wireless communication system comprising a first wireless communication device and one or more second wireless communication devices, wherein the first wireless communication device transmits a plurality of wireless frames of the same frequency, each containing a plurality of known signals, to the one or more second wireless communication devices, and the one or more second wireless communication devices each include a plurality of Fourier transform units that convert the plurality of wireless frames transmitted from the first wireless communication device, or the wireless frames received by a plurality of antennas that receive at least one of the plurality of wireless frames transmitted from the first wireless communication device and wireless frames transmitted from other second wireless communication devices, into the frequency domain at predetermined intervals, and a plurality of filters that estimate the path response in each of the plurality of frequency domains using a plurality of known signals present in the wireless frames, and perform equalization processing based on the estimated plurality of path responses.

[0009] One aspect of the present invention is a receiving method that converts each of several wireless frames of the same frequency transmitted from one or more other wireless communication devices, received by each of several antennas, into the frequency domain at predetermined intervals, estimates the path response in each of the several frequency domains using several known signals present in the wireless frames, and performs equalization processing based on the estimated several path responses. [Effects of the Invention]

[0010] This invention makes it possible to suppress inter-stream interference and improve wireless transmission quality, even when propagation path fluctuations are rapid. [Brief explanation of the drawing]

[0011] [Figure 1] This is a schematic diagram showing a typical MIMO system configuration. [Figure 2] This figure shows an example configuration of a wireless communication device that performs signal stream separation processing by frequency domain equalization. [Figure 3] This figure shows an example configuration of a wireless communication device that performs signal stream separation processing by repeated equalization in the frequency domain. [Figure 4] This figure shows an example of a conventional wireless frame. [Figure 5] This figure shows an example of a wireless frame configuration in the first embodiment. [Figure 6] This is a flowchart showing the processing flow of the wireless communication device in the first embodiment. [Figure 7] This is a schematic diagram showing an example configuration of a wireless communication system according to the second embodiment. [Figure 8] This figure shows an example configuration of a wireless communication device that performs signal stream separation processing by repeated equalization in the frequency domain. [Modes for carrying out the invention]

[0012] One embodiment of the present invention will be described below with reference to the drawings.

[0013] (Premise) Before describing the specific details of the present invention, we will explain the MIMO technology that forms the basis of the present invention and frequency domain equalization, which is its demodulation method. Figure 1 is a schematic diagram showing an example of the configuration of a typical MIMO system. The MIMO system comprises a wireless communication device 80BS and a wireless communication device 80T. In Figure 1, as an example, the wireless communication device 80BS is a base station device and the wireless communication device 80T is a terminal device. That is, although Figure 1 explains the case of downlink transmission as an example, the present invention can be similarly applied to base station-side processing in the uplink. Therefore, the wireless communication device 80BS may be a terminal device and the wireless communication device 80T may be a base station device. When the wireless communication device 80BS is a base station device and the wireless communication device 80T is a terminal device, the MIMO system may comprise multiple wireless communication devices 80T.

[0014] The base station device, wireless communication device 80BS, performs spatial multiplexing transmission with multiple wireless communication devices 80T at the same time on the same frequency in the same space. Here, "same frequency" does not mean strictly the same frequency value, but rather the same frequency band (same frequency channel) having a certain bandwidth centered on the center frequency allocated for MIMO transmission. Furthermore, if the base station device, wireless communication device 80BS, accommodates multiple terminal devices, wireless communication devices 80T, the wireless communication device 80BS may select some of the multiple wireless communication devices 80T to perform MIMO transmission.

[0015] The wireless communication device 80BS forms a plurality of directional beams using a number of antenna elements. For example, consider the case where three MIMO channels are allocated to the wireless communication device 80T and three signal sequences are transmitted. The separation process of these three signals may be performed on the wireless communication device 80BS side, on the wireless communication device 80T side, or both. In this case, either one or both of the wireless communication devices 80BS and 80T that perform the signal separation process grasp the state of the wireless propagation path between the wireless communication device 80BS and the wireless communication device 80T, and perform the separation process based on the information indicating the state of the wireless propagation path. When the signal separation process is performed on the transmission side, the process is sometimes called precoding, and when the signal separation process is performed on the reception side, the process is sometimes called postcoding.

[0016] The separation process of signals in the MIMO channel is performed as follows. Here, as shown in FIG. 1, the case where both the wireless communication device 80BS and the wireless communication device 80T each include three antenna elements will be described. For example, the channel information between the j-th (j = 1,..., 3) antenna element of the wireless communication device 80BS and the i-th (i = 1,..., 3) antenna element of the wireless communication device 80T is denoted as h ij The column vector h1 corresponding to the first antenna element (j = 1) of the wireless communication device 80BS is set to (h 11 , h 21 , h 31 ) T Similarly, the column vector h2 corresponding to the second antenna element (j = 2) of the wireless communication device 80BS is set to (h 12 , h 22 , h 32 ) T And the column vector h3 corresponding to the third antenna element (j = 3) of the wireless communication device 80BS is set to (h 13 , h 23 , h 33 ) TLet's assume that the number of column vectors is equal to the number of antenna elements of the wireless communication device 80BS, and each of the column vectors is a vector with dimensions equal to the number of antenna elements of the wireless communication device 80T. Here, the symbol "T" in the upper right corner represents the transpose of a vector or matrix.

[0017] Channel information refers to a value representing the state of the radio propagation path, that is, the actual propagation coefficient of the channel between the antenna element of the transmitting wireless communication device 80BS and the antenna element of the receiving wireless communication device 80T, which is estimated and obtained by the transmitting wireless communication device 80BS or the receiving wireless communication device 80T (strictly speaking, this includes the influence of amplifiers and filters within the transmitting and receiving wireless communication devices 80BS and 80T).

[0018] Let the three signals transmitted by the wireless communication device 80BS be denoted as t1 to t3, and let the column vector with these as components be T x =(t1,t2,t3) T This is how it is expressed. The received signals received by the three antenna elements of the wireless communication device 80T are denoted as r1 to r3, and the column vector with these as components is R x =(r1,r2,r3) T This is denoted as follows. The matrix whose column vectors h1 to h3 are the first to third columns is denoted as the channel matrix H. Noise is denoted as n. In this case, the following relationship (1) holds for the entire MIMO system.

[0019]

number

[0020] Here, when the receiving wireless communication device 80T performs signal separation processing, i.e., postcoding, the received signal vector R x From the transmitted signal vector T x The process involves extracting the weight matrix W (received weight matrix) as shown in equation (2) below, and the received signal vector R at the receiving end. x Perform a multiplication operation from left to right.

[0021]

number

[0022] When the noise n is sufficiently small, the weight matrix W follows the ZF (Zero-Forcing) standard that satisfies equation (3) below. ZF Using the (received weight matrix), the following equation (4) holds, and the received signal vector R x From the transmitted signal vector T x This can be extracted. Note that in equation (3), matrix I is a 3x3 identity matrix.

[0023]

number

[0024]

number

[0025] Weight matrix W ZF The calculation involves a weight matrix W that satisfies equation (3). ZF Any calculation method can be applied to calculate the weight matrix W. For example, when using the Moore-Penrose generalized inverse, the weight matrix W can be calculated. ZF This is calculated by the following equation (5).

[0026]

number

[0027] In equation (5), the "+" symbol in the superscript indicates a pseudo-inverse matrix, the "-1" symbol indicates an inverse matrix, and the "H" symbol indicates a conjugate transpose matrix. The elements h of the channel matrix H ij And the weight matrix W ZF element w ij The following relationship holds between them.

[0028]

number

[0029] In other words, for the desired signal stream, i.e., when i=j, the energy is controlled so that it does not become zero, and for signal streams other than the desired one, i.e., when i≠j, the energy is controlled so that it becomes zero. For the latter case, the control is performed to create a point where no signal is received (a null). For example, in a line-of-sight environment, the control is performed so that the beam shape becomes a valley, and this control is called null formation.

[0030] Note that the weight matrix W ZF The row vectors represent the weight vectors (received weight vectors) corresponding to each antenna element of the wireless communication device 80BS. That is, equation (6) is the weight matrix W, which is the received weight vector corresponding to the i-th antenna element of the wireless communication device 80BS. ZF The row vector of the i-th row (w i1 ,w i2 ,w i3 ) and the column vector of the jth column of the channel matrix H (h 1j ,h 2j ,h 3j This indicates that null formation is performed by multiplying by T such that if i≠j, the result is 0, and if i=j, the result is non-zero. MIMO transmission is performed through this orthogonalization process.

[0031] (Signal separation processing) Here, we will explain the procedure for signal separation on the receiving side. First, the wireless communication device 80T receives a known reference signal from the wireless communication device 80BS at a predetermined timing and estimates the channel matrix H based on the received known reference signal. Next, the wireless communication device 80T estimates the received weight matrix, i.e., the weight matrix W that follows the ZF standard. ZF The wireless communication device 80T calculates the received signal vector R. x For the weight matrix W ZF Multiply by the desired transmit signal vector Tx The process is terminated after obtaining the data. Note that channel information is usually different for each frequency component, so for broadband signals, such as signals using OFDM (Orthogonal Frequency Division Multiplexing) modulation, the same weight matrix W is obtained for each frequency component, i.e., for each subcarrier. ZF This will result in the calculation.

[0032] When transmitting from wireless communication device 80T to wireless communication device 80BS, the transmission weight matrix used when the transmitting wireless communication device 80T performs signal separation processing, i.e., precoding, can be calculated, for example, as follows: The received signal vector received by wireless communication device 80BS is "R x "The transmission signal vector transmitted by the wireless communication device 80T is "T x Let the channel matrix be "H" and the noise be "n". Here, "R x =H·W·T x Let us assume the relationship "+n". In this relationship, we can calculate a transmission weight matrix W such that "H·W" is the identity matrix I, and if the noise n is sufficiently small, then the relationship is R x ≒T x Therefore, in the wireless communication device 80T, the transmitted signal vector T x If the result of multiplying by the transmission weight matrix W from the left is transmitted, the received signal vector R received by the wireless communication device 80BS will be transmitted. x Transmit signal vector T x This can be considered as follows. The transmission weight matrix W is the same as the weight matrix W described above. ZF It can be calculated using a similar calculation method. In this case, the channel matrix H is a matrix obtained by converting the column vectors h1, h2, h3 into row vectors and arranging them, so the transmit weight matrix W is a matrix obtained by arranging the column vectors, the transmit weight vectors.

[0033] The above describes the signal separation process at the receiving end in a typical MIMO system. Similar signal separation processes exist in several variations, as described in Non-Patent Document 1. Note that the value of "i" used in the above explanation differs from the value of "i" used in the following explanation.

[0034] (Example configuration of the receiving wireless communication device 1) Figure 2 shows an example configuration of a wireless communication device 80T that performs signal stream separation processing by frequency domain equalization. The wireless communication device 80T includes a CP removal unit 81, an S / P unit 82, an FFT unit 83, a filter 84, an IFFT unit 85, a P / S unit 86, a BP decoding unit 87, and a hard determination unit 88. In Figure 2, an example is shown in which the wireless communication device 80T has three CP removal units 81, S / P units 82, FFT units 83, IFFT units 85, P / S units 86, BP decoding units 87, and hard determination units 88. However, the wireless communication device 80T only needs to have the same number of CP removal units 81, S / P units 82, FFT units 83, IFFT units 85, P / S units 86, BP decoding units 87, and hard determination units 88 as the number of antennas in the wireless communication device 80T. In addition, the wireless communication device 80T has filters 84 for the number of frequency components F (F is an integer of 2 or more). The following explanation uses the example of a wireless communication device 80T equipped with three antennas. The wireless communication device 80T receives the three signal streams transmitted after LDPC coding processing by the wireless communication device 80BS with each of its three antennas.

[0035] The CP removal unit 81 removes the CP (Cyclic Prefix) from the signal received by the antenna.

[0036] The S / P unit 82 performs serial-to-parallel conversion on the signal from which the CP has been removed in the CP removal unit 81. As a result, the S / P unit 82 divides the signal from which the CP has been removed into data blocks of a specific time length.

[0037] The FFT unit 83 converts each data block into a frequency domain signal by applying a Fast Fourier Transform to each data block of a specific time length divided in the S / P unit 82. The FFT unit 83 outputs each of the converted frequency domain signals to different filters 84 according to their frequency components. For example, the FFT unit 83 outputs the signal of the first frequency component to filter 84-1 and the signal of the fth frequency component to filter 84-F.

[0038] Filter 84 performs MMSE (Minimum Mean Square Error) equalization on the frequency domain signals (hereinafter referred to as "frequency signals") of each data block converted in the FFT section 83. Specifically, filter 84 performs equalization by multiplying the frequency signals of each frequency component by the MMSE weight. For example, filter 84-1 receives the frequency signals of the first frequency component from each of the FFT sections 83-1 to 83-3. Filter 84-1 performs equalization by multiplying the frequency signals of the first frequency component by the MMSE weight. For example, filter 84-F receives the frequency signals of the fth frequency component from each of the FFT sections 83-1 to 83-3. Filter 84-F performs equalization by multiplying the frequency signals of the fth frequency component by the MMSE weight.

[0039] The IFFT section 85 applies an inverse fast Fourier transform to the frequency signals of each frequency component that have undergone MMSE equalization processing in the filter 84, thereby converting the frequency signals of each frequency component into time-domain signals.

[0040] The P / S unit 86 performs parallel-to-serial conversion on the signals in each time domain converted by the IFFT unit 85. This converts the signals in each time domain into a single time domain signal.

[0041] The BP decoding unit 87 calculates the likelihood by performing Belief Propagation (BP) decoding on an LDPC code for a single time-domain signal converted in the P / S unit 86.

[0042] The hard determination unit 88 obtains the decoding result by performing a hard determination on the likelihood calculated by the BP decoding unit 87.

[0043] (Example configuration of the receiving wireless communication device 2) Figure 3 shows an example configuration of a wireless communication device 80aT that performs signal stream separation processing by repeated equalization in the frequency domain. The wireless communication device 80aT includes a CP removal unit 81, an S / P unit 82, an FFT unit 83, a filter 84, an IFFT unit 85, a P / S unit 86, a BP decoding unit 87, a hard determination unit 88, and a soft determination replica generation unit 89. The wireless communication device 80aT differs in configuration from the wireless communication device 80T in that it further includes a soft determination replica generation unit 89.

[0044] In the wireless communication device 80aT, if the BP decoding unit 87 has not performed the processing for a specified number of iterations, it outputs the calculated likelihood to the soft-decision replica generation unit 89. If the BP decoding unit 87 has performed the processing for a specified number of iterations, it outputs the calculated likelihood to the hard-decision unit 88.

[0045] The soft-decision replica generation unit 89 generates soft-decision symbols based on the likelihood output from each BP decoding unit 87, and converts the soft-decision symbols to the frequency domain by performing a Fast Fourier Transform. This allows the soft-decision replica generation unit 89 to generate soft-decision replicas of the interference signal in the frequency domain. The soft-decision replica generation unit 89 generates soft-decision symbols based on the likelihood output from BP decoding unit 87-1, and generates a first soft-decision replica of the interference signal based on the generated soft-decision symbols. Similarly, the soft-decision replica generation unit 89 generates soft-decision symbols based on the likelihood output from BP decoding units 87-2 and 87-3, and generates a second soft-decision replica and a F-th soft-decision replica of the interference signal based on the generated soft-decision symbols.

[0046] The soft-decision replica generation unit 89 outputs the first soft-decision replica of the generated interference signal to filter 84-1. The soft-decision replica generation unit 89 outputs the second soft-decision replica of the generated interference signal to filter 84-2. The soft-decision replica generation unit 89 outputs the fth soft-decision replica of the generated interference signal to filter 84-F.

[0047] Filter 84 performs equalization by subtracting the soft-decision replica of the interference signal output from the soft-decision replica generation unit 89 from the received signal (the frequency domain signal output from the FFT unit 83), and then multiplying it by the MMSE weight.

[0048] Next, the above process will be formally explained. Let's assume that the wireless communication device 80BS has N (where N is an integer greater than or equal to 2) antennas, and the wireless communication devices 80T and 80aT have M (where M is an integer greater than or equal to 2) antennas. In an environment where L (where L is an integer greater than or equal to 1) delayed waves arrive at wireless communication devices 80T and 80aT, the signal vector y(k) ∈ C received at time k. M×1 This can be expressed as shown in equation (7) below.

[0049]

number

[0050] In equation (7), h(l)∈C M×N x(k)∈C represents the propagation path matrix of the l-th path. N×1 Here, n(k)∈C represents the transmitted signal at each time point k. M×1 This represents the noise component at each time k. When frequency domain equalization is applied, the Fast Fourier Transform is applied by filter 84 to the received signal y(f)∈C at the fth frequency (f is an integer greater than or equal to 1) in the frequency domain. M×1 Therefore, the equalized signal "~" X(f)∈C is as shown in equation (8) below. N×1 This can be obtained. Filter 84 is provided for each frequency component. Therefore, each filter 84 will process a different frequency component. For example, filter 84-1 processes the received signal y(1)∈C in the first frequency component. M×1Therefore, the equalized signal "~" X(1)∈C N×1 Similarly, filter 84-2 can obtain the received signal y(2)∈C in the second frequency component. M×1 Therefore, the equalized signal "~" X(2)∈C N×1 This can be obtained. In the following explanation, the symbol "~" placed before a sign is placed above the sign. For example, the symbol "~" placed before the sign "X" is placed above the sign "X".

[0051]

number

[0052] In equation (8), W H (f)∈C N×M This represents the MMSE weight matrix in the frequency domain and is calculated based on the frequency domain propagation path response G(f) shown in equation (9) below. Note that the MMSE weight matrix in the frequency domain differs for each frequency component.

[0053]

number

[0054] In equation (9), where G(f)∈C M×N This represents the propagation path response at the fth frequency and is estimated from a known signal embedded in the transmitted signal beforehand. In equation (9), P represents the transmitted power, No represents the noise power, and I represents the M×M identity matrix.

[0055] The equalized signal is returned to the time domain by the inverse fast Fourier transform performed by the IFFT unit 85, resulting in the signal "~" x(k)∈C. N×1 The output is input to the BP decoding unit 87 and the likelihood is calculated. If iterative equalization is not performed, the final decoding result can be obtained by performing a hard decision on the obtained likelihood output. If iterative equalization is performed, and the specified upper limit of iterations has not been reached, the soft decision replica generation unit 89 generates the soft decision symbol "^" x(k)∈C based on the updated likelihood after BP decoding. N×1This is generated. In the following explanation, the symbol "^" placed before a sign is placed above the sign. For example, the symbol "^" placed before the sign "x" is placed above the sign "X".

[0056] The soft-decision replica generation unit 89 converts the soft-decision symbol "^" x(k) into a frequency-domain signal "^" X(k) ∈ C using the Fast Fourier Transform. N×1 It is converted to this. Then, the soft-decision replica generation unit 89 generates the frequency domain signal "^" X for each transmitting antenna. n For (f), the frequency response vector G for each transmitting antenna n (f)∈C M×1 Using the soft decision replica "^" Y of the interference signal n (f)∈C M×1 This is generated based on the following equation (10). In the second and subsequent passes of filter 84, the replicas of all interference signals other than the desired signal are subtracted as in the following equation (10), and then the weights are multiplied to perform equalization.

[0057]

number

[0058] Here, weight W n (f) is calculated as shown in equation (11) below.

[0059]

number

[0060] In equation (11), β i is 0≦β i This is the residual interference coefficient that satisfies ≤1. The output after this equalization process is "^"X n (f) is subjected to an inverse fast Fourier transform in the IFFT section 85 to obtain the time-domain signal "~" x n (k) is obtained. Then, the time-domain signal "~" x nThe likelihood is updated when (k) is input to the BP decoding unit 87. In the wireless communication devices 80T and 80aT, the above process is repeated until the specified upper limit of repetitions is reached, and then a hard decision is made on the obtained likelihood output to obtain the final decoding result.

[0061] Figure 4 shows an example of a conventional wireless frame. When frequency domain equalization is performed on a wireless frame basis as shown in Figure 4, the wireless communication devices 80T and 80aT perform MIMO equalization of the data portion based on known signals such as a preamble that are pre-embedded in the frame within the frame for a predetermined time length in the time domain. That is, the wireless communication devices 80T and 80aT calculate the frequency domain propagation path response G(f) and use the calculated frequency domain propagation path response G(f) to determine the weight W H (f) is generated. Then, the wireless communication devices 80T and 80aT apply the previously generated weight W to the data portion of the frame other than the known signals. H The iterative equalization process described above is performed using (f). Generally, the time length of the known signal corresponds to a specific time length of the data block divided by the S / P unit 82, so that the known signal is appropriately distributed to each component in the frequency domain.

[0062] However, generally, if the wireless propagation path does not fluctuate, the Fast Fourier Transform H(f) obtained from the time-domain propagation path matrix h(k) when data is actually transmitted will coincide with the frequency-domain propagation path response G(f) used in advance for weight generation. In this case, H(f) is W H Because it is orthogonal to (f), inter-stream interference is completely suppressed. On the other hand, if the radio propagation path changes, H(f) and G(f) will no longer coincide, so H(f) becomes W H (f) becomes non-orthogonal, and inter-stream interference occurs. Although this inter-stream interference can be suppressed to some extent by the iterative equalization process described above, it may not be sufficiently suppressed when propagation path fluctuations are severe (when transmission path fluctuations are rapid). Therefore, the present invention describes a method for suppressing inter-stream interference even when propagation path fluctuations are severe (when transmission path fluctuations are rapid).

[0063] (First Embodiment) In the present invention, a robust weight W´ corresponding to a plurality of wireless propagation path fluctuations is generated in advance, so as to suppress inter-stream interference even when the propagation path fluctuates severely (when the transmission path fluctuates rapidly). The system configuration in the first embodiment is the same as that in FIG. 1, and the device configuration is the same as that of the wireless communication device 80T or the wireless communication device 80aT. For example, when iterative equalization is not performed, the configuration similar to the wireless communication device 80T may be used, and when iterative equalization is performed, the configuration similar to the wireless communication device 80aT may be used. Hereinafter, the differences from the conventional technology will be described. H (f) is generated in advance, so as to suppress inter-stream interference even when the propagation path fluctuates severely (when the transmission path fluctuates rapidly). The system configuration in the first embodiment is the same as that in FIG. 1, and the device configuration is the same as that of the wireless communication device 80T or the wireless communication device 80aT. For example, when iterative equalization is not performed, the configuration similar to the wireless communication device 80T may be used, and when iterative equalization is performed, the configuration similar to the wireless communication device 80aT may be used. Hereinafter, the differences from the conventional technology will be described.

[0064] FIG. 5 is a diagram showing an example of a wireless frame configuration in the first embodiment. FIG. 5 shows an example in which known signals are embedded in each of the two known signal portions provided in the frame. For example, the first known signal portion is provided at the head of the wireless frame, and the second known signal portion is provided in the center of the wireless frame. Note that the number of known signal portions in the wireless frame may be plural, but in the frequency domain, it is necessary to consider that a plurality of known signals can be obtained for each frequency component. This is, for example, that the start point of the second known signal portion is a multiple of a specific time length of the data block divided by the S / P section 82. When the wireless propagation path fluctuates, the propagation path response G(f) (hereinafter referred to as "G 1 (f)") in the frequency domain obtained from the known signal included in the known signal portion provided at the head of the wireless frame and the propagation path response G(f) (hereinafter referred to as "G 2 (f)") in the frequency domain obtained from the known signal included in the known signal portion provided near the center of the frame may be different.

[0065] By generating the MMSE weight based on the extended propagation path response matrix obtained by expanding the matrix of the propagation path responses G 1 (f), G 2 (f) in the two frequency domains, a weight robust to propagation path fluctuations can be generated. Hereinafter, the propagation path responses G 1 (f), G 2When the matrix in (f) is not particularly distinguished, it is simply denoted as the frequency-domain response matrix G(f). The frequency-domain response matrix G(f) ∈ C M×N is the frequency-domain response vector G n (f) ∈ C M×1 for each transmitting antenna and can be decomposed as shown in the following equation (12).

[0066]

Equation

[0067] And G 1 (f) and G 2 (f) in equation (12) are respectively expressed as shown in the following equation (13).

[0068]

Equation

[0069] G 1 (f) and G 2 (f), the extended propagation path response matrix G (1) (f) ∈ C M×(2N-1) for the first transmitting antenna is obtained as shown in the following equation (14).

[0070]

Equation

[0071] G 1 1(f) represents the desired transmitting antenna, G 1 2(f) represents the interfering transmitting antenna (the second transmitting antenna), G 2 2(f) represents the additional vector of the interfering transmitting antenna (the second transmitting antenna), G 1 N (f) represents the interfering transmitting antenna (the Nth (N is an integer greater than or equal to 2) transmitting antenna), G 2 N(f) represents the additional vector for the interfering transmitting antenna (the Nth transmitting antenna (where N is an integer greater than or equal to 2)). Note that the accuracy of interference suppression is improved by not inserting the additional path response vector for the desired transmitting antenna, but inserting it for all other interfering transmitting antennas (the basic principle of the null space extension method).

[0072] The MMSE weight applied to filter 84 on the first run is calculated as shown in equation (15) below. The first run in filter 84 refers to the state where no soft-decision replicas have been input from the soft-decision replica generation unit 89.

[0073]

number

[0074] It should be noted that the components from the second column onward in the third side of equation (15) are meaningless weights that create nulls for antennas with propagation path variations while obtaining the combined gain for a specific transmitting antenna. On the other hand, the first column component obtained by equation (15) is a weight that obtains the combined gain for the first transmitting antenna while further suppressing interference for each subsequent antenna and their propagation path variations, and thus can be expected to have propagation path variation tolerance. Similarly, the extended propagation path response matrix G for the second transmitting antenna (2) (f) is obtained as shown in equation (16) below.

[0075]

number

[0076] Similarly, the MMSE weights applied to filter 84 on the first run are calculated using the following formula (17).

[0077]

number

[0078] Subsequently, by repeating the same process for each transmitting antenna, the null-space-extended weight matrix W'(f) is obtained as shown in equation (18) below.

[0079]

number

[0080] By applying the weight matrix W'(f) obtained by equation (18) in place of W(f) in equation (8), it is expected that the equalized signal "~" X(f) will be obtained with greater suppression of inter-stream interference caused by propagation path variations in the initial MMSE filter.

[0081] The resulting signal is then subjected to an inverse fast Fourier transform by the IFFT unit 85 to obtain a time-domain signal "~" x(k)∈C N×1 The signal is input to the BP decoding unit 87 and the likelihood is calculated. Subsequent processing is the same as in normal frequency domain equalization. That is, if repeated equalization is not performed, the BP decoding unit 87 performs a hard decision based on the likelihood obtained in the first processing to obtain the final decoding result. Alternatively, if repeated equalization is performed, a soft decision replica can be generated by the soft decision replica generation unit 89, and the soft decision replica can be subtracted from the received signal by the filter 84, and this process can be repeated according to equations (10) and (11). Note that G is used for generating the soft decision replica and generating the weights of the filter 84 from the second time onward. k n (f) can be any value of k. That is, in the examples so far, G 1 n (f) or G 2 n Either (f) may be used. However, G estimated at a time closer to the current time may be used. k n Using (f) may yield better accuracy and potentially greater profits.

[0082] (operation) Next, the processing of the wireless communication device 80aT will be described. Figure 6 is a flowchart showing the processing flow of the wireless communication device 80aT in the first embodiment. The processing in Figure 6 is performed after multiple signals (wireless frames shown in Figure 5) transmitted from the wireless communication device 80BS are received by each antenna of the wireless communication device 80aT. The CP removal units 81-1 to 81-3 remove CP from the signal received by the antenna (step S101). The CP removal units 81-1 to 81-3 output the CP-removed signal to the S / P units 82-1 to 82-3. The S / P units 82-1 to 82-3 perform a serial-to-parallel conversion on the CP-removed signal output from the CP removal units 81-1 to 81-3 (step S102). As a result, the S / P units 82-1 to 82-3 divide the CP-removed signal into data blocks of a specific time length. The S / P units 82-1 to 82-3 output each data block divided into specific time lengths to the FFT units 83-1 to 83-3.

[0083] The FFT sections 83-1 to 83-3 perform a Fast Fourier Transform on each data block, which is divided into specific time lengths and output from the S / P sections 82-1 to 82-3, thereby converting each data block into a frequency domain signal (frequency signal) (step S103). The FFT sections 83-1 to 83-3 output the frequency signals of each frequency component to different filters 84-1 to 84-F depending on the frequency component. The filters 84-1 to 84-F perform filtering on the frequency signals of each frequency component output from the FFT sections 83-1 to 83-3 (step S104). Specifically, the filters 84-1 to 84-F first use the frequency signals of each frequency component output from the FFT sections 83-1 to 83-3 to calculate the frequency domain propagation path response G 1 (f) and G 2 Estimate (f).

[0084] For example, filters 84-1 to 84-F obtain the frequency domain propagation path response G from the known signal contained in the first known signal section of the wireless frame. 1 (f) and the frequency domain propagation path response G obtained from the known signals included in the second known signal section.2 (f) and are estimated. Next, filters 84-1 to 84-F are used to estimate the propagation path response G in the frequency domain. 1 (f) and G 2 Using (f), the extended path response matrix for each transmitting antenna is calculated. For example, filters 84-1 to 84-F are the estimated frequency domain path response G 1 (f) and G 2 Using (f), the extended propagation path response matrix G for the first transmitting antenna is obtained. (1) (f) is calculated based on equation (14) above. Similarly, filters 84-1 to 84-F are used to calculate the estimated propagation path response G in the frequency domain. 1 (f) and G 2 Using (f), the extended propagation path response matrix G for the second transmitting antenna is obtained. (2) (f) is calculated based on equation (16) above. Filters 84-1 to 84-F are calculated in the same manner as the extended propagation path response matrices for the first and second transmitting antennas, and the extended propagation path response matrix G for the third transmitting antenna. (3) Calculate (f).

[0085] Subsequently, filters 84-1 to 84-F are used to obtain the extended propagation path response matrix G for each transmitting antenna obtained by calculation. (1) (f)~G (3) MMSE weights are calculated based on (f). For example, filters 84-1 to 84-F calculate the weight matrix W'(f) shown in equation (18) based on equations (15) and (17) above. Filters 84-1 to 84-F perform filtering by applying the calculated weight matrix W'(f) in place of W(f) in equation (8). For example, filter 84-1 performs filtering by applying the calculated weight matrix W'(1) in place of W(1) in equation (8). Similarly, filter 84-2 performs filtering by applying the calculated weight matrix W'(2) in place of W(2) in equation (8). Similarly, filter 84-F performs filtering by applying the calculated weight matrix W'(F) in place of W(F) in equation (8).

[0086] The IFFT sections 85-1 to 85-3 perform an inverse fast Fourier transform on the frequency signals of each frequency component that have been filtered (MMSE equalization) in filters 84-1 to 84-F, respectively, thereby converting the frequency signals of each frequency component into time-domain signals (step S105). The IFFT sections 85-1 to 85-3 output the time-domain signals to the P / S sections 86-1 to 86-3. The P / S sections 86-1 to 86-3 perform a parallel-serial conversion on the time-domain signals converted in the IFFT sections 85-1 to 85-3 (step S106). The P / S sections 86-1 to 86-3 output the signals after the parallel-serial conversion to the BP decoding sections 87-1 to 87-3.

[0087] The BP decoding units 87-1 to 87-3 calculate the likelihood by performing BP decoding on the parallel-to-serial converted signal output from the P / S units 86-1 to 86-3 (step S107). The BP decoding units 87-1 to 87-3 determine whether the process in step S107 has been repeated the specified number of times (step S108). If the BP decoding units 87-1 to 87-3 determine that the process in step S107 has been repeated the specified number of times (step S108-YES), the BP decoding units 87-1 to 87-3 output the calculated likelihood to the hard determination units 88-1 to 88-3. The hard determination units 88-1 to 88-3 obtain the decoding result by performing a hard determination on the likelihood calculated by the BP decoding units 87-1 to 87-3 (step S109).

[0088] On the other hand, if the BP decoding units 87-1 to 87-3 determine that the process in step S107 has not been repeated the specified number of times (step S108-NO), the BP decoding units 87-1 to 87-3 output the calculated likelihood to the soft-decision replica generation unit 89. The soft-decision replica generation unit 89 generates soft-decision symbols based on the likelihood output from the BP decoding units 87-1 to 87-3, and converts the generated soft-decision symbols into the frequency domain by performing a Fast Fourier Transform on them. As a result, the soft-decision replica generation unit 89 generates soft-decision replicas of the interference signal in the frequency domain for each antenna (step S110). The soft-decision replica generation unit 89 outputs the generated soft-decision replicas of the interference signal for each antenna to the filters 84-1 to 84-F.

[0089] Filters 84-1 to 84-F perform filtering by subtracting the soft-decision replica of the interference signal output from the soft-decision replica generation unit 89 from the received signal (frequency domain signal output from the FFT unit 83), and then multiplying by the MMSE weight (step S111). Subsequently, the processing from step S105 onwards is executed.

[0090] Note that the process shown in Figure 6 is the process flow when repeated equalization is performed, but the present invention can also be applied when repeated equalization is not performed. In this case, the process in step S108 shown in Figure 6 is not performed, and the process in step S109 is performed after the process in step S107.

[0091] In the MIMO system configured as described above, the wireless communication devices 80T and 80Ta include multiple antennas that receive multiple wireless frames of the same frequency transmitted from the wireless communication device 80BS, multiple Fourier transform units that convert the wireless frames received by each of the multiple antennas into the frequency domain at predetermined intervals, and multiple filters 84-1 to 84-3 that estimate the path response in each of the multiple frequency domains using multiple known signals present in the wireless frame and perform equalization processing based on the estimated multiple path responses.

[0092] Thus, in frequency domain MIMO equalization, which processes data blocks of a specific time length, one null is formed for each stream using a single known signal to suppress inter-stream interference. In contrast, in this invention, multiple nulls are formed from multiple known signals within the fluctuating propagation path. This allows for efficient suppression of inter-stream interference, thereby improving transmission capacity. Consequently, even when propagation path fluctuations are rapid, inter-stream interference can be suppressed in frequency domain MIMO equalization, which processes data blocks of a specific time length. As a result, wireless transmission quality can be improved.

[0093] Furthermore, Reference 1, shown below, describes an invention relating to a null space extension technique for separating and suppressing inter-stream interference on the transmitting (base station equipment) side when downlink transmission is performed from the same base station equipment to multiple terminal devices. On the other hand, the present invention differs in that the receiving wireless communication device (e.g., terminal equipment) processes the signal transmitted from the transmitting wireless communication device (e.g., base station equipment) by converting it into the frequency domain and forming multiple nulls. (Reference 1: Japanese Patent Publication No. 2016-136706)

[0094] (Second embodiment) In the first embodiment, the case where there is one wireless communication device as a terminal device was described as an example. On the other hand, MIMO transmission capacity can be improved by the cooperation of multiple terminal devices. For example, as shown in Reference 1 below, the number of antenna elements equipped in a terminal device can be virtually increased by transferring a signal received by one terminal device to another terminal device using a frequency different from the frequency at which the signal from the base station is transmitted. The present invention is also applicable in such a configuration. Therefore, in the second embodiment, the case where there are multiple wireless communication devices as terminal devices will be described. (Reference 1: Eiichi Murata, Daisuke Umehara, "MIMO Receiving Technology with Terminal Collaboration - Theory and Application to Mobile Devices -", ​​IEICE Fundamentals Review, Institute of Electronics, Information and Communication Engineers, January 2022, Vol. 15, No. 3, pp. 210-219)

[0095] Figure 7 is a schematic diagram showing an example configuration of a wireless communication system according to the second embodiment. The wireless communication system shown in Figure 7 is the so-called MIMO system described above. The wireless communication system comprises a base station device 80BS and a plurality of terminal devices 80aT-1 to 80aT-6. The base station device 80BS has the same configuration as the wireless communication device 80BS shown in Figure 1 with respect to MIMO transmission, and is therefore indicated with the same designation "80BS" as the wireless communication device 80BS in Figure 1. Accordingly, the base station device 80BS transmits three signals (downlink signals of three streams) from three antennas, similar to the wireless communication device 80BS in Figure 1.

[0096] Each of the terminal devices 80aT-1 to 80aT-6 has the same configuration, and hereafter, any one of the terminal devices 80aT-1 to 80aT-6 will be referred to as terminal device 80aT. In Figure 8, six terminal devices 80aT-1 to 80aT-6 are shown as an example, but the number of terminal devices 80aT-1 to 80aT-6 must be greater than or equal to the number of streams that the base station device 80BS intends to transmit, assuming that each terminal device 80aT-1 to 80aT-6 is equipped with one antenna element as an antenna element to receive signals from the base station device 80BS, and that the terminal device 80aT performs signal stream separation processing based on the ZF standard. Therefore, the number of terminal devices 80aT-1 to 80aT-6 should be greater than or equal to the number of streams that the base station device 80BS intends to transmit. Furthermore, when performing weight calculation based on additional channel vectors as described in the embodiments of the present invention, the Moore-Penrose generalized inverse can be calculated by having a number of machines equal to or greater than the number of signal streams plus the number of additional channel vectors.

[0097] (Example configuration of the receiving wireless communication device) Figure 8 shows an example configuration of a wireless communication device 80bT that performs signal stream separation processing by repeated equalization in the frequency domain. The wireless communication device 80bT comprises a CP removal unit 81, an S / P unit 82, an FFT unit 83, a filter 84, an IFFT unit 85, a P / S unit 86, a BP decoding unit 87, a hard determination unit 88, a soft determination replica generation unit 89, and a frequency conversion unit 90. The wireless communication device 80bT differs in configuration from the wireless communication device 80aT in that it further comprises a frequency conversion unit 90. Here, each antenna of the wireless communication device 80bT is called antenna NR, NR1, ... NR M Let M be an integer greater than or equal to 1.

[0098] The signal transmitted from the base station device 80BS is received by the antenna NR of the wireless communication device 80bT. The signal received by the antenna NR is processed by the CP removal unit 81 and subsequent units. The signal received by the antenna NR is also output to the frequency conversion unit 90. The frequency conversion unit 90 converts the input signal to a transmission frequency and transmits it to the other wireless communication device 80bT via the antenna NT. For example, the frequency conversion unit 90 may convert the input signal to a frequency higher than the frequency at which the base station device 80BS sends signals to the wireless communication device 80bT and transmit it to the other wireless communication device 80bT. Note that the frequency converted by the frequency conversion unit 90 is not limited to a frequency higher than the frequency at which the base station device 80BS sends signals to the wireless communication device 80bT. Also, since the signal transmitted by the frequency conversion unit 90 is the signal transmitted from the base station device 80BS itself, it contains multiple known signals.

[0099] Each wireless communication device 80bT receives signals transmitted from other wireless communication devices 80bT via antennas NR1~NR M Received via each antenna NR1,…NR M The signals received are processed by the CP removal unit 81 and beyond. That is, each antenna NR1, ... NR MThis is an antenna configured to receive signals at frequencies transmitted by other wireless communication devices 80bT. Thus, each wireless communication device 80bT has one antenna to receive signals transmitted from the base station device 80BS, and it forwards the signals received by this antenna to other wireless communication devices 80bT and processes them within its own device. Furthermore, each wireless communication device 80bT can virtually increase the number of antenna elements that receive signals from the base station device 80BS by receiving signals forwarded by other wireless communication devices 80bT. The processing after receiving a signal in each wireless communication device 80bT is the same as in the first embodiment.

[0100] In the wireless communication system of the second embodiment configured as described above, the wireless communication device 80bT that performs equalization processing does not need to receive all signals from the base station device 80BS, but performs equalization processing using signals transferred from other wireless communication devices 80bT and some signals from the base station device 80BS. When the wireless communication device 80bT is a terminal device, it is difficult to realize an antenna with as many as 100 elements due to limitations on the size and power consumption of the device. In contrast, in the configuration shown in the second embodiment, by coordinating multiple wireless communication devices 80bT, it is possible to virtually increase the number of antenna elements while limiting the number of antenna elements that each wireless communication device 80bT has for direct communication with the base station device 80BS, making it easier to realize the wireless communication device 80bT as an actual device. Furthermore, by applying null space extension technology to each of the wireless communication devices 80bT, it becomes possible to suppress inter-stream interference with respect to radio propagation path fluctuations.

[0101] Furthermore, the above-described processing may be performed by recording the programs for realizing the functions of the wireless communication devices 80BS, 80T, 80aT, and 80bT in the first and second embodiments described above onto a computer-readable recording medium, loading the programs recorded on this recording medium into a computer system, and executing them. Here, "computer system" includes hardware such as an OS (Operating System) and peripheral devices. "Computer system" also includes a WWW system equipped with a homepage provisioning environment (or display environment). "Computer-readable recording medium" refers to portable media such as flexible disks, magneto-optical disks, ROMs (Read Only Memory), CD-ROMs, and storage devices such as hard disks built into a computer system. Moreover, "computer-readable recording medium" also includes volatile memory (RAM) inside a computer system that acts as a server or client when a program is transmitted via a network such as the Internet or a communication line such as a telephone line, which holds the program for a certain period of time.

[0102] Furthermore, the above program may be transmitted from a computer system that stores the program in a memory device or the like to another computer system via a transmission medium or by transmission waves within the transmission medium. Here, the "transmission medium" for transmitting the program refers to a medium that has the function of transmitting information, such as a network (communication network) such as the Internet or a communication line (communication line) such as a telephone line. Also, the above program may be for the purpose of realizing only a part of the functions described above. Furthermore, it may be a so-called differential file (differential program) that can realize the above functions in combination with a program already recorded in the computer system.

[0103] While embodiments of this invention have been described in detail above with reference to the drawings, the specific configuration is not limited to these embodiments and includes designs and the like that do not depart from the spirit of this invention. [Explanation of Symbols]

[0104] 80BS…Wireless communication device, 80T…Wireless communication device, 81, 81-1~81-3…CP removal unit, 82, 82-1~82-3…S / P unit, 83, 83-1~83-3…FFT unit, 84, 84-1~84-3…Filter, 85, 85-1~85-3…IFFT unit, 86, 86-1~86-3…P / S unit, 87, 87-1~87-3…BP decoding unit, 88, 88-1~88-3…Hard judgment unit, 89…Soft judgment replica generation unit, 90…Frequency conversion unit

Claims

1. Multiple Fourier transform units that convert the wireless frames received by each of the multiple antennas that receive multiple wireless frames of the same frequency transmitted from one or more other wireless communication devices into the frequency domain at predetermined intervals, Multiple filters that estimate the path response in each of multiple frequency domains using multiple known signals present within the wireless frame, and perform equalization processing based on the estimated path responses, A receiving device equipped with the following features.

2. The aforementioned multiple filters are, A weight matrix is ​​generated based on the estimated multiple propagation path responses, and an equalization process is performed by multiplying the generated weight matrix by the received signal in the frequency domain. The receiving device according to claim 1.

3. When the multiple filters perform an equalization process a predetermined number of times, the system further includes a replica generation unit that generates a frequency domain replica signal based on the signal after the equalization process, The aforementioned multiple filters are, Equalization is performed by subtracting the replica signal in the frequency domain from the received signal in the frequency domain and then multiplying it by the weight matrix. The receiving device according to claim 2.

4. The aforementioned one or more other wireless communication devices include a base station device and other terminal devices. The aforementioned multiple antennas are The base station device and the other terminal devices each receive wireless frames transmitted from them. The aforementioned multiple filters are, The base station device receives wireless frames from the multiple antennas and from each of the other terminal devices, and equalization processing is performed based on these frames. A receiving device according to any one of claims 1 to 3.

5. The system further includes a frequency conversion unit that converts the frequency of the radio frame transmitted from the base station device, received by some of the plurality of antennas, and transmits it to the other terminal device. The receiving device according to claim 4.

6. A wireless communication system comprising a first wireless communication device and one or more second wireless communication devices, The first wireless communication device is Multiple wireless frames of the same frequency, each containing multiple known signals, are transmitted to one or more second wireless communication devices. The one or more second wireless communication devices are A plurality of Fourier transform units that convert the plurality of wireless frames transmitted from the first wireless communication device, or the wireless frames received by each of the plurality of antennas that receive at least one of the plurality of wireless frames transmitted from the first wireless communication device and wireless frames transmitted from another second wireless communication device, into the frequency domain at predetermined intervals, Multiple filters that estimate the path response in each of multiple frequency domains using multiple known signals present within the wireless frame, and perform equalization processing based on the estimated path responses, A wireless communication system equipped with [the following features].

7. Each of the multiple wireless frames of the same frequency, received by each of the multiple antennas and transmitted from one or more other wireless communication devices, is converted to the frequency domain at predetermined intervals. Using multiple known signals present within the aforementioned wireless frame, the propagation path response in each of the multiple frequency domains is estimated. Equalization is performed based on the estimated multiple propagation path responses. Reception method.

Images