Wireless signal processing system and wireless signal processing method

Abstract

To provide a radio signal processing system and a radio signal processing method which implement iterative MUD capable of shortening a computation time.SOLUTION: A radio signal processing system comprises: error calculation means for generating two or more patterns of estimated signals for which a waveform is estimated for each of transmitted signals transmitted from two or more second communication stations and for respectively calculating an estimation error based on a difference between the received signals and a sum of the estimated signals included in the respective generated patterns; selection means for selecting two or more patterns of the estimation signals based on the respective estimation errors calculated by the error calculation means; probability calculation means for calculating a probability distribution indicating a distribution of probabilities that the estimated signals are the transmission signals based on the estimation errors with respect to the two or more patterns selected by the selection means; and estimation means for estimating information included in the transmission signals based on the probability distribution calculated by the probability calculation means.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a wireless signal processing system and a wireless signal processing method. [Background technology]

[0002] Massive connectivity, particularly on the uplink (UL), is one of the key features of Beyond 5G. Furthermore, IoT-related technologies have developed rapidly in recent years, and it is projected that by 2030, the number of IoT devices connected to the network will reach hundreds of billions. Therefore, to support IoT applications in Beyond 5G, there is a need to enhance the capabilities of massive machine-type communication (mMTC) on the UL. Non-orthogonal Multiple Access (NOMA) is attracting attention as a promising technology for strengthening mMTC. In particular, uplink NOMA (UL-NOMA) systems exhibit high spectral efficiency by allowing multiple devices to share the same radio resource block. At this time, multi-user detection (MUD), which estimates information contained in the transmitted signals sent by each device from the received signals, is essential on the base station side. In particular, iterative multi-user detection does not require large differences in received power between devices, thus enabling a larger number of connections than methods such as successive interference cancellation (SIC). For this reason, techniques for detecting repeated multi-users, such as those described in Non-Patent Document 1, are attracting attention.

[0003] Non-Patent Document 1 proposes an iterative MUD using log-likelihood ratios. In the technology disclosed in Non-Patent Document 1, the log-likelihood ratio for each device is calculated from the received signal in the MUD block, and these are input to the decoder as prior information for error correction. In the technology disclosed in Non-Patent Document 1, the results of this error correction are fed back to the MUD block, and the iterative process is performed until the transmitted data from all devices is error-free, or up to a predetermined maximum number of times. [Prior art documents] [Non-patent literature]

[0004] [Non-Patent Document 1] HVPoor, "Iterative multiuser detection," IEEE Signal Processing Magazine, vol. 21, no. 1, pp. 81-88, 2004. [Overview of the Initiative] [Problems that the invention aims to solve]

[0005] On the other hand, the computational efficiency of iterative MUD is O(M) from the modulation order M and the number of devices K. K Since it is given by ), the MUD part, which calculates the log-likelihood ratio from the received signal for each device, requires a large amount of computation time.

[0006] Consequently, the technology disclosed in Non-Patent Document 1 requires a significant amount of computation time when used as an iterative MUD for large-scale connections. Therefore, a problem with the technology disclosed in Non-Patent Document 1 is that it is not possible to implement UL-NOMA using an iterative MUD for large-scale connections.

[0007] The present invention was derived to solve these problems and aims to provide a wireless signal processing system and wireless signal processing method that realize an iterative MUD capable of reducing computation time. [Means for solving the problem]

[0008] The wireless signal processing system according to the first invention is a wireless signal processing system that estimates information contained in a transmission signal based on a received signal received by a first communication station from a transmission signal transmitted from two or more second communication stations, and is characterized by comprising: a selection means for selecting two or more estimated signal patterns for each transmission signal transmitted from the two or more second communication stations based on an estimation error corresponding to the waveform of the transmission signal estimated from the estimated signal pattern and the received signal; a probability calculation means for calculating a probability distribution showing the probability distribution of the probability that the estimated signal is the transmission signal based on the estimation error for the two or more patterns selected by the selection means; and an estimation means for estimating information contained in the transmission signal based on the probability distribution calculated by the probability calculation means.

[0009] The wireless signal processing system according to the second invention is characterized in that, in the first invention, the estimation means calculates a log-likelihood ratio indicating the plausibility that the estimated signal is the transmission signal based on the probability distribution calculated by the probability calculation means, and estimates the information contained in the transmission signal based on the sign of the calculated log-likelihood ratio.

[0010] The wireless signal processing system according to the third invention is, in the first invention, wherein the selection means indicates the estimated signal by spin, and based on the spin kue The method is characterized by calculating the estimation error using an energy function that represents energy, and selecting two or more patterns of the estimated signal based on the calculated estimation error.

[0011] The wireless signal processing system according to the fourth invention is characterized in that, in the first invention, the selection means calculates the estimation error using quantum annealing and selects two or more patterns of the estimated signal based on the calculated estimation error.

[0012] The wireless signal processing system according to the fifth invention is a wireless signal processing system that estimates information included in the transmission signal based on a received signal r(i) received by a first communication station from two or more second communication stations. In the system, an estimated signal x that estimates the waveform of the transmission signal k (i) is the spin variable z of the spin shown by equation (1) k (i) based on the estimation error E shown by equation (2) MUD (z(i)), based on the estimated signal x that estimates the waveform of each transmission signal transmitted from two or more of the second communication stations k (i) of the spin variable z k (i), a selection means for selecting two or more patterns z(i) of the spin variable z(i), and a set Z of two or more patterns z(i) selected by the selection means QA (i), and the estimation error E MUD (z(i)), based on the estimated signal x k is a probability distribution Pr indicating the distribution of the probability that the estimated signal x is the transmission signal Approx [r(i)|x(i)] is calculated using equation (3) by a probability calculation means, and based on the probability distribution Pr Approx [r(i)|x(i)] includes an estimation means for estimating information included in the transmission signal.

[0013]

Equation

Equation

[0014]

number

[0015] The wireless signal processing method according to the sixth invention is a wireless signal processing method that estimates information contained in a transmission signal based on a received signal received by a first communication station from a transmission signal transmitted from two or more second communication stations, and is characterized in that it causes a computer to execute the following steps: a selection step of selecting two or more estimated signal patterns that estimate the waveform of each transmission signal transmitted from the two or more second communication stations based on an estimation error corresponding to the waveform of the transmission signal and the received signal; a probability calculation step of calculating a probability distribution that shows the probability distribution of the probability that the estimated signal is the transmission signal based on the estimation error for the two or more estimated signal patterns selected in the selection step; and an estimation step of estimating information contained in the transmission signal based on the probability distribution calculated in the probability calculation step. [Effects of the Invention]

[0016] According to the first to sixth inventions, the probability calculation means calculates a probability distribution showing the distribution of the probability that an estimated signal is a transmitted signal, based on the estimation errors of two or more patterns of estimated signals selected by the selection means. As a result, since only the estimated signal patterns determined by the determination means are used as samples for calculation, computation time can be reduced even when used as an iterative MUD for large-scale connections.

[0017] In particular, according to the second invention, a log-likelihood ratio indicating the plausibility that the estimated signal is the transmitted signal is calculated based on a probability distribution, and the transmitted signal is estimated based on the sign of the calculated log-likelihood ratio. This makes it possible to determine the plausibility that the estimated signal is the transmitted signal according to the sign of the log-likelihood ratio.

[0018] In particular, according to the third invention, the selection means calculates the estimation error using an energy function that represents energy. As a result, only the more appropriate estimation signal patterns are calculated, which reduces computation time even when used as an iterative MUD for large-scale connections.

[0019] In particular, according to the fourth invention, the selection means calculates the estimation error using quantum annealing. This makes it possible to select an appropriate estimated signal pattern in a shorter computation time by using quantum annealing. [Brief explanation of the drawing]

[0020] [Figure 1] Figure 1 is a schematic diagram of the wireless signal processing system in this embodiment. [Figure 2] Figure 2 is a schematic diagram showing an example of the configuration of a multi-user detector in this embodiment. [Figure 3] Figure 3 shows the processing flow of the wireless signal processing system. [Figure 4] Figure 4(a) is a graph showing the probability distribution when no pattern of estimated signal is selected. Figure 4(b) is a graph showing the probability distribution when the wireless signal processing system of this embodiment is used. [Figure 5]Figure 5(a) is a graph showing the bit error rate for the first run in this embodiment. Figure 5(b) is a graph showing the bit error rate for the second run in this embodiment. Figure 5(c) is a graph showing the bit error rate for the third run in this embodiment. Figure 5(d) is a graph showing the bit error rate for the fourth run in this embodiment. Figure 5(e) is a graph showing the bit error rate for the fifth run in this embodiment. Figure 5(f) is a graph showing the bit error rate for the sixth run in this embodiment. [Figure 6] This graph shows a comparison of the trends in mutual information volume for each of the second communications stations. [Figure 7] This graph shows a comparison of calculation times. [Modes for carrying out the invention]

[0021] The wireless signal processing system 100 to which the present invention is applied will be described in detail below with reference to the drawings.

[0022] Figure 1 is a schematic diagram of the wireless signal processing system 100 in this embodiment. As shown in Figure 1, the wireless signal processing system 100 comprises a base station 1, a user terminal 2 connected to the base station 1, and a quantum annealing machine 3. The wireless signal processing system 100 may also include two or more user terminals (2a to 2c). The wireless signal processing system 100 estimates the transmitted signals sent from the two or more user terminals 2 based on the received signals received by the base station 1, for example. Communication between the base station 1 and the user terminals 2 uses wireless communication, but is not limited to this, and may also use wired communication.

[0023] User terminal 2 is a device (Device:dev) that transmits a transmission signal. User terminal 2 consists of a wireless communication-capable terminal device used by a user, such as a notebook personal computer (PC), a mobile terminal, a smartphone, a tablet terminal, or a wearable terminal. User terminal 2 includes an encoder (ENC) 21 for encoding a digital signal and a modulator (MOD) 22, such as QPSK (quadrature phase shift keying), connected to the encoder 21 for modulating the encoded digital signal. User terminal 2 is, for example, a second communication station that transmits a transmission signal to base station 1, but it may also be a first communication station that receives a received signal. Furthermore, user terminal 2 may include an interleaver (not shown) between the encoder 21 and the modulator 22 to rearrange the order of the codeword bits.

[0024] The quantum annealing machine 3 is a computer that utilizes the principle of quantum annealing to select an estimated signal pattern that indicates the estimated waveform of the transmitted signals sent from each of the second communication stations from the received signal received by the first communication station. Quantum annealing is a method that uses quantum fluctuations to find the optimal or approximate solution to a combinatorial optimization problem. The annealing time of quantum annealing may be, for example, 20 μs or less. The quantum annealing machine 3 may also be included in the base station 1. Furthermore, the quantum annealing machine 3 is not limited to a computer that utilizes the principle of quantum annealing, but may be a computer with any computing means. The quantum annealing machine 3 may be designed, for example, to solve the energy function for the Ising model given by equation (4).

[0025]

number

[0026] Base station 1 is a communication station (BS) that acts as a wireless access point between multiple user terminals 2 and as an interface between the base station and communication lines such as the Internet. In other words, base station 1 is a relay means that enables user terminals 2 to send and receive data to and from communication lines such as the Internet. Base station 1 controls wireless signal processing. Base station 1 may use electronic equipment such as a personal computer (PC) to control wireless signal processing, or it may use electronic equipment such as a smartphone, tablet terminal, wearable terminal, IoT (Internet of Things) device, or single-board computer. Base station 1 may be a first communication station that receives received signals or a second communication station that transmits transmitted signals.

[0027] Base station 1 includes a multi-user detector (MUD) 10 and a decoder (DEC) 11 for decoding data connected to the multi-user detector 10. Base station 1 may also include multiple decoders 11 connected to the multi-user detector 10. The multi-user detector 10 estimates the information contained in the transmission signals sent from each of the second communication stations. The multi-user detector 10 may be, for example, a soft-input soft-output (SISO) multi-user detector. Furthermore, if user terminal 2 includes an interleaver (not shown) between encoder 21 and modulator 22 for rearranging the codeword bits, the output from the multi-user detector 10 is input to the decoder 11 via a deinterleaver (not shown), and the output from the decoder 11 is input to the multi-user detector 10 via an interleaver (not shown).

[0028] Figure 2 is a schematic diagram showing an example of the configuration of the multi-user detector 10 in this embodiment. The multi-user detector 10 comprises a probability distribution calculation unit 101, a likelihood ratio calculation unit 102 connected to the probability distribution calculation unit 101 and the decoder 11, and a prior probability calculation unit (APP) 103 connected to the likelihood ratio calculation unit 102 and the decoder 11.

[0029] The probability distribution calculation unit 101 calculates a probability distribution showing the probability distribution of the probability that the estimated signal, which represents the estimated waveform of the transmitted signal, is the transmitted signal, based on the received signal, for example, transmitted from the second communication station and received by the first communication station. The probability distribution calculation unit 101 generates two or more estimated signal patterns, for example, by estimating the waveforms of each transmitted signal transmitted from two or more second communication stations, and calculates an estimation error for each generated pattern based on the difference between the sum of the estimated signals included in each pattern and the received signal. The probability distribution calculation unit 101 also selects two or more estimated signal patterns based on the calculated estimation errors. The probability distribution calculation unit 101 then calculates a probability distribution based on the estimation errors of the two or more selected estimated signal patterns. The probability distribution calculation unit 101 may also be connected to a quantum annealing machine 3. In this case, the probability distribution calculation unit 101 may use the quantum annealing machine 3 to select an estimated signal pattern based on the estimation error.

[0030] The likelihood ratio calculation unit 102 calculates the log-likelihood ratio based on the probability distribution calculated by the probability distribution calculation unit 101. The likelihood ratio calculation unit 102 transmits the calculated log-likelihood ratio to the decoder 11.

[0031] The prior probability calculation unit (APP) 103 calculates a prior probability P(x) for newly calculating the log-likelihood ratio based on the log-likelihood ratio transmitted from the decoder 11, and transmits the calculated prior probability P(x) to the likelihood ratio calculation unit 102.

[0032] Next, the operation of the wireless signal processing system 100 will be explained using Figure 3. First, in step S110, the second communication station transmits a transmission signal. The first communication station receives a reception signal based on transmission signals transmitted from two or more second communication stations. The transmission signal is a signal transmitted from the second communication station. If the number of bits of information contained in the transmission signal is D, then the bits of length D are {b k (0),...,b k The information is (D-1). The received signal is a signal received by the first communication station from two or more second communication stations that have transmitted signals. The received signal is a signal that has been interfered with by the first communication station due to data collisions between the transmitted signals when it is received by the first communication station. Here, the first communication station may be, for example, base station 1, but is not limited to this, and may be any communication station. Also, the second communication station may be, for example, user terminal 2, but is not limited to this, and may be any communication station.

[0033] In step S110, the encoder 21 of the k-th second communication station transmits a cyclic redundancy check (CRC) code of a specific number of bits to the bit sequence b of the information of the transmitted signal in order to perform error detection in step S150, which will be described later. k Assigned to the error correction codeword bit string c of length N bits. k ={c k (0), ..., c k Encoded to (N-1). Next, in step S110, the modulator 22 encodes the error-corrected codeword bit sequence c k The transmitted signal a k x k It modulates to this. The second communications station transmits this signal a k x k The signal is transmitted to the first communication station, and the first communication station receives the transmitted signals from multiple second communication stations as received signals. In step S110, the transmitted signal transmitted from the k-th second communication station is the channel coefficient h which indicates the radio wave propagation characteristics between the k-th second communication station and the first communication station. k The signals are amplified twice, and the resulting interference between the amplified transmitted signals becomes the received signal, which is then received by the first communication station. Here, k represents the number of the second communication station.

[0034] Alternatively, this received signal may be calculated using the formula shown in equation (5).

[0035]

number

[0036] Next, in step S120, the probability distribution calculation unit 101 selects two or more estimated signal patterns, which are estimated waveforms of each transmission signal transmitted from two or more second communication stations, based on the estimated signal pattern obtained by estimating the waveform of the transmission signal and the estimation error corresponding to the received signal. The estimated signal is a signal obtained by estimating the waveform of each transmission signal transmitted by each second communication station. The probability distribution calculation unit 101 may estimate the estimated signal using any method. The estimated signal sample is a sample of the estimated signal for each transmission signal transmitted from two or more second communication stations. The estimation error is, for example, h for each estimated signal included in the estimated signal sample. k The error may also be based on the difference between the sum of the multiplied values ​​and the received signal. The estimation error may be shown, for example, by equation (6).

[0037]

number

[0038] In step S120, the probability distribution calculation unit 101 selects two or more estimated signal patterns based on the estimation error. The probability distribution calculation unit 101 may, for example, select the estimated signal pattern that minimizes the estimation error, or two or more patterns that are the approximate solution when the estimation error is minimized. The probability distribution calculation unit 101 may, for example, select two or more estimated signal patterns that minimize the estimation error as much as possible. By selecting two or more estimated signal patterns in step S120, it becomes possible to generate multiple estimated signal patterns as samples, enabling more accurate estimation of the transmitted signal. The multi-user detector 10 may also select a pattern in which the estimation error is below a threshold.

[0039] Furthermore, in step S120, the multi-user detector 10 may select a pattern of the estimated signal estimated by the quantum annealing machine 3 based on the received signal received in step S110. In such a case, for example, when QPSK modulation is used, the estimated signal is calculated by equation (1) with spin as a variable.

[0040]

number

[0041] From equations (1) and (6) above, the estimation error can be expressed by the energy function (2), which represents the energy based on spin.

[0042]

number

[0043] Next, in step S130, the probability distribution calculation unit 101 calculates a probability distribution showing the probability distribution of the probability that the estimated signal is the transmission signal, based on the estimation error for each of two or more samples of the estimated signal selected in step S120. The probability distribution may also be the probability distribution of the probability that the received signal is the received signal received in step S110 when the estimated signal of the sample selected in step S120 was the transmission signal. The probability distribution is the channel coefficient h k Assuming that it is fixed within the symbol, the standard deviation σ of additive white Gaussian noise w Using the channel coefficient h k and estimated signal x k It is obtained as a two-dimensional Gaussian distribution with the product of (i) as the mean. In step S130, for example, the probability distribution calculation unit 101 calculates the probability distribution using equation (7) based on the estimation error for each of two or more samples of the estimated signal selected in step S120.

number

[0044] Furthermore, for example, the probability distribution calculation unit 101 calculates the set of samples Z for the spin combination pattern obtained from the quantum annealing machine 3 for the i-th received signal r(i). QA (i) and the estimated error E calculated in step S130 MUD Based on (z(i)), the probability distribution shown in equation (3) may be calculated.

number

[0045] Alternatively, if a sample of the transmission signal is selected using the quantum annealing machine 3 in step S130, the probability distribution may be calculated using equation (8).

[0046]

number

[0047] Next, in step S140, the likelihood ratio calculation unit 102 calculates a log-likelihood ratio indicating the validity that the estimated signal is the transmission signal based on the probability distribution calculated in step S130. Also, in step S140, the multi-user detector 10 may calculate the log-likelihood ratio using equation (9).

[0048] [Number] Here, L(c k (n)) MUD indicates the log-likelihood ratio, and Pr[c k (n)=1|r] indicates the probability that the codeword bit c k (n) becomes 1 when the received signal vector r is received. Also, Pr[c k (n)=0|r] indicates the probability that the codeword bit c k (n) becomes 0 when the received signal vector r is received.

[0049] Also, the first term on the right side of equation (9) is extrinsic information, and is denoted as L(c k (n)) e MUD [[ID=and]] is used. Also, the second term on the right side of equation (9) is a priori information fed back from the decoder 11, and is denoted as L(c k (n)) a MUD k e DEC ) k MUD is set to all zeros. The log-likelihood ratio L(c k ) MUD can be calculated by equation (10).

[0050] [Number] Here, X 1 k,n and X 0 k,n are the subsets of X of x, where the pattern of the estimated signal x k is described as the vector x, and c k (n) is 1 or 0. Pr[x] is obtained from the prior information L(c k ) a MUD fed back from the decoder 11. The extrinsic information is obtained by subtracting L(c k ) MUD from L(c k ) a MUD and passed to the decoder 11 as the prior information L(c k ) a DEC .

[0051] Next, in step S150, the multi-user detector 10 estimates the information included in the transmission signal based on the log-likelihood ratio calculated in step S140. In step S150, the multi-user detector 10 estimates the information included in the transmission signal based on, for example, the sign of the log-likelihood ratio. In such a case, for example, for the bit sequence b k output from the decoder 11, error detection by CRC is performed on the code given in step S110, and the estimated signal of the bit sequence b k determined to be error-free may be estimated as the transmission signal. Also, for the transmission bit sequence b k of the estimated signal determined to be error-free, remapping is performed using the communication channel coefficient h k and the propagation delay time δ k to create a received signal replica r‘ k = h k · x k (i - δ k ). This received signal replica r‘ kBy subtracting from the received signal vector r, interference between the second communication station can be suppressed and removed. If an error is detected in step S150, the decoder 11 sends prior information L(c) to the prior probability calculation unit (APP) 103 of the multi-user detector 10. k ) a MUD The data is fed back, and the likelihood exchange process is performed again in step S140 to calculate the log-likelihood ratio. The likelihood exchange process between the multi-user detector 10 and the decoder 11 is performed using the bit sequence b of the entire estimated signal. k The process is repeated until it is determined that there are no errors, or up to a predetermined upper limit of likelihood exchanges. By performing each of the steps described above, the operation of the wireless signal processing system 100 is completed. This makes it possible to reduce computation time even when used as an iterative MUD for large-scale connections.

[0052] Next, we will describe the experimental results of estimating a transmitted signal using the wireless signal processing system 100 of this embodiment. Figure 4(a) is a graph showing the probability distribution when no pattern of estimated signal is selected. Figure 4(b) is a graph showing the probability distribution when the wireless signal processing system of this embodiment is used. In Figures 4(a) and 4(b), the vertical axis represents probability, and the horizontal axis represents estimation error. The probability distribution shown in Figure 4(a), when no pattern of estimated signal is selected, is a probability distribution calculated based on all possible patterns of the estimated signal without performing, for example, step S120 in this embodiment. As shown in Figures 4(a) and 4(b), there is no difference in the comparison of the probability distributions, so it can be considered that there is no difference in accuracy between the probability distribution when using the wireless signal processing system 100 of this embodiment and the probability distribution calculated based on all possible patterns of the estimated signal.

[0053] Figure 5(a) is a graph showing the bit error rate for the first time in this embodiment. Figure 5(b) is a graph showing the bit error rate for the second time in this embodiment. Figure 5(c) is a graph showing the bit error rate for the third time in this embodiment. Figure 5(d) is a graph showing the bit error rate for the fourth time in this embodiment. Figure 5(e) is a graph showing the bit error rate for the fifth time in this embodiment. Figure 5(f) is a graph showing the bit error rate for the sixth time in this embodiment. The bar graphs in Figures 5(a), 5(b), 5(c), 5(d), 5(e), and 5(f) show the error rate for each second communication station. Figures 5(a), 5(b), 5(c), 5(d), 5(e), and 5(f) show the bit error rate for each number of likelihood exchange processes in steps S140 and S150, respectively. As shown in Figures 5(a), 5(b), 5(c), 5(d), 5(e), and 5(f), the bit error rate decreases with each iteration of the likelihood exchange process, and after six likelihood exchange processes, all second communication stations were able to correctly detect the information contained in the transmitted signal.

[0054] Figure 6 is a graph showing a comparison of the changes in mutual information for each second communication station. In Figure 6, the vertical axis represents the mutual information, and the horizontal axis represents the number of likelihood exchange processes. In Figure 6, QA shows the changes in mutual information when using this embodiment, and EXACT shows the changes in mutual information calculated based on the probability distribution calculated based on all estimated signal patterns without performing step S120 in this embodiment. As shown in Figure 6, it can be considered that there is no difference in accuracy between the case where the wireless signal processing system 100 in this embodiment is used and the case where it is calculated based on all estimated signal patterns.

[0055] Figure 7 is a graph showing a comparison of calculation times. In Figure 7, the vertical axis represents the calculation time, and the horizontal axis represents the number of second communication stations. QA(Net) shows the calculation time when using this embodiment, and QA(Total) shows the calculation time considering all estimated signal turns without performing step S120 in this embodiment. As shown in Figure 7, the calculation time when using the wireless signal processing system 100 of this embodiment can be reduced to about one-tenth of the calculation time when considering all estimated signal patterns. From this, it is possible to reduce the calculation time without degrading accuracy even when used as an iterative MUD for large-scale connections.

[0056] While embodiments of the present invention have been described, these embodiments are presented as examples only and are not intended to limit the scope of the invention. Such novel embodiments can be implemented in a variety of other forms, and various omissions, substitutions, and modifications can be made without departing from the spirit of the invention. These embodiments and their variations are included within the scope and spirit of the invention, as well as within the scope of the invention and its equivalents as described in the claims. [Explanation of symbols]

[0057] 1 base station 2 User terminals 3. Quantum Annealing Machine 10 Multi-user detector 11 Decoder 21 Encoders 22 Modulators 100 Wireless signal processing systems 101 Probability Distribution Calculation Unit 102 Likelihood Ratio Calculation Unit 103 Prior Probability Calculation Unit

Claims

1. In a radio signal processing system that estimates information contained in a transmitted signal based on a received signal received by a first communication station from a transmitted signal transmitted from two or more second communication stations, A selection means for selecting two or more estimated signal patterns, which are estimated from the waveforms of each transmitted signal transmitted from two or more of the second communication stations, based on the estimated signal pattern obtained by estimating the waveform of the transmitted signal and the estimation error corresponding to the received signal, A probability calculation means calculates a probability distribution showing the probability distribution of the estimated signal being the transmitted signal, based on the estimation error for two or more patterns selected by the selection means. The system includes estimation means for estimating information contained in the transmitted signal based on the probability distribution calculated by the probability calculation means. A wireless signal processing system characterized by the following.

2. The estimation means calculates a log-likelihood ratio indicating the plausibility that the estimated signal is the transmitted signal, based on the probability distribution calculated by the probability calculation means, and estimates the information contained in the transmitted signal based on the sign of the calculated log-likelihood ratio. A wireless signal processing system according to claim 1, characterized by the following:

3. The selection means represents the estimated signal in terms of spin, calculates the estimation error using an energy function that represents the energy based on the spin, and selects two or more patterns of the estimated signal based on the calculated estimation error. A wireless signal processing system according to claim 1, characterized by the following:

4. The selection means calculates the estimation error using quantum annealing and selects two or more patterns of the estimated signal based on the calculated estimation error. A wireless signal processing system according to claim 1, characterized by the following:

5. In a radio signal processing system that estimates information contained in a transmitted signal based on a received signal r(i) received by a first communication station from a transmitted signal transmitted from two or more second communication stations, Estimated signal x, which is the waveform of the transmitted signal. k The spin variable z that (i) is shown by equation (1) k The estimated error E shown by equation (2) based on (i) MUD Based on (z(i)), the estimated signal x is obtained by estimating the waveform of each transmission signal transmitted from two or more of the second communication stations. k (i) The spin variable z k A selection means for selecting two or more patterns z(i) of (i), A set Z of two or more patterns z(i) selected by the selection means QA (i), and the estimation error E MUD (z(i)), based on which the estimated signal x k is a probability distribution Pr showing the distribution of the probability that the transmission signal Approx [r(i)|x(i)] is calculated using equation (3), a probability calculation means The probability distribution Pr calculated by the aforementioned probability calculation means Approx The system includes estimation means for estimating information contained in the transmitted signal based on [r(i) | x(i)]. A wireless signal processing system characterized by the following. [Math 1] Here, x k (i) shows the i-th symbol of the estimated signal for the transmitted signal sent from the k-th second communication station, z 1 k , z 2 k These represent the real and imaginary spin variables, respectively. [Math 2] Here, K indicates the number of the second communication station, k and l indicate the numbers of the second communication station, and h k This indicates the k-th channel coefficient between the second communication station and the first communication station, and J ij This shows the interaction between i and j, and a k a l is the transmitted signal x k This indicates the magnification when modulating. [Math 3] Here, σ w This indicates the standard deviation. Also, x(i) = [x 1 (i), ..., x k (i), ..., x K (i)]. Also, r(i) indicates the i-th symbol of the received signal.

6. In a radio signal processing method for estimating information contained in a transmitted signal based on a received signal received by a first communication station from a transmitted signal transmitted from two or more second communication stations, A selection step of selecting two or more estimated signal patterns, which are estimated from the waveforms of each transmitted signal transmitted from two or more of the second communication stations, based on the estimated signal pattern obtained by estimating the waveform of the transmitted signal and the estimation error corresponding to the received signal, A probability calculation step, based on the estimation errors for two or more patterns of estimated signals selected in the selection step, calculates a probability distribution showing the distribution of the probability that the estimated signal is the transmitted signal. The computer is instructed to perform an estimation step, which estimates the information contained in the transmitted signal based on the probability distribution calculated in the probability calculation step. A wireless signal processing method characterized by the following.

Images