JP5967273B2 - Wireless device and wireless device control method - Google Patents

Description

  The present invention relates to a radio apparatus and a radio apparatus control method.

  In recent years, research on MIMO (Multiple Input Multiple Output) technology has been conducted as a next-generation communication technology. In a MIMO system, a plurality of data streams are transmitted from a transmitter having a plurality of transmission antennas (for example, M), and the plurality of data streams are separated and received by a receiver having a plurality of reception antennas (for example, N). . However, M ≦ N.

  Here, for the sake of simplicity, a case where the transmitter transmits M data streams equal to the number of transmission antennas will be described as an example. On the receiver side, N received signals are received. Here, the data stream is a vector x of M rows and 1 column, the channel matrix of N rows and M columns having the propagation path gain hij between the jth transmitting antenna and the ith receiving antenna as elements, and the received signal is N. If a row 1 column vector y and noise is an N row 1 column vector n, they can be expressed as in Equation (1).

Stream separation methods on the receiving side include MMSE (Minimum Mean Square Error) and MLD (Maximum Likelihood Detection). In MLD, a metric such as a square Euclidean distance is calculated for all symbol replica candidate combinations of a plurality of stream signals, and a combination having a minimum total metric is set as a signal after stream separation. MLD can obtain superior reception characteristics as compared with linear separation methods such as MMSE. However, if the modulation multi-level number of the l-th transmission signal is m ₁ (for example, m = 4 in QPSK, m = 16 in 16QAM, m = 64 in 64QAM), the number of combinations is a number. It becomes two formulas.

  As shown in Formula 2, there is a problem that the number of metric calculations increases exponentially with the increase in the number of modulation multi-values and the number of transmission streams, and the processing amount becomes enormous. For this reason, various MLD reduction type MLDs have been proposed.

In the prior art, QRM-MLD combining QR decomposition and M algorithm has been proposed. In QRM-MLD, a metric such as a square Euclidean distance with all symbol replica candidates is calculated for the surviving symbol replica candidates in the previous stage. Number of metric calculations are k = 1, ..., the number of surviving candidates M-th stage when the S _k, the number 3 expression.

In the prior art, an ASESS (Adaptive Selection of Surviving Symbol replica candidates based on the maximum reliability) method in which a further metric calculation number reduction method is applied to QRM-MLD is shown. Symbol replica candidates in each stage are ranked by region detection, and metrics are calculated for the number of surviving symbol replica candidates in order from the symbol replica with the smallest metric accumulated value. Number of metric calculations are k = 1, ..., the number of surviving candidates M-th stage when the S _k, the Formula 4.

  In the ASESS method, the number of metric calculations increases linearly with the number of transmission streams. Patent Document 1 discloses a method in which the ranking of symbol candidates is applied to an LSD (List Sphere Decoding) method.

  The ASESS method will be described in detail. Further, in order to simplify the explanation, a case where M = N is taken as an example. In the ASESS method, the channel matrix H is QR-decomposed into a unitary matrix Q and an upper triangular matrix R as shown in Equation 5.

  o represents a zero matrix. In the ASESS method, when the received signal y is multiplied by the Hermitian conjugate of the unitary matrix Q from the left, it can be orthogonalized as shown in Equation 6.

In the first stage, in the first stage, the area detection of Equation 7 is performed for the lowest stage, and the area number ε ⁽¹⁾ to which u _N belongs is determined.

In ASESS method, area detection is made from the origin movement of _{N div} times quadrant detection and _{N div} -1 ^times, it belongs to any of the ^{2 2Ndiv} number of region detection. In the ASESS method, the symbol ranking table Ω is referred to, and a candidate replica having the number of surviving candidates S ₁ from the top of the ranking is used as a first-stage survivor path, and a metric such as a square Euclidean distance is calculated. For the metric, the surviving path is expressed by equation (8), and in the case of the square Euclidean distance, equation (9) is obtained.

However, Ω ⁽⁴⁾ , Ω ⁽¹⁶⁾ , and Ω ⁽⁶⁴⁾ represent symbol ranking tables for QPSK, 16QAM, and 64QAM, respectively. Equation ⁽¹⁰⁾ represents the symbol number of the ranking i for the region number ε ⁽¹⁾ stored in the symbol ranking table.

In the second stage, in the second stage, the region detection of Equation 11 is performed by canceling each of the S ₁ surviving path candidate replicas surviving in the first stage from the second received signal z _N−1 from the bottom. The region number ε ⁽¹⁾ (i) to which Equation 12 belongs is determined.

  The ASESS method adaptively selects the surviving path in the second stage as follows. First, in the ASESS method, the representative cumulative metric value E (i) and the current rank ρ (i) of each surviving path that survived in the first stage are initialized, and are represented by Equation 13, Equation 14, and Equation 15.

ASESS method, ρ (i) ≦ _{m N} and E (i) selects the ranking [rho _{(i min)} position candidate replica _{i min} of the minimum value from the symbol ranking table, survival of q-th second stage The path is expressed by Equation 16 and Equation 17.

  The cumulative metric is calculated as shown in Equation 18.

  Then, in the ASESS method, the accumulated metric is updated as shown in Equation 19, Equation 20, and Equation 21.

ASESS method performs the above process until q reaches the surviving path number _{S 2} of the second stage. In the ASESS method, in the subsequent k-th stage, Equation (22) is obtained by canceling each candidate replica of the SK ₋₁ surviving paths that survived in the k−1 stage from the kth received signal z _{N−K + 1} from the bottom. The region number ε ^(k) (i) to which the equation ⁽ 23 ⁾ belongs is determined.

  The ASESS method adaptively selects a survival path in the k-th stage as follows. In the ASESS method, first, the representative cumulative metric value E (i) and the current rank ρ (i) of each surviving path that has survived in the (k−1) -th stage are initialized to be expressed by Equation 24, Equation 25, and Equation 26. .

ASESS _{method, ρ (i) ≦ m N} -k + 1 and E (i) selects the ranking [rho _{(i min)} position candidate replica _{i min} of the minimum value from the symbol ranking table, q th k stage The surviving paths are expressed by Equation 27 and Equation 28.

  The cumulative metric is calculated as shown in Equation 29.

Then, ASESS method, the number 30 formula, the number 31 expression, and update as in equation 32 formula is carried out until the above processing q reaches the surviving path number _{S k} of the k stage.

JP 2006-270430 A

K. J. Kim and J. Yue, “Joint channel estimation and data detection algorithms for MIMO-OFDM systems,” in Proc. Thirty-Sixth Asilomar Conference on Signals, Systems and Computers, pp. 1857-1861, Nov. 2002. K. Higuchi, H. Kawai, N. Maeda and M. Sawahashi, “Adaptive Selection of Surviving Symbol Replica Candidates Based on Maximum Reliability in QRM−MLD for OFCDM MIMO Multiplexing,” Proc. Of IEEE Globecom 2004, pp. 2480−2486 , Nov. 2004. Higuchi, Kawai, Maeda, Sawahashi, “Adaptive Survival Symbol Replica Candidate Selection Method Using Reliability Information in OFCDM MIMO Multiplexing Using QRM-MLD,” RCS2004-69, May 2004

  As described above, in the ASESS method, signal region detection in which symbol candidates of each surviving path from the received signal to the previous stage are canceled is performed. In the ASESS method, a symbol ranking table is used to reduce the amount of processing by using a symbol candidate of the current rank of the surviving path having the smallest representative cumulative metric as a surviving candidate. The symbol ranking table is a table that stores the result of ranking the symbol candidates in order of increasing distance from the center of the region to the symbol candidate for each region prepared in advance.

  In the ASESS method, the symbol ranking table in the k-th stage is a table in which the number of words equal to Equation 33 and the bit width per word are equal to Equation 34 for each modulation method. Here, m is the modulation multi-level number of the modulation system.

Here, examples of symbol ranking tables in the conventional example are shown in FIGS. The size of the symbol ranking table is shown in FIG. Although an example of different N _div symbol ranking tables in the same modulation scheme is shown, the same modulation scheme may have at least one N _div symbol ranking table.

  As shown in FIGS. 41 to 46, the size of the symbol ranking table increases as the number of modulation multi-values and the number of regions increase. As the size of the symbol ranking table increases, the memory capacity of the device increases.

  The disclosed technology has been made in view of the above, and an object thereof is to realize a wireless device and a wireless device control method capable of reducing the size of a symbol ranking table.

  The radio | wireless apparatus which this application discloses is provided with the signal conversion part which converts the received signal received via the antenna into the signal containing the product of an upper triangular matrix and a transmission signal in one aspect. In addition, the wireless device includes an area detection unit that detects an area on the IQ plane to which the signal converted by the signal conversion unit belongs. In addition, the wireless device divides the first rank to the order equal to the modulation multi-level number into two or more ranges, and in each range, the distance from the representative point of the adjacent area is short for each adjacent area. A symbol ranking table storing symbol candidates in order is provided. In addition, the radio apparatus includes a surviving symbol selection unit that selects a symbol candidate based on the region detected by the region detection unit and the symbol ranking table.

  According to one aspect of the wireless device disclosed in the present application, the size of the symbol ranking table can be reduced.

FIG. 1 is a diagram showing the CCDF of the current rank of each surviving path at the end of the stage. FIG. 2 is a diagram illustrating a configuration of the MIMO system. FIG. 3 is a diagram illustrating a configuration example of the MIMO stream separation unit according to the first embodiment. FIG. 4 is a diagram showing a flowchart of area detection. FIG. 5 is a diagram illustrating an example of QPSK (N _div = 1) region detection. FIG. 6 is a diagram illustrating an example of a size reduction symbol ranking table of QPSK, N _div = 1. FIG. 7 is a diagram illustrating an example of area detection of QPSK (N _div = 2). FIG. 8 is a diagram illustrating an example of a size reduction symbol ranking table of QPSK, N _div = 2. FIG. 9 is a diagram illustrating an example of 16QAM (N _div = 2). FIG. 10 is a diagram illustrating an example of a 16QAM (N _div = 2) size reduction symbol ranking table. FIG. 11 is a diagram illustrating an example of 16QAM (N _div = 3). FIG. 12 is a diagram illustrating an example of a 16QAM (N _div = 3) size reduction symbol ranking table. FIG. 13 is a diagram illustrating an example of 64QAM (N _div = 3). FIG. 14 is a diagram illustrating an example of a size reduction type symbol ranking table of 64QAM and N _div = 3. FIG. 15 is a diagram illustrating an example of area detection. FIG. 16 is a flowchart illustrating the first embodiment. FIG. 17 is a diagram illustrating a flowchart of the first stage process. FIG. 18 is a diagram illustrating a flowchart of surviving symbol selection in the first stage process. FIG. 19 is a diagram illustrating a flowchart of the k-th stage process. FIG. 20 is a diagram illustrating a flowchart of surviving symbol selection in the k-th stage process. FIG. 21 is a diagram showing an outline of the second mode. FIG. 22 is a diagram illustrating a configuration example of the MIMO stream separation unit according to the second embodiment. FIG. 23 is a diagram illustrating an example of a symbol upper ranking table and a symbol lower ranking table with QPSK, N _div = 2. FIG. 24 is a diagram illustrating an example of a symbol upper ranking table and a symbol lower ranking table of 16QAM, N _div = 3. FIG. 25 is a diagram illustrating an example of a symbol upper ranking table of 64QAM, N _div = 3. FIG. 26 is a diagram illustrating an example of a symbol lower ranking table of 64QAM, N _div = 3. FIG. 27 is a diagram illustrating a flowchart of surviving symbol selection in the first stage processing according to the second embodiment. FIG. 28 is a flowchart illustrating the k-th stage process according to the second embodiment. FIG. 29 is a flowchart illustrating survival symbol selection in the k-th stage process according to the second embodiment. FIG. 30 is a diagram illustrating a configuration example of the MIMO stream separation unit according to the third embodiment. FIG. 31 is a diagram illustrating examples of a symbol upper ranking table, a symbol middle ranking table, and a symbol lower ranking table. FIG. 32 is a flowchart illustrating survival symbol selection in the first stage process according to the third embodiment. FIG. 33 is a flowchart illustrating survival symbol selection in the k-th stage process according to the third embodiment. FIG. 34 is a diagram illustrating an example of a symbol upper ranking table and a symbol lower ranking table. FIG. 35 is a flowchart illustrating the first stage process according to the fifth embodiment. FIG. 36 is a diagram illustrating an example of the symbol upper ranking table and the symbol lower ranking table for the first stage processing according to the fifth embodiment. FIG. 37 is a diagram illustrating the configuration of the MIMO stream separation unit according to the sixth embodiment. FIG. 38 is a flowchart illustrating the sixth embodiment. FIG. 39 is a diagram illustrating the configuration of the MIMO stream separation unit according to the seventh embodiment. FIG. 40 is a flowchart illustrating the seventh embodiment. FIG. 41 is a diagram illustrating an example of a symbol ranking table for QPSK, N _div = 1. FIG. 42 is a diagram illustrating an example of a symbol ranking table for QPSK, N _div = 2. FIG. 43 is a diagram illustrating an example of a symbol ranking table for 16QAM, N _div = 2. FIG. 44 is a diagram illustrating an example of a symbol ranking table for 16QAM, N _div = 3. FIG. 45 is a diagram illustrating an example of a symbol ranking table of 64QAM, N _div = 3. FIG. 46 is a diagram showing the size of the symbol ranking table.

  Embodiments of a wireless device and a wireless device control method disclosed in the present application will be described below in detail with reference to the drawings. The disclosed technology is not limited by this embodiment.

(Solution 1)
First, the basic concept of the first aspect for solving the problem of the present application will be described. First, regarding the conventional example, consider the current rank of each surviving path. FIG. 1 is a diagram showing a CCDF (Complementary Cumulative Distribution Function) of the current rank of each surviving path at the end of the stage. More specifically, FIG. 1 shows, for each surviving path, when the number of transmitting antennas is 4, the number of receiving antennas is 4, the modulation scheme of transmission data of each transmitting antenna is 64QAM, and the number of surviving candidates S1 = S2 = S3 = S4 = 64. It is a figure which shows CCDF of the present rank at the time of the end of a stage. However, since the current rank tends to increase as the surviving path with a smaller cumulative metric up to the previous stage, only the 10 surviving paths are shown in ascending order of the cumulative metric up to the previous stage. From FIG. 1, it can be seen that the probability that the current rank becomes a large value is small.

  Therefore, in the first mode, a surviving path whose current rank is less than or equal to a preset upper limit value MAX_RANK (<m) and whose representative cumulative metric is minimum is selected. Thereby, since the current rank does not become higher than MAX_RANK, the number of words in the symbol ranking table can be expressed by Equation 35. Therefore, according to the first aspect, it is possible to reduce the size of the symbol ranking table.

(Example 1)
Example 1 is an example to which the first aspect is applied. FIG. 2 is a diagram illustrating a configuration of the MIMO system. As shown in FIG. 2, the MIMO system includes a transmitter 100 and a receiver 200. The transmitter 100 includes an error correction encoding unit 102, a modulation unit 104, a plurality of transmission units 106, and a plurality of transmission antennas 108.

  The error correction coding unit 102 performs error correction coding on the transmission data. The modulation unit 104 performs modulation such as QPSK, 16QAM, and 64QAM, and separates the streams. The transmission unit 106 up-converts the signal modulated by the modulation unit 104 to a radio frequency, and transmits it simultaneously from the transmission antenna 108.

  The receiver 200 includes a plurality of reception antennas 202, a reception unit 204, a demodulation unit 208, an error correction decoding unit 214, a CPU (Central Processing Unit) 260, and a memory 262. The demodulation unit 208 includes a channel estimation unit 210 and a MIMO stream separation unit 212.

  The receiving unit 204 converts a signal received by the receiving antenna 202 into a baseband signal. Channel estimation section 210 estimates a propagation path by receiving a pilot signal or the like transmitted from transmitter 100. The MIMO stream separation unit 212 performs stream separation processing using the received signal and the channel estimation value. The error correction decoding unit 214 performs error correction decoding of the separated stream.

  The memory 262 includes a ROM (Read Only Memory) that stores data for executing various functions of the receiver 200 and various programs for executing the various functions of the receiver 200. The memory 262 has a RAM (Random Access Memory) that stores programs to be executed among various programs stored in the ROM.

  The CPU 260 is an arithmetic processing unit that executes various programs stored in the memory 262. The CPU 260 controls the receiver 200 by executing various programs stored in the memory 262. The program executed by the CPU 260 is not only stored in the memory 262 but also recorded on a storage medium that can be distributed such as a CD (Compact Disc) -ROM or a memory medium, and read from the storage medium for execution. Can do. In addition, the program is stored in a server connected via a network so that the program can be operated on the server, and a service can be requested in response to a request from the receiver 200 connected via the network. It can also be provided to the receiver 200.

  FIG. 3 is a diagram illustrating a configuration example of the MIMO stream separation unit according to the first embodiment. As illustrated in FIG. 3, the MIMO stream separation unit 212 includes a QR decomposition unit 216, a MIMO demodulation unit 222, and an LLR (Log Likelihood Ratio) calculation unit 254. The QR decomposition unit 216 includes a QR decomposition processing unit 218 and a received signal conversion unit 220. The MIMO demodulator 222 includes a first stage processor 224, a second stage processor 234,... And an Nth stage processor 244.

  The first stage processing unit 224 includes a region detection unit 226, a surviving symbol selection unit 228, a metric calculation unit 230, and a symbol ranking table 232. The second stage processing unit 234 includes a region detection unit 236, a surviving symbol selection unit 238, a metric calculation unit 240, and a size reduction type symbol ranking table 242. The Nth stage processing unit 244 includes a region detection unit 246, a surviving symbol selection unit 248, a metric calculation unit 250, and a size reduction type symbol ranking table 252.

  The QR decomposition unit 216 performs QR decomposition that decomposes the channel matrix into a unitary matrix Q and an upper triangular matrix R, as shown in Equation 36.

  Here, by appropriately selecting the unitary matrix Q, the diagonal component of the matrix R can be made a positive real number. The unitary transform unit multiplies the received signal vector y by the Hermitian conjugate of the unitary matrix Q to obtain a unitary transform vector z as shown in Equation 37.

  At this time, the relationship expressed by Equation 38 is established between the unitary conversion vector z and the transmission stream vector x.

The area detector 226 of the first stage performs the area detection of Equation 39 in the lowest stage and determines the area number ε ⁽¹⁾ to which u _N belongs.

The surviving symbol selection unit 228 ranks the symbol candidates in descending order of distance from the center of the region to each symbol candidate c _{N, i} (i = 1, 2,..., _{M N} ) of the transmission signal x _N in advance for each region number. Refer to the symbol ranking table Ω that stores the results. Then, the surviving symbol selection unit 228 sets the surviving candidate number S ₁ candidate replicas from the top ranking as surviving paths of the first stage. The surviving symbol selection unit 228 sets the surviving path to several 40 formulas.

Here, Π (k) (i) represents the i-th surviving path of the k-th stage. _{Ｊ j} ^{(k) When} expressed as (i), it represents a path in the j-th stage (j = 1, 2,..., K) of the i-th surviving path in the k-th stage. Further, when expressed as Π a to _b ^(k) (i) (a <b), it represents a partial path from the a-th stage to the b-th stage of the i-th surviving path of the k-th stage. For example, in the case of Equation 41, Equation 42, Equation 43, and Equation 44 are obtained.

The metric calculation unit 230 calculates a metric such as a square Euclidean distance. In the case of the square Euclidean distance, Equation 45 is obtained. However, Ω ⁽⁴⁾ , Ω ⁽¹⁶⁾ , and Ω ⁽⁶⁴⁾ represent symbol ranking tables for QPSK, 16QAM, and 64QAM, respectively.

Equation 46 represents the symbol number of the ranking i for the region number ε ⁽¹⁾ stored in the symbol ranking table. Here, c _{N, i} is a signal point of the modulation method used for transmission of the transmission stream x _N. Typical examples of the modulation scheme include QPSK, 16QAM, and 64QAM, but this embodiment is not limited to a specific modulation scheme.

The subsequent k-th stage (for example, second stage) processing will be described. In the k-th stage area detection unit, the area of Formula 47 is obtained by canceling each candidate replica of S _k−1 surviving paths that survived in the k−1 stage from the kth received signal Z _{N−K + 1} from the bottom. Detection is performed to determine the region number ε ^(k) (i) to which Equation 48 belongs.

  The survival symbol selection unit (for example, the survival symbol selection unit 238) adaptively selects the survival path in the k-th stage as follows. First, the representative cumulative metric value E (i) and the current rank ρ (i) of each surviving path that has survived in the k−1th stage are initialized as Equation 49, Equation 50, and Equation 51.

The surviving symbol selection unit, ρ (i) ≦ MAX_RANK _k and E (i) is selected from the Index [rho _{(i min)} position size-reduced symbol ranking table candidate replica of _{i min} of the minimum value. The survival symbol selection unit sets the survival path of the qth k-th stage to Formula 52 and Formula 53.

However, MAX_RANK _k is a rank upper limit value of the k-th stage set in advance. MAX_RANK _k may be set to a different value for each k-th stage modulation scheme. Further, in order to be able to select the _{S k} number of surviving candidates _shall meet the _{S K-1 × MAX_RANK K ≧} S K.

  A metric calculation unit (for example, metric calculation unit 240) calculates a cumulative metric as shown in Formula 54.

  Then, the metric calculation unit updates the accumulated metric as shown in Formula 55, Formula 56, and Formula 57.

The metric calculation unit performs the above processing until q reaches the number _k of surviving paths in the k-th stage. Each stage processing unit performs the same processing up to the Nth stage thereafter.

Next, details of the area detection and size reduction type symbol ranking table will be described. Symbol ranking table of the first stage processing unit may be able to have only a subset portion of the left side S ₁ row of symbol ranking table shown in FIGS. 41 45.

The details of the size reduction symbol ranking table and area detection of the kth stage will be described below. Here, examples of Formula 58, Formula 59, and Formula 60 are shown as setting values of MAX_RANK _k . However, this is an example of N = 4. Further, this parameter is merely an example, and different setting values may be used.

  Since the description of the area detection here is common to all stages, the subscript of the signal for detecting the area is omitted and u is assumed. FIG. 4 is a diagram showing a flowchart of area detection. First, the area detection unit sets t to 2 and sets χmod as shown in Formula 61 (step S101).

In addition, the region detection unit first determines whether the real part and the imaginary part of u ⁽¹⁾ = u are positive or negative (IQ component positive / negative), and to which quadrant u belongs (step S102). . The region detection unit determines whether t> N _div is satisfied (step S103). When t> N _div is satisfied (step S103, Yes), the region detection unit determines a region number (step S104). If t> N _div is not satisfied (step S103, No), the area detection unit moves the origin to the center of the quadrant determined first (formula 62) (step S105). In addition, the region detection unit performs quadrant detection of u ⁽²⁾ whose origin is moved to the center of the quadrant determined in the first time (formula 62) for the ^{second time} (step S106). However, sign (x) is as shown in Equation 63.

Subsequently, the region detection unit increments t and sets χmod / 2 to χmod (step S107). Subsequently, the region detection unit determines whether or not t ≦ N _div (step S108). If t ≦ N _div is satisfied (step S108, Yes), the region detection unit returns to step S105 and repeats the process. In addition, the area detection unit performs quadrant detection of u ^(t) in which the origin is moved to the center (formula 64) of the quadrant determined at the t−1th time.

If t ≦ N _div is not satisfied (step S108, No), the region detection unit determines a region number (step S109). That is, the region detection unit sequentially repeats the above processing N _div times to determine which region belongs to the number of regions represented by Formula 65.

Here, an example of how to _assign an area number of N _div = 1 of QPSK will be described. FIG. 5 is a diagram illustrating an example of QPSK (N _div = 1) region detection. FIG. 6 is a diagram illustrating an example of a size reduction symbol ranking table of QPSK, N _div = 1.

For each region, symbol numbers are stored in order of increasing distance between a representative point of the region (a symbol position included in the region) and the symbol. The representative point is expressed by Expression 66 using the region number ε. However, Expression 67, bit (ε, n) is the bit value of the n (1 to 2N _div ) bit from the MSB of ε represented by the bit width of 2N _div [bit]. In this case, the representative point of each region is the same as the symbol position belonging to each region.

Here, since MAX_RANK ^(QPSK) _k = 3, only the symbol numbers up to the third ranking are held. Therefore, as shown in FIG. 6, the number of words in the table is 4 (number of areas) × 3 (MAX_RANK ^(QPSK) _k ) = 12, and the table size is 3/4 compared to the conventional example. By the way, when N _div = 1, the second closest symbol and the third closest symbol from the representative point of the region cannot be distinguished. For example, in the case of area 0, it is impossible to determine which of 10 (decimal: 2) and 01 (decimal: 1) is close. Just keep it. Further, the method of assigning area numbers is also an example, and may be given arbitrarily. In that case, only the rows of the size-reduced symbol ranking table shown in FIG. 6 are exchanged. The same applies to the following examples.

Here, in the example of N _div = 1 of QPSK, the second closest symbol and the third closest symbol from the representative point of the region cannot be distinguished. It was nice to create. Ranking accuracy can also be increased by increasing the number of quadrant detections N _div .

An example of QPSK N _div = 2 will be described. FIG. 7 is a diagram illustrating an example of area detection of QPSK (N _div = 2). In this case, the representative point of each region is any one of (± {1,3} / 2√2, ± {1,3} / 2√2).

The size reduction type symbol ranking table will be described. FIG. 8 is a diagram illustrating an example of a size reduction symbol ranking table of QPSK, N _div = 2. In regions 1, 2, 4, 7, 8, 11, 13, and 14, the second closest symbol and the third closest symbol from the representative point of the region are distinguished, and the accuracy of the size reduction symbol ranking table is improved. Yes. Since MAX_RANK ^(QPSK) _k = 3, only the symbol numbers up to the third ranking are held.

Next, an example of how to assign an area number of 16QAM, N _div = 2 and a size reduction type symbol ranking table will be described. FIG. 9 is a diagram illustrating an example of 16QAM (N _div = 2). FIG. 10 is a diagram illustrating an example of a 16QAM (N _div = 2) size reduction symbol ranking table.

The white numbers in the symbols in FIG. 9 are symbol numbers (bit strings converted into decimal numbers). In this case, the representative point of each region is the same as the symbol position included in the region. As shown in FIG. 10, the size reduction type symbol ranking table holds only symbol numbers up to the 8th ranking because MAX_RANK ^(16QAM) _k = 8.

Next, an example of a region numbering method of 16QAM, N _div = 3 and a size reduction type symbol ranking table will be described. FIG. 11 is a diagram illustrating an example of 16QAM (N _div = 3). FIG. 12 is a diagram illustrating an example of a 16QAM (N _div = 3) size reduction symbol ranking table.

In this case, the representative point of each region is any one of (± {1, 3, 5, 7} / 2√10, ± {1, 3, 5, 7} / 2√10). As shown in FIG. 12, the size-reduced symbol ranking table holds only symbol numbers up to the 8th ranking because MAX_RANK ^(16QAM) _k = 8.

Next, an example of how to _assign an area number of 64QAM, N _div = 3 and a size reduction type symbol ranking table will be described. FIG. 13 is a diagram illustrating an example of 64QAM (N _div = 3). FIG. 14 is a diagram illustrating an example of a size reduction type symbol ranking table of 64QAM and N _div = 3.

In this case, the representative point of each region is the same as the symbol position included in the region. As shown in FIG. 14, the size reduction type symbol ranking table holds only symbol numbers up to the 32nd ranking because MAX_RANK ^(16QAM) _k = 32.

As described above, it can be understood that the size-reduced symbol ranking table only needs to have a subset portion of the left MAX_RANK _k columns of the symbol ranking tables shown in FIGS. That is, it can be easily applied to a setting value different from the setting value example of MAX_RANK _k given in this example.

Here, a specific example of area detection will be described by taking the case of 64QAM as an example. FIG. 15 is a diagram illustrating an example of area detection. In the example of FIG. 15, u is indicated by a star, and N _div = 3. As shown in the lower left of FIG. 15, the area is divided into 64 areas, and the area numbers are given as the numbers in the upper left of the area. In the first quadrant detection, the real part and the imaginary part are determined to be positive or negative, and both are negative, so it is determined to be the fourth quadrant. Next, since the fourth quadrant is determined in the first quadrant detection, the origin is moved to (−4 / √42, −4 / √42). This is expressed by Equation 68 by subtracting the coordinates of the next origin from u.

Similarly, in the second quadrant detection, whether the real part and the imaginary part of u ⁽²⁾ are positive or negative is determined, and as a result of the determination, the first quadrant is detected. As in the first time, the origin is moved to (−2 / √42, −2 / √42). A third quadrant is detected, and the third quadrant is detected. As a result, it is determined that u belongs to the region number 50.

Returning to the description of FIG. 3, the LLR calculation unit 254 calculates a bit LLR for each transmission stream. First, the LLR calculation unit 254 searches for the minimum value of the cumulative metric, and sets the surviving path having the minimum cumulative metric as the maximum likelihood symbol combination. Bit LLR of the n-th bit of the first stream x _l is a difference between the minimum value of the accumulated metric for the symbol with the inverted value of the n-th bit of the total metric and the maximum likelihood symbol for the maximum likelihood symbol combination. Specifically, the bit LLR is expressed by Equation 72 under the conditions of Equation 69, Equation 70, and Equation 71.

  Here, l represents a stream number, bit (x, n), and invbit (x, n) represent an nth bit value and an inverted bit value of x, respectively. In the above description, the bit LLR is used as a metric difference, but the bit LLR can also be expressed by Equation 73 by taking the square root.

  Next, a flowchart of the MIMO stream separation unit according to the first embodiment will be described. FIG. 16 is a flowchart illustrating the first embodiment. As illustrated in FIG. 16, the QR decomposition processing unit 218 performs QR decomposition (step S201). Subsequently, the reception signal conversion unit 220 performs reception signal conversion and calculates a unitary conversion vector z (step S202). Subsequently, the first stage processing unit 224 performs first stage processing (step S203). Details of the first stage process will be described later.

  Subsequently, the MIMO demodulator 222 sets k to 2 (step S204). Subsequently, the kth stage processing unit performs kth stage processing (step S205). Details of the k-th stage process will be described later. Subsequently, the MIMO demodulator 222 increments k (step S206) and determines whether or not k ≦ N (step S207).

  If k ≦ N is satisfied (step S207, Yes), the MIMO demodulator 222 returns to step S205 and repeats the process. On the other hand, when k ≦ N is not satisfied (step S207, No), the LLR calculation unit 254 calculates the LLR as described above (step S208).

Next, a flowchart of the first stage process will be described. FIG. 17 is a diagram illustrating a flowchart of the first stage process. As shown in FIG. 17, the region detecting unit 226 detects the region of _{u N} (step S301). Subsequently, the first stage processing unit 224 sets q to 1 (step S302). Subsequently, the surviving symbol selection unit 228 selects a surviving symbol (step S303). The selection of the survival symbol will be described later.

Subsequently, the metric calculation unit 230 calculates a metric as described above (step S304). Subsequently, the first stage processor 224 increments q (step S305), determines whether the q ≦ _{S 1} (step S306). The first stage processor 224, in the case of q ≦ _{S 1} (step S306, Yes), the operation returns to step S303. On the other hand, the first stage processor 224, if not the q ≦ _{S 1} (step S306, No), the process ends.

Next, a flowchart of surviving symbol selection in the first stage process will be described. FIG. 18 is a diagram illustrating a flowchart of surviving symbol selection in the first stage process. As illustrated in FIG. 18, the surviving symbol selection unit 228 selects a symbol of ε ⁽¹⁾ rows and q columns of Ω ^{(mk) in} the symbol ranking table and sets it as a surviving path (step S401).

Next, a flowchart of the kth stage process will be described. FIG. 19 is a diagram illustrating a flowchart of the k-th stage process. As shown in FIG. 19, the k-th stage processing unit sets 1 to i (step S501). Subsequently, the region detection unit performs region detection as shown in Equation 47 (step S502). Subsequently, the region detection unit increments i (step S503). Subsequently, the region detection unit determines whether i ≦ S _k−1 (step S504). If i ≦ S _k−1 is satisfied (step S504, Yes), the region detection unit returns to step S502 and repeats the process.

On the other hand, if i ≦ S _k−1 is not satisfied (step S504, No), the surviving symbol selection unit survives at the k−1 stage as shown in Equation 49, Equation 50, and Equation 51. The representative cumulative metric value E (i) and the current rank ρ (i) of each surviving path are initialized (step S505). Subsequently, the surviving symbol selection unit selects i _min where ρ (i) ≦ MAX_RANK _k and the representative cumulative metric E (i) is the minimum value (step S506).

  Subsequently, the surviving symbol selection unit selects a surviving symbol (step S507). The selection of the survival symbol will be described later. Subsequently, the metric calculation unit calculates a cumulative metric as shown in Formula 54 (step S508). Subsequently, the metric calculation unit updates the accumulated metric as shown in Formula 55, Formula 56, and Formula 57 (step S509).

Subsequently, the metric calculation unit determines whether q ≦ S _k (step S510). When q ≦ S _k is satisfied (Yes in step S510), the surviving symbol selection unit returns to step S506 and repeats the process. On the other hand, if q ≦ S _k is not satisfied (step S510, No), the metric calculation unit ends the process.

Next, a flowchart of survival symbol selection in the k-th stage process will be described. FIG. 20 is a diagram illustrating a flowchart of surviving symbol selection in the k-th stage process. As illustrated in FIG. 20, the surviving symbol selection unit adds symbols in the ρ (i _min ) column of ε ^(k) (i _min ) rows of the size reduction type symbol ranking table Ω to the surviving path (step S601).

As described above, according to the first embodiment, since the current rank of each surviving path does not become larger than the preset rank upper limit value MAX_RANK, the symbol ranking table may hold only symbol candidates equal to or lower than MAX_RANK. As a result, according to the first embodiment, the size of the symbol ranking table can be reduced as compared with the conventional example. Note that MAX_RANK _k and N _div are not limited to the examples shown in this embodiment.

(Solution Mode 2)
Next, a basic concept of the second aspect for solving the problem of the present application will be described. The second mode is divided into a symbol upper ranking table that holds symbol candidates of ranking higher than a certain rank ρ _th and a symbol lower ranking table that holds symbol candidates of ranking lower than ρ _th . In the second mode, the symbol lower ranking table holds the symbol candidates in the order of the _shortest distance from the representative point of the adjacent N _adj regions. Thereby, the number of words in the symbol lower ranking table can be reduced, and the table size can be reduced.

  FIG. 21 is a diagram showing an outline of the second mode. As shown in FIG. 21, for example, in the case of 16QAM and 64 areas, the upper symbol candidates ranked 1 to 8 in the areas 0, 1, 2, and 3 are common, and the lower symbols in the 8 to 16 positions. Candidates are also common. Therefore, in the second mode, for the upper symbol candidates, in order to maintain the accuracy of ranking, the symbol candidates ranked for each region are held in the symbol upper ranking table. In the second mode, for the lower symbol candidates, the accuracy of ranking is coarsened and the symbol candidates are held in the symbol lower ranking table in the order of the distance from the representative point of the adjacent region for each adjacent region.

(Example 2)
Example 2 is an example to which the second aspect is applied. FIG. 22 is a diagram illustrating a configuration example of the MIMO stream separation unit according to the second embodiment. As illustrated in FIG. 22, the MIMO stream separation unit 312 includes a QR decomposition unit 316, a MIMO demodulation unit 322, and an LLR calculation unit 354. The QR decomposition unit 316 includes a QR decomposition processing unit 318 and a received signal conversion unit 320. The MIMO demodulator 322 includes a first stage processor 324, a second stage processor 334,... And an Nth stage processor 344.

  The first stage processing unit 324 includes a region detection unit 326, a surviving symbol selection unit 328, a metric calculation unit 330, a symbol upper ranking table 331, and a symbol lower ranking table 332. The second stage processing unit 334 includes a region detection unit 336, a surviving symbol selection unit 338, a metric calculation unit 340, a symbol upper ranking table 341, and a symbol lower ranking table 342. The N-th stage processing unit 344 includes a region detection unit 346, a surviving symbol selection unit 348, a metric calculation unit 350, a symbol upper ranking table 351, and a symbol lower ranking table 352.

  As shown in FIG. 22, the difference from the first embodiment is that the symbol ranking table 232 and the size-reduced symbol ranking tables 242 and 252 of the first embodiment are replaced with the symbol upper ranking tables 331, 341, 351, and the symbol lower ranking table 332. , 342, 352. Therefore, the description of the same configuration as that of the first embodiment will be omitted as appropriate, and the description will focus on the configuration different from that of the first embodiment.

The surviving symbol selection unit 328 refers to the symbol upper ranking table 331 (Ω ₁ ) or the symbol lower ranking table 332 (Ω ₂ ), and sets the candidate replica of the surviving candidate number S 1 from the higher ranking as the first stage survival path. The surviving path is expressed by Equation 74.

  The survival symbol selection unit of the k-th stage processing unit adaptively selects the survival path in the k-th stage as follows. First, the surviving symbol selection unit initializes the representative cumulative metric value E (i) and the current rank ρ (i) of each surviving path that survived in the (k−1) -th stage, and formulas 75, 76, 77 And

The surviving symbol selection _{unit, ρ (i) ≦ m N} -k + 1 and Index [rho _{(i min)} position of a symbol-level candidate replica ranking table Omega ₁ or symbol lower ranking table of _{i min} of E (i) is the minimum value to choose from Ω _2. Then, the survival symbol selection unit sets the survival path of the qth k-th stage to Formula 78.

Next, details of the symbol upper ranking table and the symbol lower ranking table will be described. In the case of QPSK, N _div = 2, as shown in FIG. 42 and FIG. 7, the symbol candidates included in the first to third rankings are common among four adjacent regions, and the fourth-ranked symbol candidates are also Obviously, it is the same in the four adjacent areas. Therefore, FIG. 23 shows an example where ρ _th = 3 and N _adj = 4. FIG. 23 is a diagram illustrating an example of a symbol upper ranking table and a symbol lower ranking table with QPSK, N _div = 2. As shown in FIG. 23, the total number of words in the symbol upper ranking table and the symbol lower ranking table is 16 × 3 + 4 × 1 = 52, and it can be seen that the table size is 52/64 compared to the conventional example.

Next, a table example and a table design method when ρ _th = 8 and N _adj = 4 in the case of 16QAM, N _div = 3 will be described. First, consider regions 0, 1, 2, and 3. From FIG. 44, the symbol candidates included in the first to eighth rankings are common. Also, the symbol candidates included in the 9th to 16th rankings are common. The symbol lower table is designed so that the distance from the representative point (position of symbol number 3) of the four adjacent regions (regions 0, 1, 2, 3) is in ascending order.

Next, consider the regions 4, 5, 6, and 7. From FIG. 44, the symbol candidates included in the first to eighth rankings of each region are as follows.
Area 4: Symbol number (0, 1, 2, 3, 4, 6, 7, 8)
Area 5: Symbol number (0, 1, 2, 3, 4, 5, 6, 7)
Area 6: Symbol number (0, 1, 2, 3, 4, 6, 8, 9)
Area 7: Symbol number (0, 1, 2, 3, 4, 6, 7, 8)

  Since the areas 0, 1, 2, and 3 are not completely common, two of the symbol numbers 5, 7, 8, and 9 are dropped into the symbol lower ranking table in order to make them common. There is a need. Here, the square error with the rank in the original symbol ranking table (FIG. 44) is made small.

  Assuming that the symbol number 5 of the region 5 is the eighth place by assuming that it is the highest place in the symbol lower ranking table by dropping into the symbol lower ranking table, (9-8) ^ 2 = 1. . Symbol number 7 of region 4 is seventh, symbol number 7 of region 5 is seventh, and symbol number 7 of region 7 is eighth, so (9-7) ^ 2 + (9-7) ^ 2 + (9- 8) ^ 2 = 9.

  Symbol number 8 of region 4 is eighth, symbol number 8 of region 5 is seventh, and symbol number 8 of region 7 is seventh, so (9-8) ^ 2 + (9-7) ^ 2 + (9- 7) ^ 2 = 9. Since the symbol number 9 in the region 6 is in the eighth place, (9-8) ^ 2 = 1.

  In order to reduce the square error, it is preferable to leave the symbol numbers 7 and 8 in the symbol upper ranking table and drop the symbol numbers 5 and 9 into the symbol lower ranking table. Therefore, the symbol numbers 0, 1, 2, 3, 4, 6, 7, and 8 are held in the symbol higher ranking table in the order of the distance from the representative point of each region, and the symbol numbers other than the above are adjacent to each other. The symbol lower ranking table is designed so that the distance from the representative point (symbol number 2) of the four regions (regions 4, 5, 6, and 7) is as close as possible.

Similarly, for the regions 8, 9, 10, and 11, the symbol numbers included in the rankings 1 to 8 are region 8: symbol numbers (0, 1, 2, 3, 4, 8, 9, 11).
Area 9: Symbol number (0, 1, 2, 3, 4, 6, 8, 9)
Area 10: Symbol number (0, 1, 2, 3, 8, 9, 10, 11)
Area 11: Symbol number (0, 1, 2, 3, 4, 8, 9, 11)
It becomes.

For this reason, it is necessary to drop two of the symbol numbers 4, 6, 10, and 11 into the symbol lower ranking table. Similarly,
Symbol number 4: (9-7) ^ 2 + (9-7) ^ 2 + (9-7) ^ 2 = 12
Symbol number 6: (9-8) ^ 2 = 1
Symbol number 10: (9-8) ^ 2 = 1
Symbol number 11: (9-8) ^ 2 + (9-7) ^ 2 + (9-8) ^ 2 = 6
It becomes.

  For this reason, symbol numbers 6 and 10 are dropped into the symbol lower ranking table. Therefore, the symbol numbers 0, 1, 2, 3, 4, 8, 9, and 11 are held in the symbol upper ranking table in the order of the distance from the representative point of each region, and the symbol numbers other than the above are adjacent to each other. The symbol lower ranking table is designed so that the distance from the representative point (symbol number 1) of the four regions (regions 8, 9, 10, 11) is as close as possible.

For the regions 12, 13, 14, and 15, the symbol numbers included in the rankings 1 to 8 are region 12: symbol numbers (0, 1, 2, 3, 4, 6, 8, and 9).
Area 13: Symbol number (0, 1, 2, 3, 4, 6, 8, 12)
Area 14: Symbol number (0, 1, 2, 3, 4, 8, 9, 12)
Area 15: Symbol number (0, 1, 2, 4, 6, 8, 9, 12)
It becomes.

For this reason, it is necessary to drop one of the symbol numbers 3, 6, 9, and 12 into the symbol lower ranking table. Similarly,
Symbol number 3: (9-4) ^ 2 + (9-7) ^ 2 + (9-7) ^ 2 = 33
Symbol number 6: (9-7) ^ 2 + (9-4) ^ 2 + (9-8) ^ 2 = 30
Symbol number 9: (9-8) ^ 2 + (9-4) ^ 2 + (9-7) ^ 2 = 30
Symbol number 12: (9-8) ^ 2 + (9-8) ^ 2 + (9-4) ^ 2 = 27
It becomes.

  For this reason, the symbol number 12 is dropped into the symbol lower ranking table. Therefore, the symbol numbers 0, 1, 2, 3, 4, 6, 8, and 9 are stored in the symbol higher ranking table in the order of the distance from the representative point of each region, and the symbol numbers other than the above are adjacent to each other. The symbol lower ranking table is designed so that the distance from the representative point (symbol number 1) of the four regions (regions 8, 9, 10, and 11) is as short as possible.

Hereinafter, FIG. 24 shows a symbol upper ranking table and a symbol lower ranking table as a result of the same processing for the regions 13 to 63. FIG. 24 is a diagram illustrating an example of a symbol upper ranking table and a symbol lower ranking table of 16QAM, N _div = 3. As can be seen from FIG. 24, the total number of words in the symbol upper ranking table and the symbol lower ranking table is represented by Equation 79.

Therefore, the size is 62.5% compared with the conventional symbol ranking table. Examples of the symbol upper ranking table and the symbol lower ranking table when ρ _th = 32 and N _adj = 4 are set for 64QAM, N _div = 3 using the same method as 16QAM described above. Shown in FIG. 25 is a diagram illustrating an example of a symbol upper ranking table of 64QAM, N _div = 3. FIG. 26 is a diagram illustrating an example of a symbol lower ranking table of 64QAM, N _div = 3. The total number of words in the symbol upper ranking table and the symbol lower ranking table is 2560, which is 62.5% the size of the conventional symbol ranking table.

Here, an example is shown as ρ _th , N _adj , but if the table design method described in detail above is used, a symbol upper ranking table and a symbol lower ranking table are also designed for ρ _th , N _adj of different values. It is possible to do.

In addition, when the symbols included in the ranking from the first rank to the ρ _th rank are not common, the selection is made on the basis of the evaluation standard that reduces the square error, but other methods may be used. For example, in the case of the areas 4, 5, 6, and 7, two of the symbol numbers 5, 7, 8, and 9 need to be selected, so any two can be selected without using an evaluation criterion such as a square error. May be.

  Next, the process of each stage in Example 2 is demonstrated. A description of the same processing as in the first embodiment will be omitted. First, a flowchart of surviving symbol selection in the first stage process of the second embodiment will be described. FIG. 27 is a diagram illustrating a flowchart of surviving symbol selection in the first stage processing according to the second embodiment.

As illustrated in FIG. 27, the surviving symbol selection unit 328 determines whether or not q ≦ ρ _th (step S701). When q ≦ ρ _th is satisfied (step S701, Yes), the surviving symbol selection unit 328 adds the q column symbols in the ε ⁽¹⁾ row of the symbol higher ranking table 331 to the surviving path (step S702). On the other hand, if q ≦ ρ _th is not satisfied (No in step S701), the surviving symbol selection unit 328 sets the symbol number of the q-ρ _th column in the floor (ε ⁽¹⁾ / N _adj ) row of the symbol lower ranking table 332. It adds to a survival path (step S703).

Next, a flowchart of the kth stage process will be described. FIG. 28 is a flowchart illustrating the k-th stage process according to the second embodiment. As shown in FIG. 28, the k-th stage processing unit sets 1 to i (step S801). Subsequently, the area detection unit performs area detection as shown in Equation 47 (step S802). Subsequently, the region detection unit increments i (step S803). Subsequently, the region detection unit determines whether i ≦ S _k−1 (step S804). If i ≦ S _k−1 (step S804, Yes), the region detection unit returns to step S802 and repeats the process.

On the other hand, if i ≦ S _k−1 is not satisfied (step S804, No), the surviving symbol selection unit survived at the k−1 stage as shown in Equation 49, Equation 50, and Equation 51. The representative cumulative metric value E (i) and the current rank ρ (i) of each surviving path are initialized (step S805). Subsequently, the surviving symbol selection unit selects i _{min at} which ρ (i) ≦ m _{N−k + 1} and the representative cumulative metric E (i) has the minimum value (step S806).

  Subsequently, the surviving symbol selection unit selects a surviving symbol (step S807). The selection of the survival symbol will be described later. Subsequently, the metric calculation unit calculates a cumulative metric as shown in Formula 54 (step S808). Subsequently, the metric calculation unit updates the accumulated metric as shown in Formula 55, Formula 56, and Formula 57 (Step S809).

Subsequently, the metric calculation unit determines whether or not q ≦ S _k (step S810). When q ≦ S _k is satisfied (step S810, Yes), the surviving symbol selection unit returns to step S806 and repeats the process. On the other hand, if q ≦ S _k is not satisfied (step S810, No), the metric calculation unit ends the process.

Next, a flowchart of survival symbol selection in the k-th stage process will be described. FIG. 29 is a flowchart illustrating survival symbol selection in the k-th stage process according to the second embodiment. As illustrated in FIG. 29, the surviving symbol selection unit determines whether or not ρ (i _min ) ≦ ρ _th (step S901).

When ρ (i _min ) ≦ ρ _th is satisfied (step S901, Yes), the surviving symbol selection unit survives symbols in the ρ (i _min ) column of ε ^(k) (i _min ) row of the symbol upper ranking table. (Step S902).

On the other hand, if ρ (i _min ) ≦ ρ _th is not satisfied (step S901, No), the surviving symbol selection unit selects ρ (i) in the floor (ε ^(k) (i _min ) / N _adj ) row of the symbol lower ranking table. The symbol numbers in the _min ) -ρ _th column are added to the surviving path (step S903).

Example 3
Next, Example 3 will be described. In the second embodiment, an example is shown in which the symbol ranking table is divided into two levels, upper and lower, with _ρth as a boundary. However, it can also be divided into more than two levels. In this embodiment, an example of dividing into three levels is shown.

  FIG. 30 is a diagram illustrating a configuration example of the MIMO stream separation unit according to the third embodiment. As shown in FIG. 30, the difference from the first embodiment and the second embodiment is that a symbol intermediate table is added. A description of the same configuration as that of the first and second embodiments will be omitted, and a description will be given focusing on a configuration different from the first and second embodiments.

The surviving symbol selection unit 428 of the first stage processing unit refers to the symbol upper ranking table 421 (Ω ₁ ), the symbol middle ranking table 422 (Ω ₂ ), or the symbol lower ranking table 433 (Ω ₃ ), and starts from the higher ranking. A candidate replica with the number of surviving candidates S1 is set as a surviving pass of the first stage. The surviving symbol selection unit 428 sets the surviving path to several 80 expressions.

  The survival symbol selection unit of the k-th stage processing unit adaptively selects the survival path in the k-th stage as follows. First, the representative cumulative metric value E (i) and the current rank ρ (i) of each surviving path that has survived in the k−1th stage are initialized to formula 81, formula 82, and formula 83.

The surviving symbol selection _{unit, ρ (i) ≦ m N} -k + 1 and Index [rho _{(i min)} position candidate replica symbol upper ranking table of Omega ₁ or symbols median ranking _{i min} of E (i) is the minimum value Select from table Ω ₂ or symbol lower ranking table Ω ₃ . The survival symbol selection unit sets the survival path of the q-th k-th stage to Equation 84.

Next, details of the symbol upper ranking table 421, the symbol middle ranking table 422, and the symbol lower ranking table 433 will be described. Symbol high rank ranking table, symbol middle rank ranking table and symbol designed as ρ _{th, 1} = 4, ρ _{th, 2} = 12, ρ _{adj, 1} = 4, ρ _{adj, 2} = 16 at 16QAM, N _div = 3 An example of the lower ranking table is shown in FIG.

  As shown in FIG. 31, the symbol upper ranking table holds symbol numbers in the order of the distance from the center of the region for each region from the first to the fourth. The symbol middle ranking table holds symbol numbers in the order of the shortest distance from the representative points of the four adjacent areas for each of the four adjacent areas from the fourth to the twelfth positions. The symbol lower ranking table is a representative of eight adjacent regions for each of the 16 adjacent regions (four regions of regions 0 to 15, regions 16 to 31, regions 32 to 47, and regions 48 to 63) from the 12th place to the 16th place. The symbol numbers are held in order of increasing distance from the point.

  Next, the process of each stage in Example 3 is demonstrated. Explanation of the same processing as in the first and second embodiments is omitted. First, a flowchart of surviving symbol selection in the first stage process of the third embodiment will be described. FIG. 32 is a flowchart illustrating survival symbol selection in the first stage process according to the third embodiment.

As illustrated in FIG. 32, the surviving symbol selection unit 428 determines whether or not q ≦ ρ _{th, 1} (step S1001). If q ≦ ρ _{th, 1} (Yes in step S1001), the surviving symbol selection unit 428 adds the symbols in the q column of the ε ⁽¹⁾ row of the symbol higher ranking table 421 to the surviving path (step S1002).

On the other hand, if q ≦ ρ _{th, 1} is not satisfied (step S1001, No), the surviving symbol selection unit 428 determines whether q ≦ ρ _{th, 2} is satisfied (step S1003). If q ≦ ρ _{th, 2} (Yes in step S1003), the surviving symbol selection unit 428 sets q−ρ _{th, 1} column of floor (ε ⁽¹⁾ / N _adj1 ) row of the symbol middle ranking table 422. The symbol number is added to the surviving path (step S1004).

On the other hand, if q ≦ ρ _{th, 2} is not satisfied (step S1003, No), the surviving symbol selection unit 428 sets q−ρ _{th, 2} columns of the floor (ε ⁽¹⁾ / N _adj2 ) row of the symbol lower ranking table. The symbol number is added to the surviving path (step S1005).

Next, a flowchart of survival symbol selection in the k-th stage process will be described. FIG. 33 is a flowchart illustrating survival symbol selection in the k-th stage process according to the third embodiment. As shown in FIG. 33, the surviving symbol selection unit determines whether or not ρ (i _min ) ≦ ρ _{th, 1} (step S1101).

When ρ (i _min ) ≦ ρ _{th, 1} is satisfied (step S1101, Yes), the surviving symbol selection unit selects symbols in the ρ (i _min ) column of ε ^(k) (i _min ) row of the symbol upper ranking table. It adds to a survival path (step S1102).

On the other hand, when ρ (i _min ) ≦ ρ _{th, 1} is not satisfied (step S1101, No), the surviving symbol selection unit determines whether ρ (i _min ) ≦ ρ _{th, 2} is satisfied (step S1103). . When ρ (i _min ) ≦ ρ _{th, 2} is satisfied (step S1103, Yes), the surviving symbol selection unit selects ρ () in the floor (ε ^(k) (i _min ) / N _adj1 ) row of the symbol intermediate ranking table. i _min ) −ρ _th, The symbol number of _one column is added to the surviving path (step S1104).

On the other hand, if ρ (i _min ) ≦ ρ _{th, 2} is not satisfied (step S1103, No), the surviving symbol selection unit ρ in the floor (ε ^(k) (i _min ) / N _adj2 ) row of the symbol lower ranking table. (I _min ) −ρ _th, The symbol numbers in the _two columns are added to the surviving path (step S1105).

  In the present embodiment, only the symbol upper ranking table, symbol middle ranking table, and symbol lower ranking table in a specific parameter example of 16QAM are shown, but the present invention is not limited to this. For example, the table can be designed using the design methods shown in the second and third embodiments in other modulation schemes and other parameters. It is also possible to design the table by dividing it into four or more levels.

Example 4
Next, Example 4 will be described. Example 4 is an example in which Solution 1 and Solution 2 described above are combined. Any of the first, second, and third embodiments may be used for the first stage process. The survival symbol selection unit of the k-th stage processing unit adaptively selects the survival path in the k-th stage as follows. First, the representative cumulative metric value E (i) and the current rank ρ (i) of each surviving path that survived in the k−1th stage are initialized, and are expressed as Expressions 85, 86, 87.

The surviving symbol selection unit, ρ (i) ≦ MAX_RANK _k and E (i) ranking is _{i min} to a minimum value [rho _{(i min)} position symbol upper candidate replica ranking table Omega ₁ or symbol lower ranking table Omega ₂ Select from. In addition, the surviving symbol selection unit sets the survival path of the qth k-th stage to Formula 88.

  The flowchart of the k-th stage is the same as that in the first embodiment. The flowchart for selecting the surviving symbols in the k-th stage is the same as that in the second embodiment, and a description thereof will be omitted.

Here, a table example of 16QAM, N _div = 3, MAX_RANK ^(16QAM) _k = 8, ρ _th = 4, N _adj = 4 will be described. FIG. 34 is a diagram illustrating an example of a symbol upper ranking table and a symbol lower ranking table. In the present embodiment, only the symbol upper ranking table and the symbol lower ranking table in the specific parameter example of 16QAM are shown, but the present invention is not limited to this. For example, the table can be designed by using the design method shown in the first and second embodiments in other modulation schemes and other parameters. It is also possible to design a table by dividing MAX_RANK _k or less into three or more levels.

(Example 5)
Next, Example 5 will be described. In the fifth embodiment, an example of a symbol ranking table suitable for the first stage process is shown. In the first stage process, the symbol ranking table is required to have a lower accuracy than the higher accuracy in order to leave symbol candidates with as small a metric as possible. Contrary to Solution 2 and Examples 2, 3, and 4 described above, the symbol upper ranking table is arranged in the order of the distance from the representative point of the adjacent region, and the symbol lower ranking table is the distance from the representative point for each region. The order is as follows.

The surviving symbol selection unit of the first stage processing unit refers to the symbol upper ranking table Ω ₁ or the symbol lower ranking table Ω ₂ and sets the candidate replica of the surviving candidate number S ₁ from the higher ranking as the first stage survival path. The surviving symbol selection unit sets the surviving path to formula 89.

  Since the flowchart of the first stage process is the same as that of the first embodiment, description thereof is omitted. A flowchart of surviving symbol selection in the first stage process will be described. FIG. 35 is a flowchart illustrating the first stage process according to the fifth embodiment.

As shown in FIG. 35, the surviving symbol selection unit determines whether or not q ≦ ρ _th (step S1201). When q ≦ ρ _th is satisfied (step S1201, Yes), the surviving symbol selection unit adds q columns of symbols in the floor (ε ⁽¹⁾ / N _adj ) row of the symbol upper ranking table to the surviving path (step S1202). ). On the other hand, if q ≦ ρ _th is not satisfied (step S1201, No), the surviving symbol selection unit adds the symbol number of the q-ρ _th column of the ε ⁽¹⁾ row of the symbol lower ranking table to the surviving path (step S1203). ).

Next, a table of ρ _th = 8 and N _adj = 4 with 16QAM, N _div = 3 will be described. FIG. 36 is a diagram illustrating an example of the symbol upper ranking table and the symbol lower ranking table for the first stage processing according to the fifth embodiment. In the fifth embodiment, any of the configurations shown in the first to fourth embodiments may be used for the configuration after the second stage. Further, it can be divided into 3 levels as in the third embodiment. Further, in the fifth embodiment, only the 16QAM table example is shown, but it is obvious that the present invention can also be applied to 64QAM.

Example 6
Next, Example 6 will be described. Example 6 is an example in which the above-described solution 1 is applied to the LSD method. First, the configuration of the MIMO stream separation unit according to the sixth embodiment will be described. FIG. 37 is a diagram illustrating the configuration of the MIMO stream separation unit according to the sixth embodiment. The description of the same configuration as in the first to fifth embodiments is omitted.

  As shown in FIG. 37, the MIMO stream separation unit 512 includes a QR decomposition unit 516, a MIMO demodulation unit 522, and an LLR calculation unit 554. The QR decomposition unit 516 includes a QR decomposition processing unit 518 and a received signal conversion unit 520. The MIMO demodulator 522 includes an LSD processor 524.

  The LSD processing unit 524 includes a region detection unit 526, a surviving symbol selection unit 528, a metric calculation unit 530, and a size reduction type symbol ranking table 532.

Next, processing of the MIMO demodulator 522 of the sixth embodiment will be described. FIG. 38 is a flowchart illustrating the sixth embodiment. As shown in FIG. 38, the region detection unit 526 performs region detection as shown in Equation 47 (step S1301). Subsequently, the LSD processing unit 524 sets ρ (k) to 1 (step S1302). Subsequently, the surviving symbol selection unit 528 selects symbols of ρ (k) columns in ε ^(k) rows of the size reduction type symbol ranking table Ω (step S1303).

  Subsequently, the metric calculation unit 530 calculates a cumulative metric as shown in Formula 54 (step S1304). Subsequently, the metric calculation unit 530 determines whether or not the cumulative metric ≦ Threshold (step S1305). If the cumulative metric ≦ Threshold is not satisfied (step S1305, No), the metric calculation unit 530 returns to the k−1th stage (step S1306).

  On the other hand, when the cumulative metric ≦ Threshold is satisfied (step S1305, Yes), the metric calculation unit 530 determines whether it is the Nth stage (step S1307). If the metric calculation unit 530 is not in the Nth stage (step S1307, No), the metric calculation unit 530 proceeds to the k + 1th stage (step S1308).

On the other hand, when it is the Nth stage (step S1307, Yes), the metric calculation unit 530 adds symbol candidates to the list (step S1309). Subsequently, the metric calculation unit 530 determines whether or not ρ (k) ≦ MAZ_RANK _k is satisfied (step S1310).

If ρ (k) ≦ MAZ_RANK _k is not satisfied (step S1310, No), the metric calculation unit 530 returns to the k−1th stage (step S1311). On the other hand, when ρ (k) ≦ MAZ_RANK _k is satisfied (step S1310, Yes), the metric calculation unit 530 increments ρ (k) (step S1312) and returns to step S1303.

(Example 7)
Next, Example 7 will be described. Example 7 is an example in which Solution 2 described above is applied to the LSD method. First, the configuration of the MIMO stream separation unit according to the seventh embodiment will be described. FIG. 39 is a diagram illustrating the configuration of the MIMO stream separation unit according to the seventh embodiment. The description of the same configuration as in the first to sixth embodiments is omitted.

  As illustrated in FIG. 39, the MIMO stream separation unit 612 includes a QR decomposition unit 616, a MIMO demodulation unit 622, and an LLR calculation unit 654. The QR decomposition unit 616 includes a QR decomposition processing unit 618 and a received signal conversion unit 620. The MIMO demodulator 622 includes an LSD processor 624.

  The LSD processing unit 624 includes a region detection unit 626, a surviving symbol selection unit 628, a metric calculation unit 630, a symbol upper ranking table 631, and a symbol lower ranking table 632.

Next, processing of the MIMO demodulator 622 of Example 7 will be described. FIG. 40 is a flowchart illustrating the seventh embodiment. As shown in FIG. 40, the area detection unit 626 performs area detection as shown in Equation 47 (step S1401). Subsequently, the LSD processing unit 624 sets ρ (k) to 1 (step S1402). Subsequently, the surviving symbol selection unit 628 determines whether or not ρ (k) ≦ ρ _th (step S1403).

Subsequently, when ρ (k) ≦ ρ _th is satisfied (step S1403, Yes), the surviving symbol selection unit 628 selects symbols in the ρ (k) column of the ε ^(k) row of the symbol upper ranking table 631 ( Step S1404). On the other hand, if ρ (k) ≦ ρ _th is not satisfied (No in step S <b> 1403), the surviving symbol selection unit 628 ρ (k) −ρ of the floor (ε ^(k) / N _adj ) row of the symbol lower ranking table 632. Symbols in the _th column are selected (step S1405).

  Subsequently, the metric calculation unit 630 calculates a cumulative metric as shown in Formula 54 (step S1406). Subsequently, the metric calculation unit 630 determines whether or not cumulative metric ≦ Threshold (step S1407). If the cumulative metric ≦ Threshold is not satisfied (step S1407, No), the metric calculation unit 630 returns to the k-1th stage (step S1408).

  On the other hand, when the cumulative metric ≦ Threshold is satisfied (step S1407, Yes), the metric calculation unit 630 determines whether it is the Nth stage (step S1409). If it is not the Nth stage (step S1409, No), the metric calculation unit 630 proceeds to the k + 1th stage (step S1410).

On the other hand, when it is the Nth stage (step S1409, Yes), the metric calculation unit 630 adds symbol candidates to the list (step S1411). Subsequently, the metric calculation unit 630 determines whether or not ρ (k) ≦ m _{N−k + 1} (step S1412).

If ρ (k) ≦ m _{N−k + 1} is not satisfied (step S1412, No), the metric calculation unit 630 returns to the k−1th stage (step S1413). On the other hand, when ρ (k) ≦ m _{N−k + 1} is satisfied (step S1412, Yes), the metric calculation unit 630 increments ρ (k) (step S1414) and returns to step S1403.

  In addition, although Example 7 showed the example which divides | segments a symbol ranking table into two levels, you may divide | segment not only into this but 3 levels or more like Example 3. FIG. Further, as in the fourth embodiment, it can be combined with the sixth embodiment.

  Although seven examples have been described above, the present embodiment is not limited to application to the ASSES method or the LSD method. The present invention is applicable to all MLD systems that perform area detection and have a symbol ranking table.

200 Receiver 202 Reception antenna 204 Reception unit 208 Demodulation unit 210 Channel estimation unit 212, 312, 512, 612 MIMO stream separation unit 214 Error correction decoding unit 216, 316, 516, 616 QR decomposition unit 218, 318, 518, 618 QR Decomposition processing units 220, 320, 520, 620 Received signal conversion units 222, 322, 522, 622 MIMO demodulation units 226, 236, 246, 326, 336, 346, 526, 626 Region detection units 228, 238, 248, 328, 338, 348, 428, 528, 628 Surviving symbol selection unit 230, 240, 250, 330, 340, 350, 530, 630 Metric calculation unit 232 Symbol ranking table 242, 252, 532 Size reduction type symbol ranking table 2 54, 354, 554 LLR calculation unit 262 Memory 331, 341, 351, 631 Symbol upper ranking table 332, 342, 352, 632 Symbol lower ranking table 421 Symbol upper ranking table 422 Symbol middle ranking table 433 Symbol lower ranking table

Claims (4)

A signal converter that converts a received signal received via an antenna into a signal including a product of an upper triangular matrix and a transmission signal;
An area detector for detecting an area on the IQ plane to which the signal converted by the signal converter belongs;
The rank from the first place to the rank equal to the modulation multi-level number is divided into two or more ranges, and in each range, symbol candidates are stored in order of increasing distance from the representative point of the adjacent area for each adjacent area. Symbol ranking table,
A surviving symbol selection unit that selects a symbol candidate based on the region detected by the region detection unit and the symbol ranking table;
A wireless device comprising: A signal converter that converts a received signal received via an antenna into a signal including a product of an upper triangular matrix and a transmission signal;
An area detector for detecting an area on the IQ plane to which the signal converted by the signal converter belongs;
The rank equal to the rank upper limit value set smaller than the modulation multi-value number of the transmission signal is divided into two or more ranges, and in each range, for each adjacent region, from the representative point of the adjacent region. A symbol ranking table storing symbol numbers in order of increasing distance;
A surviving symbol selection unit that selects a symbol candidate based on the region detected by the region detection unit and the symbol ranking table;
A wireless device comprising: Computer
The received signal received through the antenna is converted into a signal including the product of the upper triangular matrix and the transmission signal,
Detect the area on the IQ plane to which the converted signal belongs,
The detected area and the rank from the first place to the order equal to the modulation multi-level number are divided into two or more ranges, and in each range, the distance from the representative point of the adjacent area is short for each adjacent area. A wireless device control method, comprising: executing a process of selecting a symbol candidate based on a symbol ranking table in which symbol candidates are stored in order. Computer
The received signal received through the antenna is converted into a signal including the product of the upper triangular matrix and the transmission signal,
Detect the area on the IQ plane to which the converted signal belongs,
The detected region and the rank equal to the rank upper limit value set smaller than the modulation multi-level number of the transmission signal are divided into two or more ranges, and the adjacent regions are adjacent to each other in each range. A wireless device control method, comprising: executing a process of selecting a symbol candidate based on a symbol ranking table storing symbol numbers in order of increasing distance from a representative point of an area.

Images