HK1145234B - Multiple antenna spatial multiplexing optimal detection method and wireless receiver - Google Patents

Description

Multi-antenna space multiplexing optimal detection method and wireless receiver

Technical Field

The present invention relates generally to wireless communications, and more particularly to multi-antenna spatial multiplexing optimal sounding (MASMOD) for soft bit (soft bit) output in a multiple-input multiple-output (MIMO) system.

Background

With the development of wireless communication, the demand for increased bandwidth and increased functionality is becoming stronger. With each generation of wireless systems, speed, bandwidth, and reliability are enhanced. One popular technique for improving overall performance of a wireless system is to use multiple antennas. A multiple-input multiple-output (MIMO) system provides multiple antennas for each transmitter and receiver. For example, a receiver has two or more receive antennas, and a transmitter has two or more transmit antennas. Such a MIMO system is applied to various digital communication module schemes such as Orthogonal Frequency Division Multiple Access (OFDMA), Code Division Multiple Access (CDMA), etc. In addition, such systems provide a basis for new generation wireless local area networks, such as Wi-Fi technology, and broadband wireless access systems, such as Worldwide Interoperability for Microwave Access (WiMAX) technology, and Long Term Evolution (LTE) technology as defined by the third generation partnership project (3 GPP).

One of the most common and promising transmission mechanisms for MIMO systems is spatial multiplexing (spatial multiplexing). Spatial multiplexing is a transmission technique that enables the independent and separate transmission of encoded data signals, referred to as "signal streams," from each transmit antenna. Therefore, the transmission channel space is reused or multiplexed. The maximum order of spatial multiplexing is limited to the minimum number of transmit or receive antennas. Therefore, when there are three receiving antennas and four transmitting antennas connected, the order of spatial multiplexing is 3, and conversely, when there are four receiving antennas and three transmitting antennas connected, the order of spatial multiplexing is also 3. Such multiplexing order means that the multiple by which the signal streams can be transmitted in parallel is equal to the multiplexing order, thereby greatly improving the spectral efficiency (i.e. the number of bits per second that can be transmitted over the radio channel, in hertz (Hz)).

Fig. 1 is a block diagram of a typical MIMO wireless system 10 using spatial multiplexing. The transmitter 100 comprises a processing module 101 which inputs I_nIs processed into bit b₁,b₂，...b_nFor transmission to the receiver 102. Quadrature modulator mod₁,mod₂，...mod_nWill bit b₁,b₂，...b_nModulated to modulation symbols x₁,x₂，...x_nThrough an antenna a_t1,a_t2，...a_tnThe formed transmitting antenna array is transmitted. Quadrature modulator mod₁,mod₂，...mod_nSymbol modulation is performed such that the in-phase/quadrature (I/Q) portion of each constellation point is modulated with two carrier signals that are 90 degrees out of phase. Examples of quadrature modulation schemes are Quadrature Amplitude Modulation (QAM), Quadrature Phase Shift Keying (QPSK), Quadrature Amplitude Shift Keying (QASK), etc.

Due to modulation symbol x₁Is through an antenna a_t1Transmitting, receiving each antenna a of the array_r1,a_r2，...a_rnReception of a received signal y via a particular channel between the antennas₁. The ad hoc channel is represented as a channel matrix h₁₁,h₁₂，...h_1nWherein the channel matrix h_ijRefers to the transmitter antenna a_tjAnd receiver antenna a_riThe channel response in between. Receiving signal y₁,y₂，...y_nExpressed by the mathematical relationship:

wherein n is₁,n₂，...n_nIs formed by an antenna a of a receiver antenna array_r1,a_r2，...a_rnAdditive White Gaussian Noise (AWGN) values.

At the receiver 102, in some types of receivers, from the received signal y₁,y₂，...y_nIn which the transmission bit b is retrieved₁,b₂，...b_nThe first step of (a) is to obtain soft bit outputs by the MIMO detection unit 103. MIMO sounding produces soft bit outputs, i.e., log-likelihood ratios (LLRs) for the bits. Thus, MIMO detection unit 103 generates soft bit output LLR (b)₁),LLR(b₂)，...LLR(b_n). The generation of soft LLR bit outputs can provide optimal sounding results for the next decoding stage of the receiver 102. Decoder 104 decodes and performs Forward Error Correction (FEC) on the soft bit output to retrieve bits b₁,b₂，...b_nThereafter, a Cyclic Redundancy Check (CRC) 105 performs an error check, and then an output O is generated at a processing stage 106. However, this is a MIMO detection process at MIMO detection unit 103, which typically introduces significant complexity to any MIMO system.

At present, there are someThe method is used to perform MIMO sounding in a spatially multiplexed MIMO system. A receiver that decodes or detects symbols in a MIMO system may be considered optimal or suboptimal. An optimal receiver can ensure that two distance variables are detected for calculating soft bit outputs, i.e., LLRs for bits. Any given bit b_iIs determined by the following equation:

where x is a vector of size equal to the number of transmit antennas and y is a vector of size equal to the number of receive antennas. As shown in equation (2), two terms, each of which finds one with a constantValue b_iThe maximum of all possible combinations of x is used to accurately calculate the soft bit output. One of the terms is calculated when b_iWhen =0, and the other term is calculated when b_iAnd (= 1). The two maximum terms are typically calculated using the two minimum distance values between y and Hx, where y is the received signal, H is the channel matrix, and x is the transmitted signal, which is represented by one constellation point. Whereas a sub-optimal receiver, at best, ensures that only one of the two required range variables can be detected or determined. These suboptimal techniques typically estimate and compute a second distance from the determined first distance. An optimal receiver typically examines all possible symbol combinations in each constellation set of the modulation scheme used at transmission. By examining all of these possible combinations, it can be guaranteed that two minimum distance variables are obtained, and using these two variables, the optimal soft bit output can be derived. However, as mentioned above, such a large number of possible combinations is examined, especially in the modulation rates of higher order numbers, which increases the computational complexity of the receiver. An example of an optimal receiver is a receiver that performs MIMO detection using a maximum log-likelihood ratio (MLLR) method. MLLR examines all possible received symbol combinations, and as expected, the optimal solution comes at the cost of high computational complexity.

Many existing detection techniques are suboptimal in view of practical applications because they do not guarantee the detection of the two required distance variables and thus do not produce an optimal solution each time. The difference between these many existing detection techniques is simply how to perform the sub-optimal detection process. These suboptimal receivers can be used in various wireless networks if, on average, the reliability level can provide sufficiently accurate results to meet wireless communication quality standards. For example, a receiver using the enumerated sphere decoder (LSD) method produces a sub-optimal result by examining a variable number of symbol combinations that guarantee only one of the two required distance variables. The second desired variable, however, can only be selected on the basis of one probability, which means that it cannot be guaranteed to be found, that only a certain probability can be found, or that it can be estimated using one or more estimation algorithms on the basis of the determined first distance. In either approach, the second distance variable cannot be guaranteed. The resulting reduction in the combination of required inspections would greatly simplify the computational requirements, but at the expense of probe quality. However, as described above, the LSD technique examines a variable number of symbol combinations. The worse (lower) SNR, the more combinations the LSD needs to examine at the same detection quality requirement. In the worst case, the LSD checks all x combinations, as in the MLLR procedure. The number of combinations examined is also influenced by how a particular QR decomposition/pruning algorithm (tree parsing/pruning algorithms) fits the time-varying channel characteristics.

Another aspect of the computational complexity in the LSD detection method is the use of a tree structure to represent the constellation signals. The LSD represents constellation points in a tree structure to exploit known tree search algorithms to compute distance metrics. However, the process of just parsing these tree structures adds a lot of additional complexity to the receiver.

Between the optimal MLLR and the sub-optimal LSD is the sub-optimal MIMO detection process described in us patent application 2007/0268813 (' 813 application) to Sequans Communications, which is incorporated herein by reference. The process described in the' 813 application is sub-optimal in that it does not examine all possible combinations. It does not have all the computational complexity of MLLR, but it can produce an optimal result from MLLR. However, to reduce computational complexity, the process of the' 813 application admittedly sacrifices some optimality.

Summary of The Invention

Embodiments of the present invention relate to a method of sounding MIMO communications. The method comprises the following steps: selecting one antenna from a plurality of antennas transmitting a received MIMO signal as a first group corresponding to the received MIMO signal at a receiver, wherein the selected antenna is transmitted using a modulation scheme represented by a first constellation diagram having a generally square shape. The method also comprises the following steps: the second set is assigned to include one or more other constellations representing the remaining antennas and a set of distances is estimated between the received MIMO signal and the corresponding transmitted symbols. The estimation includes: a first partial distance between each constellation point in the second set and the received MIMO signal is calculated for each constellation point in the second set, and a second partial distance between each constellation point in the second set and each point in the first set of selected constellation points is calculated. The set of selected constellation points includes a minimum distance constellation point and m additional constellation points, the m additional constellation points are arranged in a crossed manner to enclose the minimum distance point in the constellation diagram, wherein m is a modulation order of the modulation mode. The method also includes: a set of distances is calculated based on a sum of corresponding distances in the first and second partial distances, and two minimum distances are found based on the set of distances.

Other exemplary embodiments relate to a wireless receiver including a main processor, memory coupled to the main processor, receiver circuitry coupled to the main processor, an antenna array having a plurality of antennas, the antenna array coupled to the receiver circuitry, and a MIMO detector in the receiver circuitry. The MIMO detectors are arranged to form a first group comprising a first constellation of generally square shapes. The first constellation group represents a modulation scheme of one of the plurality of antennas. The MIMO detector is also arranged to form a second group comprising one or more further constellations representing the remaining antenna modulation schemes. For each constellation point in the second set, the MIMO detector is further configured to: calculating a first Partial Euclidean Distance (PED) based on each constellation point in the second set₂) Selecting a minimum distance point in the first set, finding m additional points from a lookup table stored in memory, where m is the modulation order of the first set of antennas, calculating a second Partial Euclidean Distance (PED) between each constellation point and each minimum distance point and m additional points in the second set₁) And mix PED₂And PED₁Together into a distance. The m additional points are indexed according to the minimum distance point and cross to enclose the minimum distance point. The MIMO detector is then arranged to determine a soft bit output based on the distance.

Another exemplary embodiment of the present invention relates to a method of sounding MIMO signals in a 2x2 spatial multiplexing MIMO wireless network. The method comprises the following steps: receiving a MIMO signal at a receiver, decomposing a channel matrix using a scaled Givens operator, inserting a plurality of 0's into the channel matrix, and splitting the two transmit antennas into two groups, wherein a first group comprises a generally square-shaped constellation representing a first transmit antenna of the two antennas and a second group comprises another constellation representing the other transmit antenna. The invention also includes: the method further includes mixing the received MIMO signal with the decomposed channel matrix, estimating a set of distances between the decomposed received MIMO signal and the first and second sets of the plurality of constellation points, and calculating a set of distances based on a sum of respective ones of the first and second partial distances. The estimation includes: for each constellation point in the second group, a first partial distance between each constellation point of the second resistance and the decomposed received MIMO signal is calculated, and a second partial distance between each constellation point in the second group and each constellation point in a selected set of constellation points in the first group is calculated. The set of selected constellation points includes a minimum distance point and m additional constellation points, where m is a modulation order of a modulation scheme used by the first transmit antenna.

The foregoing has outlined rather broadly the features and technical advantages of the present invention in order that the detailed description of the invention that follows may be better understood. Additional features and advantages of the invention will be described hereinafter which form a part of the claims. It should be appreciated by those skilled in the art that the conception and specific embodiment disclosed may be readily utilized as a basis for modifying or designing other structures for carrying out the same purposes of the present invention. It should also be realized by those skilled in the art that such equivalent constructions do not depart from the spirit and scope of the invention as set forth in the appended claims. The novel features which are believed to be characteristic of the technology, both as to its organization and method of operation, together with further objects and advantages will be better understood from the following description when considered in connection with the accompanying figures. It is to be expressly understood, however, that each of the figures is provided for the purpose of illustration only and is not intended as a definition of the limits of the present invention.

Drawings

For a more complete understanding of the present invention, reference is now made to the following descriptions taken in conjunction with the accompanying drawing, in which:

fig. 1 is a block diagram of an exemplary MIMO wireless system using spatial multiplexing;

fig. 2 is a block diagram of a spatial multiplexing MIMO wireless system in accordance with an embodiment of the present invention;

FIG. 3 is a detailed block diagram of an enhanced MIMO detector in accordance with an embodiment of the present invention;

fig. 4A is a 16-QAM constellation used in an LCCS procedure during MIMO sounding in accordance with an embodiment of the present invention;

fig. 4B is a flow chart of exemplary steps for implementing a MIMO sounding procedure in connection with the constellation diagram depicted in fig. 4A;

FIG. 5 is a detailed block diagram of an enhanced MIMO detector in accordance with an embodiment of the present invention;

FIG. 6 is a block diagram of a mobile device in accordance with one embodiment of the present invention;

FIG. 7 is a block diagram of a base station in accordance with an embodiment of the present invention;

figure 8 is a block diagram of a MIMO OFDMA wireless network in accordance with one embodiment of the present invention; and

fig. 9 is a flow chart of exemplary steps for operation of a 2x2MIMO system in accordance with an embodiment of the present invention.

Detailed Description

Fig. 2 is a block diagram of a spatial multiplexing MIMO wireless system 20 in accordance with one embodiment of the present invention. The transmitter 200 and the receiver 201 are each provided with an antenna array, each antenna array consisting of two antennas, the transmitter 200 having two antennas a_t1And a_t2The receiver 201 has two antennas a_r1And a_r2. The antenna array and transmission capabilities of transmitter 200 are separated into two separate devices, transmitter 200-1 and transmitter 200-2, where antenna a_t1Is located on the transmitter 200-1 and the antenna a_t2Located at transmitter 200-2. This 2x2 antenna array configuration provides a particular example MIMO setup, operating the MASMOD embodiment, a two-antenna spatial multiplexing optimal detection (DASMOD) system. The operation of transmitter 200 and receiver 201 in a 2x2 configuration is similar to the operation of transmitter 100 and receiver 102 of fig. 1, except that the number of antennas is different. Data bit b₁And b₂At modulator mod₁And mod₂Is modulated to generate a modulation symbol x₁And x₂。mod₁And mod₂Is arranged to use a quadrature modulation scheme having a constellation diagram that generally appears square. Examples of such square constellation modulation schemes are QAM, some low order QPSK schemes, etc. Then, the symbol x is modulated₁And x₂By means of an antenna a_t1And a_t2Transmitted over a frequency channel. The receiver 201 receives the transmission signal as a received signal y₁And y₂. Enhanced MIMO detector 202 will receive signal y₁And y₂Processing into soft bit output LLR (b)₁) And LLR (b)₂) So that the transmission symbols can be finally decoded. The MIMO detection method employed by enhanced MIMO detector 202 examines a fixed number of possible combinations and produces an optimal result that is comparable to the result obtained by the MLLR method, but does not examine all possible combinations or the computational complexity of resolving the large tree structure.

Fig. 3 is a detailed block diagram of enhanced MIMO detector 202 in accordance with one embodiment of the present invention. Enhanced MIMO detector 202, operating under control of processor 30 of receiver 201 (fig. 2), first processes received signal y_iQR decomposition of the channel matrix H is performed. The channel matrix is an expression of the analog channel over which the modulation symbols are transmitted. Which represents the channel gain between the transmitter antenna array and the receiver antenna array. QR decomposition of the channel matrix typically yields an identity orthogonal matrix Q: (unity orthogonal matrix), and an upper triangular matrix R such that the channel matrix can be set according to the following equation:

H=Q*R (3)

in many existing suboptimal MIMO sounding techniques, QR decomposition becomes very complex because in high bandwidth mobile systems, such as WiMAX, the rate of change of the channel matrix H is nearly identical to the received signal y₁The rate of change of (c) is the same. In addition, the decomposition of such equations typically involves a square root operator, which is relatively complex and time consuming process. The LSD method and certain embodiments of the method described in the' 813 application handle these square root operations, performing QR decomposition on the H matrix during its MIMO detection, and therefore, require a high computational complexity.

In the described process of enhancing MIMO detector 202, without performing a typical QR decomposition, enhanced MIMO detector 202 performs a scaled Givens rotation (Givens rotation) on channel matrix H, resulting in upper triangular matrix R. Givens rotations are typically used in numerical linear algebra to produce zero vectors or matrices. However, the standard Givens rotation also handles square root and division operators. In contrast, a scaled Givens rotation zeroes out the elements in the vector or matrix, but does not require square root and division operators. Therefore, applying the scaled Givens rotation operator (SG) to the channel matrix H will produce the upper triangular matrix SR, but without the computational complexity of the standard Givens rotation or standard QR decomposition process. Furthermore, the multiplication of SG by H results in an upper triangular matrix SR in which there is only one element (SR)₁₁) Is a real number, as described in equations (9) and (10). This can be shown by the following formula, from the received signal y_iThe basic equation of (1) begins:

consider that:

multiplying the scaled Givens rotation operator SG on both sides yields:

consider that:

Y_i=SG*y_i(ii) a And (7)

SR=SG*H (8)

Let

And

SR₁₁＝|h₁₁|²+|h₂₁|² (10)

rewriting equation (6) using the relationships established in equations (7), (9), and (10) yields:

as described above, SR₁₁Is a real number which makes the resulting scaling simpler, as shown in equation (17).

For any information bit b_iAll receiver antennas (y) need to be determined₁,y₂，...y_n) And all transmitter antennas (x)₁,x₂，...x_n) The distance between the transmitted modulation symbols. This "determination" is typically made using some distance metric, such as Euclidean Distance (ED), Manhattan Distance (MD), Chebyshev Distance (CD), MD-CD, and the like. The distance metric is typically calculated by finding or estimating the received signal y_iAnd transmitting the modulation symbol x_iThe distance between, or the amount of error or noise, itself using the constellation point s for a particular modulation scheme_iAnd (6) estimating. After determining the ED, the ED²The probability calculation used to find the correct point makes finding a minimum ED equivalent to finding X with the maximum possible.

It should be noted that ED may be the most common distance metric for MIMO detection calculations. However, while ED determinations generally lead to more accurate results, it should be noted that the embodiments of the invention described herein are not limited to the use of only ED. Additional and/or other embodiments may determine distance by MD, CD, MD-CD measurement/calculation, etc.

Unlike the MLLR detection method, which directly calculates ED for all symbol combinations, the method employed by enhanced MIMO detector 202 divides the distance calculation into two independent groups of antennas at the expense of significant computational complexity, and performs partial ED calculation (partial edc) for one of the partial groups using a Last Cross Constellation Set (LCCS) search. Separating ED subset lookups makes the method easier to adapt to (be scaled to) a larger number of antennas. Furthermore, by dividing the antenna groups into a first group of antennas with square shaped modulation constellations and a second group of remaining antennas, some useful calculations can be derived from the previously feasible techniques in single-input single-output (SISO) systems.

To directly solve for ED in MLLR detection methods, for each desired constellation point combination, the following equation is solved:

in a 2x 216-QAM communication link, each constellation has 16 constellation points, which requires solving the above equation 256 times (i.e., 16x 16). In a 64-QAM 2x2 link with 64 constellation points, equation (12) will be solved 4096 times (i.e., 64x 64) by the MLLR method. However, when the ED calculation is split into two antenna groups, where at least one antenna produces a generally square set of modulation constellations, as in QAM modulation, a SISO subset search technique is used, which may save calculations when calculating part of the ED.

In operation, one antenna is selected as the first group C_-last. Generally, it is preferable to select the antenna with the highest modulation order. Antenna a with the highest modulation order_-lastThere are usually the most constellation points, and therefore in MIMO detection methods such as MLLR, the maximum number of point combinations is considered. a is_-lastA modulation scheme with a square constellation set, such as a constellation set of QAM, will also be used. All other antennas, whether or not they are transmitted in QAM, QPSK, etc., are placed in group C_-restAnd (c) removing the residue. In a 2X2 configuration, C_-restConsisting of only one antenna, but for other constructions with more than two antennas, C_-restWill include in addition to a_-lastAll of the antennas of (1).

Returning to FIG. 3, SGR QR decomposition module 300 multiplies the scaled Givens rotation operator SG by the channel matrix H to produce an upper triangular matrix SR. It also generates an orthogonal matrix Q^-1I.e., the inverse of the orthogonal matrix Q, is typically generated during QR decomposition. Considering the relationship between H, Q, SR and SG defined in equations (3) and (8), Q^-1In effect equal to SG, the scaled Givens rotation operator. Therefore, the received signal y_iMixed with SG at multiplier 301, orthogonal matrix Q from SGR QR decomposition module 300^-1. This multiplication follows the relationship of equation (7) to produce Y_i. Based on these relationships, the received signal is now represented by equation (11), as follows:

to begin determining the portion ED, the second line of equation (11) is rewritten as:

Y₂-SR₂₂x₂=SG₂₁*n₁+SG₂₂*n₂ (13)

take the absolute square value of both sides of equation (13) and let the junctionFruit Right side as first Part ED (PED)₂),PED₂Will be calculated according to the following equation:

PED₂＝|Y₂-SR₂₂x₂|² (14)

as shown in equation (14), PED₂Relying only on x₂。

Continue to calculate the second fraction ED (PED)₁) Similarly, the first row of equation (11) is rewritten as:

Y₁-SR₁₁x₁-SR₁₂x₂=SG₁₁*n₁+SG₁₂*n₂ (15)

the absolute square value is again taken for both sides of equation (15) and let equation right side be PED₁，PED₁Is calculated according to the following equation:

PED₁＝|(Y₁-SR₁₂x₂)-SR₁₁x₁|² (16)

wherein x₂Is C_-restConstellation points within the constellation diagram. As shown in equation (16), PED₁Dependent on x₁And x₂However, to find the relevant x₂=s₂All PEDs in time₁，PED₁Only depend on x₁. The combined distance ED is then equal to PED₂And PED₁The sum divided by SR₁₁：

ED＝(PED₁+PED₂)/SR₁₁ (17)

PED₁And PED₂The sum divided by SR₁₁Due to Q generated at SGR QR decomposition 300^-1The matrix is an orthogonal matrix, not an identity matrix. However, in practice, when the following receiver stages, such as the FEC stage, are linear, and SR₁₁Is constant if SR₁₁Is almost constant, thenSo as not to require division by SR₁₁。

When calculating the optimal result for ED, at C_-restEach constellation point combination in the constellation will be used to calculate the partial distance PED₂And C is_-lastThe interior is irrelevant, as shown in formula (14). However, in computing PED₁Then, it is not necessary to calculate C_-lastBecause it is generally a square geometry. Because of this square geometry, the LCCS method can be guaranteed to find two minimum distance calculations that must satisfy equation (2) by only computing a defined constellation point around a conditional minimum distance center point. With two guaranteed minimum ED metrics, an optimal maximum LLR (M-LLR) result can be derived.

Note that a similar approach, Cross Constellation Set (CCS) lookup, has been used previously for SISO systems. However, since there are a plurality of antennas in the MIMO system, it cannot be practically applied to the MIMO system until the present invention is disclosed. Because splitting ED calculates allowed C_-lastConstellation correspondence C for a group considering only one antenna_-restFor each constellation point of a group, the SISO technique may be applied to a MIMO application.

Fig. 4A is a diagram of 16-QAM constellations 400 and 402 used in a 16-QAM x 16-QAMMIMO detection process, in accordance with one embodiment of the present invention. Fig. 4B is a flow chart illustrating exemplary steps for implementing the MIMO sounding process with respect to constellations 400 and 401 in fig. 4A. In step 409, the receiver assigns transmit antennas a with square constellations_t1To the first group C_-last. In step 410, the remaining transmit antennas a_t2Assigned by the receiver to a second group C_-rest. Constellation diagram 400 denotes a_t2And constellation 401 represents a_t1Of the constellation of (a). In step 411, a constellation point s is selected from the constellation diagram 400₂. At step 412, PED is calculated according to equation (14)₂(s₂). At step 413, a conditional minimum distance point tmpY is calculated according to the following equation₁：

tmpY₁=(Y₁-SR₁₂x₂) (18)

At step 414, a near tmpY is selected in the constellation 401₁Point P of_m404. Once P is selected_m404, and then in step 415, a subset P of relevant points in the constellation 401 is determined by looking up the subsets in a lookup table_nm(b_1i). Since there are a limited number of constellation points in the constellation diagram 401, and since the constellation diagram 401 is a generally square shape, a lookup table of relevant points can be generated in advance, which can improve the speed and efficiency of the MIMO detection process.

At step 416, from the point subset and P_m404 selects a point. At step 417, PED is calculated according to the following equation₁(x₁|x₂)：

PED₁(x₁|x₂)=|tmpY₁–SR₁₁x₁|² (19)

For any fixed x₂PED, as shown above in relation to equation (16)₁The calculation depends only on x₁. Therefore, for a fixed x₂Corresponding PED₁The same tmpY can be used for the calculation₁This helps simplify the PED₁And (4) calculating. Further, as shown above, the value SR₁₁Is a real number, which also helps to simplify the PED₁And (4) calculating. Computing PED₁Thereafter, at step 418, the PED is processed in accordance with equation (17)₂And PED₁If the sum is divided by SR₁₁Linear relation and SR depending on next stage₁₁To calculate the overall ED. In step 419, a subset of points or P is determined_mWhether there are other points left to choose. If so, the process from step 416 is repeated. If no points remain, then a determination is made at step 420 as to whether there are any more points in the second group to consider. If so, the process from step 411 is repeated again. Otherwise, the process terminates at step 421. Then, for b_i=1 and b_i=0 calculating the minimum ED distance so as toThe soft bit output can be optimally calculated.

It should be noted that the lookup table of the desired point can be generated in advance by various methods according to the mobile communication system of the particular embodiment of the present invention. For example, consider the number of elements in FIG. 4A, for P_m404 each bit b_1iSelecting points of constellation 401 such that point P_nm(b_1i) Bit b of_1iIs not equal to P_m404, respectively. For example, P_m404 is bit b₁₃,b₁₂,b₁₁,b₁₀= 1010. Therefore, point P_nm(b_1i) At least one bit b of_1iAnd P_mIs not matched (e.g. P)_nm(b_1i) At least one bit b of_1iIs b₁₃=0,b₁₂=1,b₁₁=0, and b₁₀= 1). As shown in FIG. 4A, the conditional minimum distance point P_mThe bitmap of 404 is 1010. Therefore, there will always be an inverse bit compared to the bitmap 1010 for four other constellation points preselected from the lookup table, and the point P_mThe closest. P_mThe relationship between 404 and the bit maps of the other m correlation points is shown in table 1 below.

Table 1

Table 1 above can be used directly as the other m P' s_nmAnd (4) point. This particular choice of the other constellation points of minimum distance depends on the antenna a_-lastThe modulation order of (2). Most wireless networks typically have a modulation rate of 2 orders of magnitude, e.g., 4-QAM,16-QAM,64-QAM,8-QPSK, etc. The modulation order m is equal to the index of the modulation rate, i.e. 2^mThe type of modulation. For example, the modulation order of the 16-QAM system depicted in FIG. 4A is 4, i.e., 2⁴= 16. Thus, the other constellation point P of the constellation 401 to be selected_nmIs 4 (wherein P is_nmIs different from the modulation order'm').

As for generating the lookup table for constellation 401 of FIG. 4A, two (m/2) constellation points, points 406 and 408, will be selected, and the point of minimum distance P_m404 are in the same row. The other two (m/2) constellation points, points 405 and 407, are selected and summed with P_m404 are in the same column. These constellation points 405-408 are then stored in a lookup table by the conditional minimum distance point P_m404 index. P_mThe selection process of these m other constellation points of the lookup table 404 will ensure that the two minimum distance values used to calculate LLR (bi) equation (19) are included in the constellation point P using these four constellation points 405-408 plus the conditional minimum distance_m404 calculated in two combinations. The same process is then repeated for each point on the constellation diagram 401 to complete the lookup table. Thus, regardless of which point on constellation 401 is determined to be conditional minimum distance constellation point P_mThere will be m suitable other points which ensure that two minimum distance values can be obtained.

Referring again to fig. 4B, once 5 points on constellation 401 have been selected, PED proceeds₁Calculation, as shown in step 417 of FIG. 4B. For each point combination, s₂403 and conditional minimum distance point P_m 404,s₂403 and constellation points 405, s₂404 and constellation points 406, s₂404 and constellation points 407, and s₂404 and constellation points 408, both for PED₁And (4) calculating. Then selects the next point of the constellation point 400 and proceeds with PED again between this next point and each point 404-408 of the constellation 401₁And (4) calculating. This process continues until PED is performed between each point in constellation 400 and each constellation point 404-408 in constellation 401₁And (4) calculating. Thus, in general, for a similar 16-QAM 2x2 application, PED is being calculated₁Once, one MLLR process will check 256 constellation pairs, while the example embodiment only checks 80 constellation pairs, both processes being to find the same four soft bit outputs.

The SISO CCS procedure has not been used for MIMO applications because all antennas and their constellations need to be considered when comparing ED's corresponding to different constellation point combinations. The SISO cross-lookup case in CCS involves only one antenna. Therefore, it is not easy to apply it to one MIMO in general. However, since the ED calculations are divided into two groups, where one group has only one antenna, the LCCS search process can now be implemented in MIMO applications as long as one antenna in the group uses a square constellation (e.g., QAM) modulation scheme.

After completing the LCCS procedure, enhanced MIMO detector 202 (fig. 2) pairs each bit b_iA minimum ED value lookup is performed. These minimum ED values are used to determine bit b according to the following formula_iLLR of (a):

LLR(b_i)＝min(ED(b_i＝1))-min(ED(b_i＝0)) (20)

as shown in equation (20), two independent minimum ED values are used to determine an optimal LLR: when b is_iMinimum ED when =1, and b_iMinimum ED when = 0. Without these two minimum ED values, only sub-optimal LLR results can be determined, which can result in higher error rates and lower reliability. Unlike the process employed by enhanced MIMO detector 202 (fig. 2), the LSD MIMO detection process finds only one guaranteed minimum ED value. At LSD, there will be at least one minimum ED, or b_iOf either =0 or b_iOf = 1. Then this minimum value missing will be estimated based on that optimal result. Therefore, the LSD process produces sub-optimal LLR results.

The MIMO detection process described in the' 813 application also produces a suboptimal result because it computes only one range value and then uses that range value to estimate a second range value. The process selects a set of possible symbol values from a constellation diagram, similar to the LSD method, selecting those points within a certain radius. For each selected symbol, a second symbol value is estimated using the values of the first possible symbol. An ED is then generated by calculating the distance between the received signal and a virtual received signal. The virtual received signal is determined by the selected and estimated values of the first and second symbols, respectively. Next, the smallest ED is selected based on that ED, and thereafter, the hard output first and second symbols are selected by decoding the first and second symbols as a function of the smallest ED.

In operation, the process subtracts the contribution of the first selected symbol from the received signal and expresses the resulting signal with a second symbol. Then, an intermediate signal is calculated based on the signal expressed by the second symbol. The intermediate signal is then processed by a threshold detector, sign detector, threshold comparator, etc. This solution or detection provided by the threshold detector is an estimate of the second symbol, which is then used to calculate ED. Thus, the process produces a final result that is suboptimal.

Although the features and advantages described have been disclosed with respect to a 2x2MIMO implementation in fig. 3 and 4, the principles of the present invention may be applied to other MIMO implementations.

Fig. 5 is a detailed block diagram of the enhanced MIMO detector 51 according to an embodiment of the present invention. Enhanced MIMO detector 51 operates in a similar manner to enhanced MIMO detector 202 (fig. 2-4), except that it operates on more than two antennas. To facilitate the present example, the enhanced MIMO detector 51 operates in a three antenna application under the control of the processor 50. Enhanced MIMO detector 51 receives received signal y from its receiver antenna array_i. The channel matrix H is QR decomposed into an orthogonal matrix Q and an upper triangular matrix R at the modulator 500. Q matrix and received signal y_iAre mixed together to form Y_i. The decomposed channel matrix is then summed with the received vector signal Y at element 502_iMix, where ED calculation is performed. In a 2x2 application, the ED calculations are divided into two groups. The first group comprises a square-shaped constellation set for the first transmit antenna, while the second group comprises a constellation set for all other transmit antennas, regardless of their constellation shape. In a 2x2 application, PED is calculated₂And PED₁The LCCS lookup procedure is used to find the appropriate ED value, which can be used to ensure that two optimal minimum ED metrics are found for LLR computation. Like the 2x2 application, LCCS in the 3x3 application also requires the use of no more than a fixed numberThe constellation points of (a), the constellation points crossing around the minimum PED constellation point. Finally, at element 503, a minimum ED lookup is performed for soft bit result LLR (b)_i) And (4) calculating.

The difference between the operation of enhanced MIMO detector 51 (fig. 5) and enhanced MIMO detector 202 (fig. 2& 3) is QR decomposition. Due to the increased number of antennas, enhanced MIMO detector 51 cannot simplify QR decomposition by performing a scaled Givens rotation order. However, QR decomposition may be performed by any known method, including those using square root operators, which may increase computational complexity over particular 2x2MIMO applications. However, even with the added computational complexity, the overall computation is still simpler than the procedure in the MLLR, LSD and' 813 applications, and the procedure can still provide two guaranteed minimum EDs, yet provide optimal soft bit results. Neither LSD nor the process of the' 813 application is capable of providing such optimal results.

It should be noted that processor 30 (fig. 3) and processor 50 (fig. 5) may be configured as a Digital Signal Processor (DSP), Field Programmable Gate Array (FPGA), microprocessor, microcontroller, or the like.

Fig. 6 is a block diagram of a mobile device 60 according to an embodiment of the present invention. The mobile device 60 is a MIMO arrangement having an antenna array 600 of antennas 601-1-601-n. The antenna array 600 provides a transmit/receive mechanism to a wireless Radio Frequency (RF) transceiver 601. Signals received by the antenna array 600 are processed by receiver processing circuitry 603. Various embodiments of the present invention that provide enhanced MIMO sounding are included in the receiver processing circuitry 603. For example, receiver processing circuitry 603 includes a MIMO detector 606, such as those described in fig. 2, 3, and 5, a decoder 607, and a CRC unit 608. The various functional stages provided by receiver processing circuitry 603 operate in conjunction with DSP/FPGA 605. DSP/FPGA 605 may comprise one or more individual DSP/FPGAs 605-1-605-N as a single processor or integrated into a single chip DSP/FPGA having multiple processing cores, and main processor 604.

The main processor 604 also controls and manages the operation of user interface features of the mobile device 60, such as a keypad 609, display 610, microphone 612 and speaker 613, as well as other features of the mobile device 60. The mobile device 60 is operated by the main processor 604 in conjunction with communication of applications and information stored on the memory 614. Memory 614 stores, for example, a basic operating system (O/S) 615, a lookup table 614, and an application 617. One or more of the lookup tables 616 or an application under the application 617 may supplement or configure the operation of the receiver processing circuitry 603 through implementation controlled by the host processor 604 and/or the DSP/FPGA 605.

The main processor 604 and DSP/FPGA 605 also operate to process input information from any user input device, such as a keyboard 609, display 610, and microphone 612 or internal system input/output (I/O) 611, which provides results that are processed locally by an application 617 running on the mobile device 60. This processed input may then be transmitted to transmitter processing circuitry 602 in preparation for transmission of data and information via RF transceiver 601 and antenna array 600.

It will be noted that the present disclosure is not limited by the structure of the mobile device 60. Any combination of hardware and software capable of performing the functions described in connection with enhancing MIMO detection may be used to implement embodiments of the present invention.

Fig. 7 is a block diagram of a base station 70 according to an embodiment of the present invention. As just one mobile end, such as mobile device 60 (fig. 6), may be configured with enhanced MIMO sounding as described herein with respect to the present invention, as well as a base station. The base station 70 includes a main processor 700 that controls the operation and processing of its computer driven functions. An antenna array 700 having antennas 701-1-701-n transmits and receives signals to form a communications network of which the base station 70 is a part. The RF transceiver 702 operates to transmit signals to the antenna array 700 and receive signals from the antenna array 700. Signals to be transmitted are processed or controlled by the main processor 700 and one or more DSP/FPGA cores 706-1-706-n of the DSP/FPGA assembly 706 and pass through transmit circuitry 703 and then prepared by the RF transceiver 702 to the antenna array 701. These transmitted signals are processed by hardware, firmware, and software stored in memory 707.

Signals received by RF transceiver 702 from antenna array 701 are processed by receiver circuitry 704 which includes enhanced MIMO detector 705. The operation of enhanced MIMO detector 705 is consistent with the methods and operations described in the examples of fig. 3 and 5. The detected and decoded signals are then processed or controlled by the main processor 700 and one or more of the DSP/FPGA cores 706-1-706-n using hardware, firmware, and/or software stored in the memory 707. After being so processed, the signals are further transmitted from the base station 70 through the transmitting circuit 703, the RF transceiver 702 and the antenna array 701. However, as previously described, the operation of enhanced MIMO detector 705 is still consistent with the various embodiments described herein.

Fig. 8 is a block diagram of a MIMO OFDMA wireless network according to an embodiment of the present invention. Fig. 8 depicts one base station 800 of a plurality of base stations (not shown) forming MIMO OFDMA wireless network 80. The base station 800 transmits spatially multiplexed MIMO signals to and receives spatially multiplexed MIMO signals from the mobile units MU 802-805 through the antenna array 801. Base station 800 and MU 802-805 each include an enhanced MIMO detector configured to operate in accordance with the operational description provided in connection with fig. 3 and 5. MIMO OFDMA wireless network 80 will provide efficient and reliable communications because each enhanced MIMO detector will provide optimal results for symbol detection with minimal computational complexity.

Fig. 9 is a flowchart illustrating exemplary steps performed in a 2x2MIMO system in accordance with an embodiment of the present invention. In step 900, a MIMO signal is received at a receiver. In step 901, the channel matrix is decomposed into an upper triangular matrix (R) and an inverse (Q) of an orthogonal matrix (Q) by applying a scaled Givens rotation operator (SG)^-1). In step 902, one of two transmit antennas for transmitting MIMO signals is selected as a first group at a receiver, wherein the selected antenna is transmitted using a modulation scheme of a first constellation of square shape. In step 903, a second is established at the receiverA group comprising constellations representing other transmit antennas. At step 904, for each constellation point on the second set, a first partial distance (PED) is calculated according to equation (14)₂). Also, at step 905, a conditional minimum distance point is calculated according to equation (18) for each constellation point in the second set. At step 906, a minimum distance point is selected in the first group that is closest to the conditional minimum distance point. At step 907, other constellation points in the first group are selected using a lookup table based on the selected minimum distance point. At step 908, a second partial distance (PED) is calculated between each constellation point in the second set and each other constellation point selected plus the minimum distance point₁). At step 909, a set of distances is calculated based on the sum of the respective first and second partial distances. At step 910, two minimum distances are found based on the set of distances, one of which is at b_i=0, the other is at b_iAnd = 1. At step 911, the soft bit output is calculated according to equation (19).

Although the present invention and its advantages have been described in detail, it should be understood that various changes, substitutions and alterations can be made herein without departing from the invention as defined by the appended claims. Moreover, the scope of the present application is not intended to be limited to the particular embodiments of the process, machine, manufacture, composition of matter, means, methods and steps described in the specification. As one of ordinary skill in the art will readily appreciate from the disclosure of the present specification, processes, machines, manufacture, compositions of matter, means, methods, or steps, presently existing or later to be developed that perform substantially the same function or achieve substantially the same result as the corresponding embodiments described herein may be utilized according to the present invention. Accordingly, the appended claims are intended to include within their scope such processes, machines, manufacture, compositions of matter, means, methods, or steps.

Claims (14)

1. A method of detecting multiple-input multiple-output (MIMO) communications, the method comprising:

selecting, for a MIMO signal received at a receiver, an antenna from a plurality of antennas from which the received MIMO signal is transmitted as a first group, wherein the selected antenna is transmitted using a modulation scheme represented by a first constellation diagram of generally square shape;

designating a second group using one or more other constellations representing remaining antennas of the plurality of antennas;

estimating a distance between a set of said received MIMO signals and a corresponding transmitted symbol, wherein for each constellation point in said second set said estimating comprises:

calculating a first partial distance between said each constellation point in said second set and said received MIMO signal;

in the estimation, a channel matrix is decomposed into an upper triangular R matrix and an inverse Q matrix of an orthogonal Q matrix before calculating the first partial distance^-1Wherein the decomposition comprises applying a scaled Givens rotation operator to the channel matrix to create an R matrix and a Q matrix^-1The matrix is a matrix of a plurality of matrices,

rotating the received MIMO signal and the R matrix based on the scaling Givens rotation operator,

a conditional minimum distance point tmpY is calculated according to the following equation₁：

tmpY₁=(Y₁–SR₁₂x₂)

Y₁，SR₁₂，x₂As parameters in equation (11) representing the received MIMO signal, the following:

(11) wherein SG is a Givens rotation operator, whereinIs an upper triangular matrix generated by multiplying the Givens rotation operator SG by the channel matrix H; and

selecting a minimum Euclidean distance constellation point by selecting one of a plurality of points in the first constellation map which is closest to the calculated conditional minimum distance point;

calculating a second partial distance between each constellation point in the second set and each constellation point in a set of selected constellation points in the first set, wherein the set of selected constellation points includes the minimum euclidean distance constellation point and m other constellation points arranged crosswise around the minimum euclidean distance constellation point in the constellation diagram, where m is a modulation order of the modulation scheme;

calculating a set of distances based on respective distance sums of the first and second partial distances; and

based on the set of distances, finding two minimum distances, the minimum distances being minimum distances determined according to an Euclidean Distance (ED) metric;

an optimal soft bit result is calculated from the two minimum distances.

2. The method of claim 1, wherein the receiver is positioned at:

a mobile station; or

A base station.

3. The method of claim 1, wherein the m other constellation points are selected from a lookup table indexed by the minimum euclidean distance constellation point.

4. A wireless receiver, comprising:

a main processor;

a memory connected to the main processor;

a receiver circuit connected to the main processor;

an antenna array having a plurality of antennas, said antenna array being connected to said receiver circuit;

a multiple-input multiple-output, MIMO, detector in the receiver circuit, the MIMO detector dividing the distance calculations into two separate sets of antennas, arranged to:

forming a first group comprising a first constellation diagram having a generally square shape, the first constellation diagram representing a modulation scheme of one of the plurality of antennas;

forming a second group comprising one or more other constellations representing the modulation schemes for the remaining ones of the plurality of antennas;

for each constellation point in the second set, the MIMO detector is further configured to:

calculating a first partial Euclidean distance, PED, based on said each constellation point in said second set₂;

A conditional minimum distance point tmpY is calculated according to the following equation₁：

tmpY₁=(Y₁–SR₁₂x₂)

Y₁，SR₁₂，x₂As parameters in equation (11) representing the received MIMO signal, the following:

(11) wherein SG is a Givens rotation operator, whereinIs an upper triangular matrix generated by multiplying the Givens rotation operator SG by the channel matrix H; selecting a point closest to the calculated conditional minimum distance point from among the plurality of points in the first constellation, and selecting a minimum Euclidean distance constellation point;

deriving m additional points from a lookup table stored in memory, the m additional points indexed according to the minimum euclidean distance constellation point, wherein the m additional points intersect to enclose the minimum euclidean distance constellation point, and m is a modulation order of the one of the first plurality of antennas;

calculating a second partial Euclidean distance PED between each constellation point of the second group and the minimum Euclidean distance constellation point and each m additional points₁(ii) a And

european distance PED of the first part₂And said second part Euclidean distance PED₁Converge to a distance; and

determining a soft bit output based on the distance.

5. The wireless receiver of claim 4, wherein the wireless receiver is a 2x2 receiver-transmitter approach, the MIMO probe further comprising:

a decomposition section arranged to apply a scaled Givens operator to a channel matrix representing a channel on which received MIMO signals are received by said antenna array, wherein application of said scaled Givens rotation operator results in an upper triangular R matrix and an inverse orthogonal Q matrix of an orthogonal Q matrix of said channel matrix^-1。

6. The wireless receiver of claim 4, wherein the MIMO probe is further configured to:

finding two minimum Euclidean distances based on the distance, wherein one minimum Euclidean distance in the two minimum Euclidean distances represents b_iA value of =1, and the other of the two minimum euclidean distances represents when b is_iValue when = 0; and

calculating a minimum likelihood ratio LLR based on the two minimum Euclidean distances, wherein LLR comprises the soft bit output.

7. The wireless receiver of claim 4, wherein the MIMO probe is further configured to:

the modulation schemes of the plurality of antennas transmitting the received MIMO signal are detected.

8. The wireless receiver of claim 4, wherein the wireless receiver provides receive functionality to:

a mobile unit; and

a base station.

9. The wireless receiver of claim 4, further comprising:

one or more additional processors coupled to the main processor, wherein the main processor supplementally controls the MIMO probe using the one or more additional processors.

10. The wireless receiver of claim 9, wherein the one or more additional processors comprise one of:

a Digital Signal Processor (DSP); or

A Field Programmable Gate Array (FPGA).

11. A method of detecting multiple-input multiple-output (MIMO) signals in a 2x2 spatial multiplexing MIMO wireless network, the method comprising:

receiving a MIMO signal at a receiver;

inserting a plurality of 0's into a channel matrix by using a scaled Givens rotation operator to decompose the channel matrix;

splitting the two transmit antennas into two groups, wherein a first group comprises a generally square constellation representing a first transmit antenna of the two transmit antennas and a second group comprises another constellation representing another transmit antenna of the two transmit antennas;

mixing the received MIMO signal with the decomposed channel matrix;

estimating a set of distances between the mixed received MIMO signal and the first and second sets of the plurality of constellation points, wherein for each constellation point in the second set the estimating comprises:

calculating a first partial distance between said each constellation point in said second set and said mixed received MIMO signal;

calculating a conditional minimum distance point tmpY according to the following equation based on said decomposed channel matrix and said received MIMO signal rotated by said scaled Givens rotation operator₁：

tmpY₁=(Y₁–SR₁₂x₂)

Y₁，SR₁₂，x₂As parameters in equation (11) representing the received MIMO signal, the following:

(11) which isWherein SG is Givens rotation operator, whereinIs an upper triangular matrix generated by multiplying the Givens rotation operator SG by the channel matrix H; and

selecting the minimum Euclidean distance constellation point by selecting one point closest to the calculated conditional minimum distance point from a plurality of points in the common square constellation map of the first group;

calculating a second partial distance between each constellation point in the second set and each constellation point in a set of selected constellation points in the first set, wherein the set of selected constellation points includes the minimum euclidean distance constellation point and m other constellation points arranged crosswise around the minimum euclidean distance constellation point, where m is a modulation order of a modulation scheme used by the first transmit antenna; and

calculating a set of distances based on a sum of respective ones of the first and second partial distances;

and searching two minimum Euclidean distances based on the distance set, and calculating an optimal soft bit result according to the two minimum Euclidean distances.

12. The method of claim 11, wherein the m other constellation points are selected from a lookup table using the minimum euclidean distance constellation point.

13. The method of claim 11, further comprising:

for in the second groupAnd finding out two minimum Euclidean distances according to the distance set for each constellation point in the constellation point, wherein one minimum Euclidean distance in the two minimum Euclidean distances represents b_iA value of =1, and the other represents when b_iValue at = 0.

14. The method of claim 11, wherein the receiver is configured to:

a mobile station; or

A base station.