Title of Invention	"AN APPARATUS AND METHOD FOR TRANSMITTING INDEPENDENT DATA STREAMS FROM A PLURALITY OF RADIO ANTENNAS"
Abstract	A system and method for transmitting multiple independent data streams from subsets of a plurality of transmit antennas. The method includes determining (11) by a receiver (50), a transmit antenna partitioning, including relative phase rotations to be applied to each transmit antenna, that results in the highest channel capacity among the possible partitionings. The receiver (50) then provides (12) partitioning information to a transmitter (40), including the number of transmit antenna subsets, which transmit antennas are included in each subset, the capacity of the data stream to be transmitted from each antenna subset, and the relative phase rotations to be applied to the antennas in each subset. The transmitter (40) partitions (35) the plurality of transmit antennas into mutually exclusive subsets in accordance with the partitioning information, applies (38) the relative phase rotation to each transmit antenna, and transmits (39) an independent data stream from each subset of transmit antennas with a rate not greater than the stream capacity.

Title of Invention

"AN APPARATUS AND METHOD FOR TRANSMITTING INDEPENDENT DATA STREAMS FROM A PLURALITY OF RADIO ANTENNAS"

Abstract

A system and method for transmitting multiple independent data streams from subsets of a plurality of transmit antennas. The method includes determining (11) by a receiver (50), a transmit antenna partitioning, including relative phase rotations to be applied to each transmit antenna, that results in the highest channel capacity among the possible partitionings. The receiver (50) then provides (12) partitioning information to a transmitter (40), including the number of transmit antenna subsets, which transmit antennas are included in each subset, the capacity of the data stream to be transmitted from each antenna subset, and the relative phase rotations to be applied to the antennas in each subset. The transmitter (40) partitions (35) the plurality of transmit antennas into mutually exclusive subsets in accordance with the partitioning information, applies (38) the relative phase rotation to each transmit antenna, and transmits (39) an independent data stream from each subset of transmit antennas with a rate not greater than the stream capacity.

Full Text	TECHNICAL FIELD OF THE INVENTION The present invention relates to radio communication systems. More particularly, and not by way of limitation, the present invention is directed to an apparatus and method for transmitting independent data streams from a plurality of radio antennas. BACKGROUND The capacity of a wireless channel can be enhanced greatly by utilizing multiple transmit and multiple receive antennas. This has been demonstrated by open-loop multiple-input-multiple-output (MIMO) schemes, such as the Bell Labs Layered Space-Time (BLAST) architecture. Open-loop MIMO schemes do not require any feedback link from the receiver to the transmitter, as in closed-loop MIMO schemes. However, by exploiting instantaneous channel-state information provided to the transmitter by the receiver as in closed-loop MIMO schemes, the channel capacity is further enhanced. Closed-loop MIMO schemes include Eigenmode-BLAST, which is an optimum closed-loop MIMO scheme, and Per Antenna Rate Control (PARC), which achieves the open-loop MIMO capacity. In PARC, an independently coded and modulated data stream is transmitted from each transmit antenna. The coding rate for each data stream is provided by the receiver via a feedback mechanism. While optimum closed-loop MIMO schemes require a large amount of instantaneous channel information to be fed back from the receiver to the transmitter, PARC requires much less feedback. However, PARC does not perform well in certain conditions, such as low signal-to-noise ratio (SNR) conditions, systems with correlated antennas, systems with fewer receive antennas than transmit antennas, and channels with Ricean fading. In such conditions, the MIMO channel often cannot support the transmission of as many data streams as the number of transmit antennas (as done in PARC). This limits the capacity of PARC. This problem was partly solved by Selective-PARC (S- PARC), which is an extension of PARC. S-PARC transmits as many data streams as can be supported by the channel by using the same number of transmit antennas as there are data streams. S-PARC performs better than PARC in the above-mentioned conditions. However, since S-PARC may not utilize all transmit antennas, its capacity is significantly lower than optimum closed-loop MIMO. Thus, what is needed in the art is a closed-loop MIMO scheme that overcomes the deficiencies of conventional systems and methods by utilizing all transmit antennas to transmit an adaptive number of data streams. The present invention provides such a system and method. SUMMARY OF THE INVENTION The present invention is a system and method for implementing in a radio communication system, a closed-loop MIMO scheme that utilizes, in one embodiment, all transmit antennas to transmit an adaptive number of data streams. This is done by partitioning the set of transmit antennas into mutually exclusive subsets to transmit independent data streams after "co-phasing". The term "co-phasing" refers to a procedure in which relative phase rotations are applied to a data stream transmitted from multiple antennas. The concept of co- phasing is similar to beamforming, except that co-phasing is performed based on instantaneous channel-state information fed back to the transmitter by the receiver. The relative phase rotations (co-phasing angles) are designed to maximize the received signal power, which approximately maximizes channel capacity. An iterative algorithm may be used for finding the co-phasing angles. The scheme, referred to herein as "Multiple Stream Co-phasing" (MSC), requires similar complexity and feedback as S-PARC. However, MSC significantly outperforms S-PARC and achieves close to the capacity of optimum closed-loop MIMO for many channels of interest. Thus, in one aspect, the present invention is directed to a method of transmitting multiple independent data streams from subsets of a plurality of transmit antennas. The method includes determining by the receiver, a partitioning of the plurality of transmit antennas that provides the highest channel capacity among all possible partitionings, and providing information regarding the determined partitioning to the transmitter. The transmitter partitions the plurality of transmit antennas into mutually exclusive subsets in accordance with the determined partitioning information, applies a relative phase rotation to each antenna, and transmits an independent data stream from each partitioned subset of antennas. In another aspect, the present invention is directed to a method of transmitting a data stream from a selected subset of a plurality of transmit antennas. The method includes partitioning by the receiver, the plurality of transmit antennas into a plurality of subsets; and calculating by the receiver for each subset, a relative phase rotation to be applied to each transmit antenna in the subset, and a stream capacity for the subset with the relative phase rotations applied. The receiver then selects a subset having the highest stream capacity and sends information to the transmitter regarding the selected subset. The transmitter then transmits the data stream from the selected subset of transmit antennas after applying the relative phase rotations to the transmit antennas in the selected subset. In yet another aspect, the present invention is directed to a radio system for transmitting multiple independent data streams from subsets of a plurality of transmit antennas. A receiver includes means for determining a partitioning of the plurality of transmit antennas that provides the highest channel capacity among all possible partitionings, and means for sending information regarding the determined antenna partitioning to a transmitter. The transmitter includes means for partitioning the plurality of transmit antennas into mutually exclusive subsets in accordance with the determined partitioning information; means for applying a relative phase rotation to each transmit antenna; and means for transmitting an independent data stream from each partitioned subset of transmit antennas. In still yet another aspect, the present invention is directed to a radio system for transmitting a data stream from a selected subset of a plurality of transmit antennas. The system includes a receiver having means for partitioning the plurality of transmit antennas into a plurality of subsets; and means for calculating for each subset, a relative phase rotation to be applied to each transmit antenna in the subset, and a stream capacity for the subset with the relative phase rotations applied. The receiver also selects a subset having the highest stream capacity, and sends information to the transmitter regarding the selected subset. The system also includes a transmitter comprising means for transmitting the data stream from the selected subset of transmit antennas after applying the co-phasing angles to the transmit antennas in the selected subset. BRIEF DESCRIPTION OF THE DRAWINGS In the following section, the invention will be described with reference to exemplary embodiments illustrated in the figures, in which: In the following section, the invention will be described with reference to exemplary embodiments illustrated in the figures, in which: FIG. 1 is a flowchart illustrating the steps of an exemplary overall multi- stream co-phasing process performed by a transmitter and receiver in accordance with the teachings of the present invention; FIG. 2 is a flowchart illustrating the steps of an exemplary process for computing channel-state feedback information by the receiver in accordance with the teachings of the present invention; FIG. 2B is a flowchart illustrating the steps of an alternative exemplary process for computing channel-state feedback information by the receiver in accordance with the teachings of the present invention; FIG. 3 is a flowchart illustrating the steps of an exemplary process for utilizing channel-state feedback information by the transmitter to maximize transmission capacity in accordance with the teachings of the present invention; FIG. 4 is a simplified block diagram of an embodiment of the transmitter of the present invention; FIG. 5 is a simplified block diagram of an embodiment of the receiver of the present invention; and FIG. 6 is a simplified block diagram of an embodiment of the system of the present invention. DETAILED DESCRIPTION FIG. 1 is a flowchart illustrating the steps of an exemplary overall multi- stream co-phasing process performed by a transmitter and receiver in accordance with the teachings of the present invention. At step 11, the receiver determines channel-state information in a process shown in more detail in FIG. 2 and 2B. In a preferred embodiment, the channel-state information includes an optimum antenna partitioning into K subsets, co-phasing angles for each antenna, and stream capacity (i.e., rate) information for the optimum antenna partitioning. At step 12, the receiver sends the channel-state information to the transmitter. At step 13, the transmitter de-multiplexes an input signal into K independent streams based on the antenna-partitioning information received from the receiver. Each stream may be coded and modulated at a rate which is not greater than the stream capacity. Each stream may also be scaled. At step 14, the transmitter partitions the antennas into K subsets and applies the co-phasing angles to each antenna in accordance with the channel-state information. At step 15, the transmitter transmits the K data streams from the K subsets of antennas. Thus, based on instantaneous channel-state information received over a feedback channel from the receiver, the transmitter divides an input data signal into K separate streams and sends parallel sequences of data symbols (streams) to multiple antennas. Different phase rotations applied to each antenna eliminate mutual interference. By transmitting K streams in parallel, the transmission time required to send K symbols is reduced to one channel use. The total transmission rate is the sum of all of the separate stream rates. FIG. 2 is a flowchart illustrating the steps of an exemplary process for computing channel-state feedback information by the receiver in accordance with the teachings of the present invention. The receiver analyzes each possible partition and determines a partition that provides the highest channel capacity (i.e., total transmission rate) among all possible partitionings. If the transmitter does not support all possible partitions, the receiver may analyze only those partitions that are supported by the transmitter. Each partition identifies the number of antenna subsets, which antennas are in each subset, the capacity of the data stream to be transmitted from each antenna subset, and what co-phasing angles are to be applied to the antennas in each subset. The process begins at step 21 with the first partition n out of J partitions to be analyzed. Given the number of transmit and receive antennas, the receiver determines all possible (or desirable) transmit antenna set partitions. The receiver also knows (or estimates) the channel response matrix H and the noise variance are determined using equation (9) N0. At step 22, the co-phasing angles and an iterative technique described below. At step 23, a prefilter weight matrix W(n) is computed using equation (4). At step 24, a prefiltered channel G(n) is computed as the product of W(n) and a channel impulse response matrix H using equation (5). At step 25, the channel capacity C(n) for the first partition is evaluated using equation (6). At step 26, the process determines whether all J partitions have been analyzed. If not, the process iterates n = n+1 at step 27, and returns to step 21 to analyze the next possible partition. If all of the J partitions have been analyzed, the process moves to step 28, where the receiver selects the partition that maximizes capacity At step 29, it is determined whether is equal to one, where is the number of input streams corresponding to the selected partition. If the process moves to step 30 and determines the stream capacity. If the method moves to step 31 where the stream capacity for each input stream is obtained using equation (12). At step 32, the receiver sends predefined channel-state information to the transmitter. FIG. 2B is a flowchart illustrating the steps of an alternative exemplary process for computing channel-state feedback information by the receiver in accordance with the teachings of the present invention. The receiver analyzes each possible partition and determines a partition that provides the highest channel capacity (i.e., total transmission rate) among all possible partitionings. If the transmitter does not support all possible partitions, the receiver may analyze only those partitions that are supported by the transmitter. Each partition identifies the number of antenna subsets, which antennas are in each subset, the capacity of the data stream to be transmitted from each antenna subset, and what co-phasing angles are to be applied to the antennas in each subset. The process begins at step 210 with the first partition n out of J partitions to be analyzed. Given the number of transmit and receive antennas, the receiver determines all possible (or desirable) transmit antenna set partitions. The receiver also knows (or estimates) the channel response matrix H and the noise variance N0. At step 220, the receiver begins with the last data stream in the n-th partition. At step 230, the co-phasing angles are determined using equation (14) and an iterative technique described below. At step 240, the capacity of the k-th data stream is estimated using (17). At step 250, the process determines if all data streams have been evaluated. If not, the process iterates k = kA at step 260. At step 270, the total capacity for the n-th partition is computed using (18). At step 280, the process determines whether all J partitions have been analyzed. If not, the process iterates n = n+1 at step 290, and returns to step 210 to analyze the next possible partition. If all of the J partitions have been analyzed, the process moves to step 300, where the receiver selects the partition that maximizes capacity At step 310, the receiver sends predefined channel-state information to the transmitter. FIG. 3 is a flowchart illustrating the steps of an exemplary process for utilizing channel-state feedback information by the transmitter to maximize transmission capacity in accordance with the teachings of the present invention. At step 34, the transmitter receives the instantaneous channel-state information from the receiver. In one embodiment, the channel-state information includes an optimum antenna partitioning into K subsets, co-phasing angles for each antenna, and stream capacity (i.e., rate) information for the optimum antenna partitioning. At step 35, the transmitter partitions the transmit antennas into K subsets as indicated by the channel-state information. At step 36, an input signal is de- multiplexed into K independent data streams, which are optionally coded, modulated, and scaled at step 37. At step 38, the transmitter applies the co- phasing information to the antennas in each subset, and at step 39, transmits independent data streams in parallel through each antenna subset. The following sections describe in more detail, the processes performed in both the receiver and the transmitter. Consider a communication system with M transmit antennas and N receive antennas (i.e., an M x N MIMO system). Suppose that the channel is quasi-static and flat fading. The baseband system model can be written as: where [x1,x2,...,xM]r is the transmitted signal, [y1,y2,...,yN]r is the received signal, H = {h1.1} is the N x M channel impulse response matrix, and [w1,w2,...,wN]T is a Gaussian noise vector with mean zero and covariance matrix N0IN, where IN denotes an N x N identity matrix. The total average transmit power is pr.. The transmitted signal is obtained as: where [S1,S2,...,SK]1 is the input signal demultiplexed into K„TM streams which can be independently coded and modulated and W is an M x K prefilter weight matrix. The following describes a method for determining the prefilter weight matrix. It is assumed that the channel impulse response matrix H and the noise variance N0 are known at the receiver. A person skilled in the art will appreciate that the true values can be replaced by their estimates. Let A(n) = {ƒÑl(n),ƒÑ2(n),...,ƒÑK(n)} be the n-th partition of the set of transmit antenna indices {1,2,...,M}. The set ƒÑk(n) contains the indices of the antennas used for the transmission of the k-th input stream in the n-th partition. The sets contained in a partition are constrained to be mutually exclusive and have at least one element each. The prefilter weight matrix for the n-th antenna set partition is given by: where is an M x 1 column vector with non-zero elements in rows is the co-phasing mƒÕƒÑk(n) and zero elements in the remaining rows, where angle in radians for the k-th data stream and the m-th transmit antenna (corresponding to the m-th row and the /c-th column of the prefilter weight matrix), and \|ƒÑ\| refers to the size of the set ƒÑ. The prefiltered channel is given by: The capacity of this scheme is given by: The co-phasing angles ƒákm(n) can be found by maximizing channel capacity. It can be seen that at low SNR, this is equivalent to maximizing the trace of the prefiltered channel autocorrelation matrix G"(n)G(n) , given by: to find the co-phasing angles. The trace metric, which equals the total received signal power, is also a non-linear function of the angles. However, the trace metric decouples in the input streams, which enables the co-phasing angles to be found independently for each data stream. Based on the trace metric, the co- phasing angles for the k-th stream can be found by using the partial derivatives: for each mƒÕƒÑk(n). It can be seen that the optimum co-phasing angles satisfy the equations: where Note that for each data stream, the co-phasing angle for one of the transmit antennas can be arbitrarily set to zero (or another value), i.e. for one m'ƒÕƒÑk(n). Also note that equation (9) is non-linear in the co-phasing angles, except when In this case, we have and for ƒÑk(n) = {m',m"}. An iterative algorithm can be used to determine the co-phasing angles for \|ƒÑk{n)\|>2. Initially, all co-phasing angles may be set to zero (or some other value). In each iteration, the angles are updated using equation (9) in a parallel or serial fashion. In the parallel case, the angles in the current iteration are determined using the angles in the previous iteration only, while in the serial case, the angles in the current iteration are determined using the angles in the previous iteration and the previously-updated angles in the current iteration. The receiver performs the process shown above in FIG. 2, and if it is determined at step 29 that the stream capacity for each input stream is obtained at step 31 as: where is the k-th column of the matrix matrix given by The selected partition is signaled to the transmitter along with the is the largest rate in bits per symbol less than or equal to which is supported by the transmitter, set). The amount of feedback required is thus M real coefficients plus a log2 J bit integer, where J is the number of possible (or desirable) partitions. Next we describe an alternative method for finding the co-phasing angles in the present invention. The channel capacity can also be expressed as where Ck(n) is the capacity of the k-th data stream (stream capacity), given by where wk(n) is the k-th column of the matrix W(n), and wk(n) is an Nx(K-k + 1) can be matrix given by wk(n) = [wk(n),wk+1(n),...,wK(n)]. The co-phasing angles found to maximize the channel capacity. However, since the channel capacity is a complex non-linear function of the angles, the optimization is hard to achieve. Instead, we propose to find the co-phasing angles for a data stream that maximize the stream capacity, given the co-phasing angles of the higher-order streams. The scheme is described below in detail. The co-phasing angles for the K-th data stream are first found by maximizing the capacity of the K-th data stream, given by This is equivalent to maximizing the received signal gain for the K-th data stream, given by Before we describe how the angles are actually computed for the K-th data stream, let us consider the metric that has to be maximized in order to compute the angles for the other data streams. Suppose that the co-phasing angles for the k+1,k+2,...,K data streams have been estimated and are given by respectively. Given these estimates, the angles for the k-th data stream can be found by maximizing the corresponding stream capacity, which is equivalent to maximizing the whitened received signal gain, given by data stream, the received signal is white and therefore the whitened received signal gain equals the received signal gain. Next we describe how the co-phasing angles are computed using the above metric. The co-phasing angles for the k-th data stream can be found by using the partial derivatives: Note that for each data stream k, the co-phasing angle for one of the transmit antennas (say m) in the subset ƒÑk(n)can be arbitrarily set to zero (or for ƒÑk(n) = {m',m"}. An iterative algorithm can be used to determine the co-phasing angles for\|ak(n)\|>2 as described earlier. Using the co-phasing angle estimates, the capacity of the k-th data stream is estimated as The total capacity for the n-th partition is estimated as the sum of all stream capacities The partition that maximizes the total capacity is selected by the receiver. FIG. 4 is a simplified block diagram of an embodiment of the transmitter 40 of the present invention. As shown, an input signal 41 is passed to a rates, and co-phasing angles are provided by the receiver. FIG. 5 is a simplified block diagram of an embodiment of the receiver 50 of the present invention. As shown, an input signal y is processed by the receiver utilizing a Minimum Mean Square Error-Decision Feedback Equalizer (MMSE- DFE) successive decoder 511-51 . Bold lines in FIG. 5 denote vectors, and regular lines denote scalars. The MMSE feedforward filter and the feedback filter for the k-th stream (stage) are given by: respectively. The input to the k-th decoder is: The k-th decoder produces an estimate of the k-th input stream ŝk. Example Consider the example of a 4 x 2 MIMO system. From the theory of MIMO channels, we know that an optimum number of input streams that can be transmitted is less than or equal to min(M,N)=2. There are eight possible transmit antenna set partitions in this case: one with single input stream (K=1), the rest with two input streams (K=2). The partitions are: A(1)={ {1,2,3,4}}, A(2)={ {1,2,3}, {4}}, A(3)={ {1,2,4}, {3}}, A(4)={ {1,3,4}, {2}}, A(5)={ {2,3,4}, {1}}, A(6)={ {1,2}, {3,4}}, A(7)={ {1,3}, {2,4}} and A(8)={{1,4},{2,3}}. The prefilter weight matrices corresponding to these antenna set partitions are given below in order. An integer with 3 bits can be used to signal the selected partition to the transmitter. The eight possible outcomes of this integer corresponding to the respectively. The receiver feeds back the channel state information to the transmitter. In a preferred embodiment of the present invention, the feedback information comprises the partition selection (which can be signaled by means of an integer as described above) and the rates and angles of the selected partition. FIG. 6 is a simplified block diagram of an embodiment of the system of the present invention. The system includes the receiver 50 in a first node (Node-1) 52 and the transmitter 40 in a second node (Node-2) 53. Functionally, the receiver may include an optimum partitioning determination unit 54 and a feedback channel 55 in addition to normal receiver components. The optimum partitioning determination unit determines channel-state information for a received signal 56. The channel-state information may include partitioning information relating to an optimum partitioning of the transmit antennas 57 into K subsets. The partitioning determination unit 54 also determines the capacity of the data streams to be transmitted from each subset, and co-phasing angles for the transmit antennas in each subset. The feedback channel 55 sends this information 58 to the transmitter 40. In the transmitter 40, an antenna partitioning demultiplexer 42 de- multiplexes the input signal 41 into K independent streams based on the antenna- partitioning information received from the receiver. Each stream may be coded and modulated at a rate which is not greater than the stream capacity. Each stream may also be scaled. The transmitter partitions the transmit antennas 57 into K subsets, and antenna phase rotation units 45 apply the co-phasing angles to each transmit antenna in accordance with the information provided by the receiver 50. The transmitter then transmits the K data streams 59 from the K subsets of antennas 57. When comparing the complexity of Multiple Stream Co-phasing (MSC) with S-PARC, it can be concluded that the complexities are of the same order. The transmitters for the two schemes require similar complexity, with the only additional requirement for MSC being phase rotation. The receivers for the two schemes are identical. The only difference in complexity between MSC and S- PARC is in how the feedback coefficients are computed. For S-PARC, the feedback coefficients are computed by computing the capacity for all possible subsets of the set of transmit antennas and then computing the individual capacity of each stream (stream capacities) for the chosen subset. For MSC, the feedback coefficients are computed by computing the capacity for all possible partitions of the set of transmit antennas and then computing the stream capacities for the chosen partition. The number of subsets of the set of transmit antennas (except the null set) is 2M-1. It has been found that the number, of antenna set partitions for 2 x 2, 3 x 3, and 4x4 MIMO systems are 2, 5, and 15, respectively. For example, the partitions for a 4 x 4 system (in addition to the partitions given earlier for a 4 x 2 system) are: A(9)={ {1,2}, {3}, {4}}, A(10)={ {1,3}, {2}, {4}}, A(11 )={ {1,4}, {2}, {3}}, A(12)={{2,3}, {1}, {4}}, A(13)={{2,4}, {1}, {3}}, A(14)={{3,4}, {1}, {2}}, and 4(15)={{1},{2},{3},{4}}. For 2 x 2, 3 x 3, and 4x4 MIMO systems, the number of partitions (2, 5, and 15, respectively) is less than or equal to the number of subsets (3, 7, and 15, respectively). The two schemes thus require a similar number of capacity evaluations. An additional requirement for MSC is the computation of co-phasing angles for each partition. A closed-form expression, equation (11), is used for the computation of a single co-phasing angle. An iterative algorithm is used for the computation of multiple co-phasing angles. It has been found that two iterations of the serial iterative algorithm described above provides very good estimates. Thus, the complexities of MSC and S-PARC are of the same order. In another embodiment of the present invention, referred to as selective- MSC (S-MSC), the transmit antennas are utilized selectively as in selective-PARC (S-PARC). In S-MSC, all possible (or desirable) subsets of the set of transmit antennas are considered for transmitting input streams. The subset that achieves the highest capacity is selected. The capacity for each subset S„~{i,2 A/} (except the null set) is found by partitioning the set S into all possible (or desirable) partitions (as in MSC) and evaluating the capacity for the partitions. The capacity for the set S is the maximum capacity over all partitions. S-MSC is useful at low SNR, when it may be better to not waste transmit power on an antenna that has weak signal power (below or not much higher than noise level). The receiver performs the antenna selection for S-MSC by hypothesizing each (or each desirable) subset of transmit antennas and finding the the best partition and the capacity for that subset. The subset that maximizes capacity is selected along with the partition. The receiver then signals this information to the transmitter. In yet another embodiment, only certain (desirable) partitions are considered. For example, for a 4 x 4 MIMO system, one may consider the partitions A(1)={ {1}, {2}, {3}, {4} } and A(2)={1,2,3,4} only. The first partition corresponds to sending four independent input streams on four antennas, and the second partition corresponds to sending one input stream on all four antennas after co-phasing. Other embodiments of the present invention include: (a) using alternative methods (metrics) for generating co-phasing angles, (b) quantizing rate and/or angle information for feedback, (c) using different power scaling of the streams based on the knowledge or an estimate of antenna gain disparity; and (d) using antenna correlation information. It has been found that MSC greatly outperforms S-PARC under many practical conditions including low SNR, correlated receive antennas, systems in which there are fewer receive antennas than transmit antennas, and Ricean fading. This is because, unlike S-PARC, MSC uses all antennas to transmit an adaptive number of data streams with co-phasing, which achieves high SNR. For a 4 x 4 MIMO system with Ricean fading (8 dB line of sight component), MSC obtains a gain of 3.8 dB over S-PARC at the rate of 3 bits per symbol. For many MIMO channels of interest, MSC gets very close to the capacity of an optimum closed-loop MIMO scheme for the entire SNR range. This is achieved with similar complexity and feedback requirements as S-PARC (which are much less than those for optimum closed-loop MIMO). As will be recognized by those skilled in the art, the innovative concepts described in the present application can be modified and varied over a wide range of applications. Accordingly, the scope of patented subject matter should not be limited to any of the specific exemplary teachings discussed above, but is instead defined by the following claims. WE CLAIM : 1. A method in a receiver (50) comprised in a first node (52) in communication with a transmitter (40) in a second node (53) in a radio system, wherein multiple independent data streams are transmitted from subsets of a plurality of transmit antennas (57), said method comprising: determining (22) for each transmit antenna, a co-phasing angle that maximizes the received signal power, said step of determining a co-phasing angle for each transmit antenna being characterized by: computing a prefilter weight matrix for each transmit antenna subset; computing a prefilter channel as a product of the prefilter weight matrix and a channel impulse response matrix; computing a trace metric of a prefilter channel autocorrelation matrix; maximizing the trace metric to decouple the input streams; and determining the co-phasing angles independently for each data stream; determining (28) a partitioning of the plurality of transmit antennas (57) that provides the highest channel capacity among all possible partitionings, wherein the step of determining a partitioning of the plurality of transmit antennas (57) is further characterized by determining the number of transmit antenna subsets, which transmit antennas are included in each subset, and the capacity of the data stream to be transmitted from each transmit antenna subset; and providing (32) to the transmitter (40), information regarding the determined co-phasing angle for each transmit antenna, and information-regarding the transmit antenna partitioning such that the transmitter can partition the plurality of transmit antennas (57) into mutually exclusive subsets in accordance with the determined partitioning information. 2. The method according to claim 1, wherein the step of determining a partitioning of the plurality of transmit antennas (57) is further characterized by: hypothesizing every possible partition of the set of transmit antennas into subsets; for each antenna partitioning, determining the co-phasing angles to be applied to each transmit antenna in each subset; for each transmit antenna partitioning, determining a total channel capacity; and selecting the transmit antenna partitioning that provides the highest channel capacity. 3. The method according to claim 2, whereby every possible partition of the set of transmit antennas into subsets implies every transmit antenna subset that is supported by the transmitter. 4. The method according to claim 1, wherein the step of determining a partitioning of the plurality of transmit antennas (57) into a plurality of subsets is further characterized by partitioning the plurality of transmit antennas into every possible antenna subset. 5. The method according to claim 1, wherein the step of calculating a relative phase rotation for each transmit antenna in each partitioned subset is further characterized by utilizing an iterative algorithm to find the relative phase rotation for each transmit antenna that maximizes received signal strength at the receiver for the subset. 6. A radio receiver in a first node for communicating with a transmitter in a second node, said transmitter having a plurality of transmit antennas, said receiver characterized in comprising: means for determining (511-51K) for each transmit antenna, a co-phasing angle that maximizes the received signal power, said means for determining a co- phasing angle for each transmit antenna being adapted to: compute a prefilter weight matrix for each transmit antenna subset; compute a prefilter channel as a product of the prefilter weight matrix and a channel impulse response matrix; compute a trace metric of a prefilter channel autocorrelation matrix; maximize the trace metric to decouple the input streams; and determine the co-phasing angles independently for each data stream; means for determining (511-51K) a partitioning of the plurality of transmit that provides the highest channel capacity among all possible partitionings, wherein the means for determining the partitioning is adapted to determine the number of transmit antenna subsets, which transmit antennas are included in each subset, and the capacity of the data stream to be transmitted from each transmit antenna subset; and means for selecting a subset having the highest stream capacity; and means for sending to the transmitter, information regarding the determined co-phasing angle for each transmit antenna and the determined partitioning. 7. The radio receiver according to claim 6, wherein the means for determining a partitioning of the plurality of transmit antennas is further characterized by: means for determining a stream capacity for every possible transmit antenna subset; means for determining a total channel capacity for each transmit antenna partitioning; and means for selecting the transmit antenna partitioning that provides the highest channel capacity. 8. The radio receiver according to claim 7, whereby every possible transmit antenna subset implies every transmit antenna subset that is supported by the transmitter. 9. A radio transmitter (40) in a second node (53) for communicating with a receiver (50) in a first node (52), said transmitter having a plurality of transmit antennas (57), wherein the transmitter is characterized in comprising: means for receiving from the receiver, information regarding a determined a co-phasing angle for each transmit antenna and a partitioning of the plurality of transmit antennas (57) into subsets determined by the receiver after determining the co-phasing angles, the determined partitioning providing the highest channel capacity among all possible partitionings; means for partitioning (42) the plurality of transmit antenna's into mutually exclusive subsets in accordance with the determined partitioning information; means for applying (45) a relative phase rotation to each transmit antenna in each subset in accordance with the co-phasing angles received from the receiver; and means for transmitting an independent data stream from each partitioned subset of transmit antennas. 10. A method in a transmitter (40) comprised in a second node (53) in communication with a receiver (50) in a first node (52) in a radio system, wherein multiple independent data streams are transmitted from subsets of a plurality of transmit antennas, said method comprising: receiving from the receiver (50), information regarding a determined a co- phasing angle for each transmit antenna and a partitioning of the plurality of transmit antennas (57) into subsets determined by the receiver after determining the co-phasing angles, the determined partitioning providing the highest channel capacity among all possible partitionings; partitioning (42) the plurality of transmit antennas (57)' into mutually exclusive subsets in accordance with the determined partitioning information; applying (45) a relative phase rotation to each transmit antenna in each subset in accordance with the co-phasing angles received from the receiver; and transmitting an independent data stream from each partitioned subset of transmit antennas. 11. A radio system comprising a receiver (50) according to claim 6 in a first node (52), a transmitter (40) according to claim 9 in a second node (53) in communication with the first node, and a plurality of transmit antennas (57), said transmitter transmitting multiple independent data streams from subsets of the plurality of transmit antennas. A system and method for transmitting multiple independent data streams from subsets of a plurality of transmit antennas. The method includes determining (11) by a receiver (50), a transmit antenna partitioning, including relative phase rotations to be applied to each transmit antenna, that results in the highest channel capacity among the possible partitionings. The receiver (50) then provides (12) partitioning information to a transmitter (40), including the number of transmit antenna subsets, which transmit antennas are included in each subset, the capacity of the data stream to be transmitted from each antenna subset, and the relative phase rotations to be applied to the antennas in each subset. The transmitter (40) partitions (35) the plurality of transmit antennas into mutually exclusive subsets in accordance with the partitioning information, applies (38) the relative phase rotation to each transmit antenna, and transmits (39) an independent data stream from each subset of transmit antennas with a rate not greater than the stream capacity.

Full Text

TECHNICAL FIELD OF THE INVENTION
The present invention relates to radio communication systems. More
particularly, and not by way of limitation, the present invention is directed to an
apparatus and method for transmitting independent data streams from a plurality
of radio antennas.
BACKGROUND
The capacity of a wireless channel can be enhanced greatly by utilizing
multiple transmit and multiple receive antennas. This has been demonstrated by
open-loop multiple-input-multiple-output (MIMO) schemes, such as the Bell Labs
Layered Space-Time (BLAST) architecture. Open-loop MIMO schemes do not
require any feedback link from the receiver to the transmitter, as in closed-loop
MIMO schemes. However, by exploiting instantaneous channel-state information
provided to the transmitter by the receiver as in closed-loop MIMO schemes, the
channel capacity is further enhanced. Closed-loop MIMO schemes include
Eigenmode-BLAST, which is an optimum closed-loop MIMO scheme, and Per
Antenna Rate Control (PARC), which achieves the open-loop MIMO capacity. In
PARC, an independently coded and modulated data stream is transmitted from
each transmit antenna. The coding rate for each data stream is provided by the
receiver via a feedback mechanism.
While optimum closed-loop MIMO schemes require a large amount of
instantaneous channel information to be fed back from the receiver to the
transmitter, PARC requires much less feedback. However, PARC does not
perform well in certain conditions, such as low signal-to-noise ratio (SNR)
conditions, systems with correlated antennas, systems with fewer receive
antennas than transmit antennas, and channels with Ricean fading. In such
conditions, the MIMO channel often cannot support the transmission of as many
data streams as the number of transmit antennas (as done in PARC). This limits
the capacity of PARC. This problem was partly solved by Selective-PARC (S-

PARC), which is an extension of PARC. S-PARC transmits as many data streams
as can be supported by the channel by using the same number of transmit
antennas as there are data streams. S-PARC performs better than PARC in the
above-mentioned conditions. However, since S-PARC may not utilize all transmit
antennas, its capacity is significantly lower than optimum closed-loop MIMO.
Thus, what is needed in the art is a closed-loop MIMO scheme that
overcomes the deficiencies of conventional systems and methods by utilizing all
transmit antennas to transmit an adaptive number of data streams. The present
invention provides such a system and method.
SUMMARY OF THE INVENTION
The present invention is a system and method for implementing in a radio
communication system, a closed-loop MIMO scheme that utilizes, in one
embodiment, all transmit antennas to transmit an adaptive number of data
streams. This is done by partitioning the set of transmit antennas into mutually
exclusive subsets to transmit independent data streams after "co-phasing". The
term "co-phasing" refers to a procedure in which relative phase rotations are
applied to a data stream transmitted from multiple antennas. The concept of co-
phasing is similar to beamforming, except that co-phasing is performed based on
instantaneous channel-state information fed back to the transmitter by the
receiver. The relative phase rotations (co-phasing angles) are designed to
maximize the received signal power, which approximately maximizes channel
capacity. An iterative algorithm may be used for finding the co-phasing angles.
The scheme, referred to herein as "Multiple Stream Co-phasing" (MSC), requires
similar complexity and feedback as S-PARC. However, MSC significantly
outperforms S-PARC and achieves close to the capacity of optimum closed-loop
MIMO for many channels of interest.
Thus, in one aspect, the present invention is directed to a method of
transmitting multiple independent data streams from subsets of a plurality of
transmit antennas. The method includes determining by the receiver, a partitioning

of the plurality of transmit antennas that provides the highest channel capacity
among all possible partitionings, and providing information regarding the
determined partitioning to the transmitter. The transmitter partitions the plurality of
transmit antennas into mutually exclusive subsets in accordance with the
determined partitioning information, applies a relative phase rotation to each
antenna, and transmits an independent data stream from each partitioned subset
of antennas.
In another aspect, the present invention is directed to a method of
transmitting a data stream from a selected subset of a plurality of transmit
antennas. The method includes partitioning by the receiver, the plurality of
transmit antennas into a plurality of subsets; and calculating by the receiver for
each subset, a relative phase rotation to be applied to each transmit antenna in
the subset, and a stream capacity for the subset with the relative phase rotations
applied. The receiver then selects a subset having the highest stream capacity
and sends information to the transmitter regarding the selected subset. The
transmitter then transmits the data stream from the selected subset of transmit
antennas after applying the relative phase rotations to the transmit antennas in the
selected subset.
In yet another aspect, the present invention is directed to a radio system for
transmitting multiple independent data streams from subsets of a plurality of
transmit antennas. A receiver includes means for determining a partitioning
of the plurality of transmit antennas that provides the highest channel capacity
among all possible partitionings, and means for sending information regarding the
determined antenna partitioning to a transmitter. The transmitter includes means
for partitioning the plurality of transmit antennas into mutually exclusive subsets in
accordance with the determined partitioning information; means for applying a
relative phase rotation to each transmit antenna; and means for transmitting an
independent data stream from each partitioned subset of transmit antennas.
In still yet another aspect, the present invention is directed to a radio
system for transmitting a data stream from a selected subset of a plurality of

transmit antennas. The system includes a receiver having means for partitioning
the plurality of transmit antennas into a plurality of subsets; and means for
calculating for each subset, a relative phase rotation to be applied to each
transmit antenna in the subset, and a stream capacity for the subset with the
relative phase rotations applied. The receiver also selects a subset having the
highest stream capacity, and sends information to the transmitter regarding the
selected subset. The system also includes a transmitter comprising means for
transmitting the data stream from the selected subset of transmit antennas after
applying the co-phasing angles to the transmit antennas in the selected subset.
BRIEF DESCRIPTION OF THE DRAWINGS
In the following section, the invention will be described with reference to
exemplary embodiments illustrated in the figures, in which:
In the following section, the invention will be described with reference to
exemplary embodiments illustrated in the figures, in which:
FIG. 1 is a flowchart illustrating the steps of an exemplary overall multi-
stream co-phasing process performed by a transmitter and receiver in accordance
with the teachings of the present invention;
FIG. 2 is a flowchart illustrating the steps of an exemplary process for
computing channel-state feedback information by the receiver in accordance with
the teachings of the present invention;
FIG. 2B is a flowchart illustrating the steps of an alternative exemplary
process for computing channel-state feedback information by the receiver in
accordance with the teachings of the present invention;
FIG. 3 is a flowchart illustrating the steps of an exemplary process for
utilizing channel-state feedback information by the transmitter to maximize
transmission capacity in accordance with the teachings of the present invention;

FIG. 4 is a simplified block diagram of an embodiment of the transmitter of
the present invention;
FIG. 5 is a simplified block diagram of an embodiment of the receiver of the
present invention; and
FIG. 6 is a simplified block diagram of an embodiment of the system of the
present invention.
DETAILED DESCRIPTION
FIG. 1 is a flowchart illustrating the steps of an exemplary overall multi-
stream co-phasing process performed by a transmitter and receiver in accordance
with the teachings of the present invention. At step 11, the receiver determines
channel-state information in a process shown in more detail in FIG. 2 and 2B. In a
preferred embodiment, the channel-state information includes an optimum
antenna partitioning into K subsets, co-phasing angles for each antenna, and
stream capacity (i.e., rate) information for the optimum antenna partitioning. At
step 12, the receiver sends the channel-state information to the transmitter. At
step 13, the transmitter de-multiplexes an input signal into K independent streams
based on the antenna-partitioning information received from the receiver. Each
stream may be coded and modulated at a rate which is not greater than the
stream capacity. Each stream may also be scaled. At step 14, the transmitter
partitions the antennas into K subsets and applies the co-phasing angles to each
antenna in accordance with the channel-state information. At step 15, the
transmitter transmits the K data streams from the K subsets of antennas.
Thus, based on instantaneous channel-state information received over a
feedback channel from the receiver, the transmitter divides an input data signal
into K separate streams and sends parallel sequences of data symbols (streams)
to multiple antennas. Different phase rotations applied to each antenna eliminate
mutual interference. By transmitting K streams in parallel, the transmission time

required to send K symbols is reduced to one channel use. The total transmission
rate is the sum of all of the separate stream rates.
FIG. 2 is a flowchart illustrating the steps of an exemplary process for
computing channel-state feedback information by the receiver in accordance with
the teachings of the present invention. The receiver analyzes each possible
partition and determines a partition that provides the highest channel capacity
(i.e., total transmission rate) among all possible partitionings. If the transmitter
does not support all possible partitions, the receiver may analyze only those
partitions that are supported by the transmitter. Each partition identifies the
number of antenna subsets, which antennas are in each subset, the capacity of
the data stream to be transmitted from each antenna subset, and what co-phasing
angles are to be applied to the antennas in each subset.
The process begins at step 21 with the first partition n out of J partitions to
be analyzed. Given the number of transmit and receive antennas, the receiver
determines all possible (or desirable) transmit antenna set partitions. The receiver
also knows (or estimates) the channel response matrix H and the noise variance

are determined using equation (9)
N0. At step 22, the co-phasing angles
and an iterative technique described below. At step 23, a prefilter weight matrix
W(n) is computed using equation (4). At step 24, a prefiltered channel G(n) is
computed as the product of W(n) and a channel impulse response matrix H using
equation (5). At step 25, the channel capacity C(n) for the first partition is
evaluated using equation (6). At step 26, the process determines whether all J
partitions have been analyzed. If not, the process iterates n = n+1 at step 27, and
returns to step 21 to analyze the next possible partition. If all of the J partitions
have been analyzed, the process moves to step 28, where the receiver selects the

partition that maximizes capacity

At step 29, it is determined whether

is equal to one, where

is the number of input streams corresponding to the

selected partition. If

the process moves to step 30 and determines the

stream capacity. If

the method moves to step 31 where the stream capacity

for each input stream

is obtained using equation (12). At step 32, the

receiver sends predefined channel-state information to the transmitter.
FIG. 2B is a flowchart illustrating the steps of an alternative exemplary
process for computing channel-state feedback information by the receiver in
accordance with the teachings of the present invention. The receiver analyzes
each possible partition and determines a partition that provides the highest
channel capacity (i.e., total transmission rate) among all possible partitionings. If
the transmitter does not support all possible partitions, the receiver may analyze
only those partitions that are supported by the transmitter. Each partition
identifies the number of antenna subsets, which antennas are in each subset, the
capacity of the data stream to be transmitted from each antenna subset, and what
co-phasing angles are to be applied to the antennas in each subset.
The process begins at step 210 with the first partition n out of J partitions to
be analyzed. Given the number of transmit and receive antennas, the receiver
determines all possible (or desirable) transmit antenna set partitions. The receiver
also knows (or estimates) the channel response matrix H and the noise variance
N0. At step 220, the receiver begins with the last data stream in the n-th partition.

At step 230, the co-phasing angles
are determined using equation (14) and
an iterative technique described below. At step 240, the capacity of the k-th data
stream is estimated using (17). At step 250, the process determines if all data
streams have been evaluated. If not, the process iterates k = kA at step 260. At
step 270, the total capacity for the n-th partition is computed using (18). At step
280, the process determines whether all J partitions have been analyzed. If not,
the process iterates n = n+1 at step 290, and returns to step 210 to analyze the
next possible partition. If all of the J partitions have been analyzed, the process
moves to step 300, where the receiver selects the partition that maximizes

capacity

At step 310, the receiver sends predefined channel-state

information to the transmitter.

FIG. 3 is a flowchart illustrating the steps of an exemplary process for
utilizing channel-state feedback information by the transmitter to maximize
transmission capacity in accordance with the teachings of the present invention.
At step 34, the transmitter receives the instantaneous channel-state information
from the receiver. In one embodiment, the channel-state information includes an
optimum antenna partitioning into K subsets, co-phasing angles for each antenna,
and stream capacity (i.e., rate) information for the optimum antenna partitioning.
At step 35, the transmitter partitions the transmit antennas into K subsets as
indicated by the channel-state information. At step 36, an input signal is de-
multiplexed into K independent data streams, which are optionally coded,
modulated, and scaled at step 37. At step 38, the transmitter applies the co-
phasing information to the antennas in each subset, and at step 39, transmits
independent data streams in parallel through each antenna subset. The following
sections describe in more detail, the processes performed in both the receiver and
the transmitter.
Consider a communication system with M transmit antennas and N receive
antennas (i.e., an M x N MIMO system). Suppose that the channel is quasi-static
and flat fading. The baseband system model can be written as:

where [x1,x2,...,xM]r is the transmitted signal, [y1,y2,...,yN]r is the received signal,
H = {h1.1} is the N x M channel impulse response matrix, and [w1,w2,...,wN]T is a
Gaussian noise vector with mean zero and covariance matrix N0IN, where
IN denotes an N x N identity matrix. The total average transmit power is pr..

The transmitted signal is obtained as:

where [S1,S2,...,SK]1 is the input signal demultiplexed into K„TM streams which can
be independently coded and modulated and W is an M x K prefilter weight matrix.
The following describes a method for determining the prefilter weight matrix. It is

assumed that the channel impulse response matrix H and the noise variance N0
are known at the receiver. A person skilled in the art will appreciate that the true
values can be replaced by their estimates.
Let A(n) = {ƒÑl(n),ƒÑ2(n),...,ƒÑK(n)} be the n-th partition of the set of transmit
antenna indices {1,2,...,M}. The set ƒÑk(n) contains the indices of the antennas
used for the transmission of the k-th input stream in the n-th partition. The sets
contained in a partition are constrained to be mutually exclusive and have at least
one element each. The prefilter weight matrix for the n-th antenna set partition is
given by:

where

is an M x 1 column vector with

non-zero elements

in rows

is the co-phasing
mƒÕƒÑk(n) and zero elements in the remaining rows, where
angle in radians for the k-th data stream and the m-th transmit antenna
(corresponding to the m-th row and the /c-th column of the prefilter weight matrix),
and |ƒÑ| refers to the size of the set ƒÑ.
The prefiltered channel is given by:

The capacity of this scheme is given by:

The co-phasing angles ƒákm(n) can be found by maximizing channel capacity.
It can be seen that at low SNR, this is equivalent to maximizing the trace of the
prefiltered channel autocorrelation matrix G"(n)G(n) , given by:

to find the co-phasing angles. The trace metric, which equals the total received
signal power, is also a non-linear function of the angles. However, the trace
metric decouples in the input streams, which enables the co-phasing angles to be
found independently for each data stream. Based on the trace metric, the co-
phasing angles for the k-th stream can be found by using the partial derivatives:

for each mƒÕƒÑk(n).

It can be seen that the optimum co-phasing angles

satisfy the

equations:

where

Note that for each data stream, the co-phasing angle for one of the transmit

antennas can be arbitrarily set to zero (or another value), i.e.

for one

m'ƒÕƒÑk(n). Also note that equation (9) is non-linear in the co-phasing angles,

except when

In this case, we have

and

for ƒÑk(n) = {m',m"}. An iterative algorithm can be used to determine the co-phasing
angles for |ƒÑk{n)|>2. Initially, all co-phasing angles may be set to zero (or some
other value). In each iteration, the angles are updated using equation (9) in a
parallel or serial fashion. In the parallel case, the angles in the current iteration
are determined using the angles in the previous iteration only, while in the serial
case, the angles in the current iteration are determined using the angles in the
previous iteration and the previously-updated angles in the current iteration.
The receiver performs the process shown above in FIG. 2, and if it is

determined at step 29 that

the stream capacity for each input stream

is obtained at step 31 as:

where

is the k-th column of the matrix

matrix given by

The selected partition

is signaled

to the transmitter along with the

is the largest rate in

bits per symbol less than or equal to

which is supported by the transmitter,

set). The amount of feedback required is thus M real coefficients plus a log2 J bit
integer, where J is the number of possible (or desirable) partitions.
Next we describe an alternative method for finding the co-phasing angles in the
present invention. The channel capacity can also be expressed as

where Ck(n) is the capacity of the k-th data stream (stream capacity), given by

where wk(n) is the k-th column of the matrix W(n), and wk(n) is an Nx(K-k + 1)

can be
matrix given by wk(n) = [wk(n),wk+1(n),...,wK(n)]. The co-phasing angles
found to maximize the channel capacity. However, since the channel capacity is a
complex non-linear function of the angles, the optimization is hard to achieve.
Instead, we propose to find the co-phasing angles for a data stream that maximize
the stream capacity, given the co-phasing angles of the higher-order streams. The
scheme is described below in detail.
The co-phasing angles for the K-th data stream are first found by maximizing the
capacity of the K-th data stream, given by

This is equivalent to maximizing the received signal gain for the K-th data stream,
given by

Before we describe how the angles are actually computed for the K-th data
stream, let us consider the metric that has to be maximized in order to compute
the angles for the other data streams. Suppose that the co-phasing angles for the

k+1,k+2,...,K data streams have been estimated and are given by

respectively. Given these estimates, the angles for the k-th data stream

can be found by maximizing the corresponding stream capacity, which is
equivalent to maximizing the whitened received signal gain, given by

data stream, the received signal is white and therefore the whitened received
signal gain equals the received signal gain.
Next we describe how the co-phasing angles are computed using the
above metric. The co-phasing angles for the k-th data stream can be found by
using the partial derivatives:

Note that for each data stream k, the co-phasing angle for one of the
transmit antennas (say m) in the subset ƒÑk(n)can be arbitrarily set to zero (or

for ƒÑk(n) = {m',m"}. An iterative algorithm can be used to determine the co-phasing
angles for|ak(n)|>2 as described earlier.
Using the co-phasing angle estimates, the capacity of the k-th data stream
is estimated as

The total capacity for the n-th partition is estimated as the sum of all stream
capacities

The partition

that maximizes the total capacity is selected by the receiver.

FIG. 4 is a simplified block diagram of an embodiment of the transmitter 40
of the present invention. As shown, an input signal 41 is passed to a

rates, and co-phasing angles are provided by the receiver.

FIG. 5 is a simplified block diagram of an embodiment of the receiver 50 of
the present invention. As shown, an input signal y is processed by the receiver
utilizing a Minimum Mean Square Error-Decision Feedback Equalizer (MMSE-
DFE) successive decoder 511-51 . Bold lines in FIG. 5 denote vectors, and
regular lines denote scalars. The MMSE feedforward filter and the feedback filter
for the k-th stream (stage) are given by:

respectively. The input to the k-th decoder is:

The k-th decoder produces an estimate of the k-th input stream ŝk.
Example
Consider the example of a 4 x 2 MIMO system. From the theory of MIMO
channels, we know that an optimum number of input streams that can be
transmitted is less than or equal to min(M,N)=2. There are eight possible transmit
antenna set partitions in this case: one with single input stream (K=1), the rest
with two input streams (K=2). The partitions are:
A(1)={ {1,2,3,4}}, A(2)={ {1,2,3}, {4}}, A(3)={ {1,2,4}, {3}}, A(4)={ {1,3,4}, {2}},
A(5)={ {2,3,4}, {1}}, A(6)={ {1,2}, {3,4}}, A(7)={ {1,3}, {2,4}} and
A(8)={{1,4},{2,3}}.
The prefilter weight matrices corresponding to these antenna set partitions
are given below in order.

An integer with 3 bits can be used to signal the selected partition to the
transmitter. The eight possible outcomes of this integer corresponding to the

respectively.
The receiver feeds back the channel state information to the transmitter. In
a preferred embodiment of the present invention, the feedback information
comprises the partition selection (which can be signaled by means of an integer
as described above) and the rates and angles of the selected partition.
FIG. 6 is a simplified block diagram of an embodiment of the system of the
present invention. The system includes the receiver 50 in a first node (Node-1) 52
and the transmitter 40 in a second node (Node-2) 53. Functionally, the receiver
may include an optimum partitioning determination unit 54 and a feedback
channel 55 in addition to normal receiver components. The optimum partitioning
determination unit determines channel-state information for a received signal 56.
The channel-state information may include partitioning information relating to an
optimum partitioning of the transmit antennas 57 into K subsets. The partitioning
determination unit 54 also determines the capacity of the data streams to be

transmitted from each subset, and co-phasing angles for the transmit antennas in
each subset. The feedback channel 55 sends this information 58 to the
transmitter 40.
In the transmitter 40, an antenna partitioning demultiplexer 42 de-
multiplexes the input signal 41 into K independent streams based on the antenna-
partitioning information received from the receiver. Each stream may be coded
and modulated at a rate which is not greater than the stream capacity. Each
stream may also be scaled. The transmitter partitions the transmit antennas 57
into K subsets, and antenna phase rotation units 45 apply the co-phasing angles
to each transmit antenna in accordance with the information provided by the
receiver 50. The transmitter then transmits the K data streams 59 from the K
subsets of antennas 57.
When comparing the complexity of Multiple Stream Co-phasing (MSC) with
S-PARC, it can be concluded that the complexities are of the same order. The
transmitters for the two schemes require similar complexity, with the only
additional requirement for MSC being phase rotation. The receivers for the two
schemes are identical. The only difference in complexity between MSC and S-
PARC is in how the feedback coefficients are computed. For S-PARC, the
feedback coefficients are computed by computing the capacity for all possible
subsets of the set of transmit antennas and then computing the individual capacity
of each stream (stream capacities) for the chosen subset. For MSC, the feedback
coefficients are computed by computing the capacity for all possible partitions of
the set of transmit antennas and then computing the stream capacities for the
chosen partition.
The number of subsets of the set of transmit antennas (except the null set)
is 2M-1. It has been found that the number, of antenna set partitions for 2 x 2, 3 x
3, and 4x4 MIMO systems are 2, 5, and 15, respectively. For example, the
partitions for a 4 x 4 system (in addition to the partitions given earlier for a 4 x 2
system) are:

A(9)={ {1,2}, {3}, {4}}, A(10)={ {1,3}, {2}, {4}}, A(11 )={ {1,4}, {2}, {3}},
A(12)={{2,3}, {1}, {4}}, A(13)={{2,4}, {1}, {3}}, A(14)={{3,4}, {1}, {2}}, and
4(15)={{1},{2},{3},{4}}.
For 2 x 2, 3 x 3, and 4x4 MIMO systems, the number of partitions (2, 5,
and 15, respectively) is less than or equal to the number of subsets (3, 7, and 15,
respectively). The two schemes thus require a similar number of capacity
evaluations. An additional requirement for MSC is the computation of co-phasing
angles for each partition. A closed-form expression, equation (11), is used for the
computation of a single co-phasing angle. An iterative algorithm is used for the
computation of multiple co-phasing angles. It has been found that two iterations
of the serial iterative algorithm described above provides very good estimates.
Thus, the complexities of MSC and S-PARC are of the same order.
In another embodiment of the present invention, referred to as selective-
MSC (S-MSC), the transmit antennas are utilized selectively as in selective-PARC
(S-PARC). In S-MSC, all possible (or desirable) subsets of the set of transmit
antennas are considered for transmitting input streams. The subset that achieves
the highest capacity is selected. The capacity for each subset S„~{i,2 A/}
(except the null set) is found by partitioning the set S into all possible (or
desirable) partitions (as in MSC) and evaluating the capacity for the partitions.
The capacity for the set S is the maximum capacity over all partitions. S-MSC is
useful at low SNR, when it may be better to not waste transmit power on an
antenna that has weak signal power (below or not much higher than noise level).
The receiver performs the antenna selection for S-MSC by hypothesizing each (or
each desirable) subset of transmit antennas and finding the the best partition and
the capacity for that subset. The subset that maximizes capacity is selected along
with the partition. The receiver then signals this information to the transmitter.
In yet another embodiment, only certain (desirable) partitions are
considered. For example, for a 4 x 4 MIMO system, one may consider the
partitions A(1)={ {1}, {2}, {3}, {4} } and A(2)={1,2,3,4} only. The first partition
corresponds to sending four independent input streams on four antennas, and the

second partition corresponds to sending one input stream on all four antennas
after co-phasing.
Other embodiments of the present invention include: (a) using alternative
methods (metrics) for generating co-phasing angles, (b) quantizing rate and/or
angle information for feedback, (c) using different power scaling of the streams
based on the knowledge or an estimate of antenna gain disparity; and (d) using
antenna correlation information.
It has been found that MSC greatly outperforms S-PARC under many
practical conditions including low SNR, correlated receive antennas, systems in
which there are fewer receive antennas than transmit antennas, and Ricean
fading. This is because, unlike S-PARC, MSC uses all antennas to transmit an
adaptive number of data streams with co-phasing, which achieves high SNR. For
a 4 x 4 MIMO system with Ricean fading (8 dB line of sight component), MSC
obtains a gain of 3.8 dB over S-PARC at the rate of 3 bits per symbol. For many
MIMO channels of interest, MSC gets very close to the capacity of an optimum
closed-loop MIMO scheme for the entire SNR range. This is achieved with similar
complexity and feedback requirements as S-PARC (which are much less than
those for optimum closed-loop MIMO).
As will be recognized by those skilled in the art, the innovative concepts
described in the present application can be modified and varied over a wide range
of applications. Accordingly, the scope of patented subject matter should not be
limited to any of the specific exemplary teachings discussed above, but is instead
defined by the following claims.

WE CLAIM :
1. A method in a receiver (50) comprised in a first node (52) in
communication with a transmitter (40) in a second node (53) in a radio system,
wherein multiple independent data streams are transmitted from subsets of a
plurality of transmit antennas (57), said method comprising:
determining (22) for each transmit antenna, a co-phasing angle that
maximizes the received signal power, said step of determining a co-phasing angle
for each transmit antenna being characterized by:
computing a prefilter weight matrix for each transmit antenna
subset;
computing a prefilter channel as a product of the prefilter weight
matrix and a channel impulse response matrix;
computing a trace metric of a prefilter channel autocorrelation
matrix;
maximizing the trace metric to decouple the input streams; and
determining the co-phasing angles independently for each data
stream;
determining (28) a partitioning of the plurality of transmit antennas (57) that
provides the highest channel capacity among all possible partitionings, wherein
the step of determining a partitioning of the plurality of transmit antennas (57) is
further characterized by determining the number of transmit antenna subsets,
which transmit antennas are included in each subset, and the capacity of the data
stream to be transmitted from each transmit antenna subset; and
providing (32) to the transmitter (40), information regarding the determined
co-phasing angle for each transmit antenna, and information-regarding the
transmit antenna partitioning such that the transmitter can partition the plurality of
transmit antennas (57) into mutually exclusive subsets in accordance with the
determined partitioning information.

2. The method according to claim 1, wherein the step of determining a
partitioning of the plurality of transmit antennas (57) is further characterized by:
hypothesizing every possible partition of the set of transmit antennas into
subsets;
for each antenna partitioning, determining the co-phasing angles to be
applied to each transmit antenna in each subset;
for each transmit antenna partitioning, determining a total channel capacity;
and
selecting the transmit antenna partitioning that provides the highest channel
capacity.
3. The method according to claim 2, whereby every possible partition
of the set of transmit antennas into subsets implies every transmit antenna subset
that is supported by the transmitter.
4. The method according to claim 1, wherein the step of determining a
partitioning of the plurality of transmit antennas (57) into a plurality of subsets is
further characterized by partitioning the plurality of transmit antennas into every
possible antenna subset.
5. The method according to claim 1, wherein the step of calculating a
relative phase rotation for each transmit antenna in each partitioned subset is
further characterized by utilizing an iterative algorithm to find the relative phase
rotation for each transmit antenna that maximizes received signal strength at the
receiver for the subset.
6. A radio receiver in a first node for communicating with a transmitter
in a second node, said transmitter having a plurality of transmit antennas, said
receiver characterized in comprising:
means for determining (511-51K) for each transmit antenna, a co-phasing
angle that maximizes the received signal power, said means for determining a co-
phasing angle for each transmit antenna being adapted to:

compute a prefilter weight matrix for each transmit antenna subset;
compute a prefilter channel as a product of the prefilter weight
matrix and a channel impulse response matrix;
compute a trace metric of a prefilter channel autocorrelation
matrix;
maximize the trace metric to decouple the input streams; and
determine the co-phasing angles independently for each data
stream;
means for determining (511-51K) a partitioning of the plurality of transmit
that provides the highest channel capacity among all possible partitionings,
wherein the means for determining the partitioning is adapted to determine the
number of transmit antenna subsets, which transmit antennas are included in
each subset, and the capacity of the data stream to be transmitted from each
transmit antenna subset; and
means for selecting a subset having the highest stream capacity; and
means for sending to the transmitter, information regarding the determined
co-phasing angle for each transmit antenna and the determined partitioning.
7. The radio receiver according to claim 6, wherein the means for
determining a partitioning of the plurality of transmit antennas is further
characterized by:
means for determining a stream capacity for every possible transmit
antenna subset;
means for determining a total channel capacity for each transmit antenna
partitioning; and
means for selecting the transmit antenna partitioning that provides the
highest channel capacity.
8. The radio receiver according to claim 7, whereby every possible
transmit antenna subset implies every transmit antenna subset that is supported
by the transmitter.

9. A radio transmitter (40) in a second node (53) for communicating
with a receiver (50) in a first node (52), said transmitter having a plurality of
transmit antennas (57), wherein the transmitter is characterized in comprising:
means for receiving from the receiver, information regarding a determined a
co-phasing angle for each transmit antenna and a partitioning of the plurality of
transmit antennas (57) into subsets determined by the receiver after determining
the co-phasing angles, the determined partitioning providing the highest channel
capacity among all possible partitionings;
means for partitioning (42) the plurality of transmit antenna's into mutually
exclusive subsets in accordance with the determined partitioning information;
means for applying (45) a relative phase rotation to each transmit antenna
in each subset in accordance with the co-phasing angles received from the
receiver; and
means for transmitting an independent data stream from each partitioned
subset of transmit antennas.
10. A method in a transmitter (40) comprised in a second node (53) in
communication with a receiver (50) in a first node (52) in a radio system, wherein
multiple independent data streams are transmitted from subsets of a plurality of
transmit antennas, said method comprising:
receiving from the receiver (50), information regarding a determined a co-
phasing angle for each transmit antenna and a partitioning of the plurality of
transmit antennas (57) into subsets determined by the receiver after determining
the co-phasing angles, the determined partitioning providing the highest channel
capacity among all possible partitionings;
partitioning (42) the plurality of transmit antennas (57)' into mutually
exclusive subsets in accordance with the determined partitioning information;

applying (45) a relative phase rotation to each transmit antenna in each
subset in accordance with the co-phasing angles received from the receiver; and
transmitting an independent data stream from each partitioned subset of
transmit antennas.
11. A radio system comprising a receiver (50) according to claim 6 in a
first node (52), a transmitter (40) according to claim 9 in a second node (53) in
communication with the first node, and a plurality of transmit antennas (57), said
transmitter transmitting multiple independent data streams from subsets of the
plurality of transmit antennas.

A system and method for transmitting multiple independent data streams
from subsets of a plurality of transmit antennas. The method includes determining
(11) by a receiver (50), a transmit antenna partitioning, including relative phase
rotations to be applied to each transmit antenna, that results in the highest
channel capacity among the possible partitionings. The receiver (50) then
provides (12) partitioning information to a transmitter (40), including the number of
transmit antenna subsets, which transmit antennas are included in each subset,
the capacity of the data stream to be transmitted from each antenna subset, and
the relative phase rotations to be applied to the antennas in each subset. The
transmitter (40) partitions (35) the plurality of transmit antennas into mutually
exclusive subsets in accordance with the partitioning information, applies (38) the
relative phase rotation to each transmit antenna, and transmits (39) an
independent data stream from each subset of transmit antennas with a rate not
greater than the stream capacity.

Documents:

http://ipindiaonline.gov.in/patentsearch/GrantedSearch/viewdoc.aspx?id=h7SM1qqMzgdpXTYFWoQcRw==&loc=wDBSZCsAt7zoiVrqcFJsRw==

« Previous Patent

Next Patent »

Patent Number

270670

Indian Patent Application Number

2654/KOLNP/2008

PG Journal Number

02/2016

Publication Date

08-Jan-2016

Grant Date

07-Jan-2016

Date of Filing

01-Jul-2008

Name of Patentee

TELEFONAKTIEBOLAGET LM ERICSSON (PUBL)

Applicant Address

S-164 83 STOCKHOLM

Inventors:

#	Inventor's Name	Inventor's Address
1	HAFEEZ, ABDULRAUF	307 PARK YORK LANE, CARY NORTH CAROLINA 27519

PCT International Classification Number

H04B 7/04

PCT International Application Number

PCT/SE2006/050539

PCT International Filing date

2006-12-05

PCT Conventions:

#	PCT Application Number	Date of Convention	Priority Country
1	11/275,068	2005-12-07	U.S.A.