HK1122922B - Method and system for processing communication signals in mimo precoder - Google Patents

Description

Method and system for processing communication signals in MIMO precoder

Technical Field

The present invention relates to signal processing for communication systems. More particularly, the present invention relates to a method and system for codebook (codebook) design for MIMO precoders with finite rate channel state information feedback.

Background

Mobile communications have changed the way people communicate, and mobile phones have become a necessity in their daily lives from their original luxury. The use of mobile phones today is dictated by social status rather than being hindered by geographic location or technology. While voice connections fulfill the basic needs of communications, with mobile voice connections continuing to penetrate even further into each day of life, the mobile internet will be the next step in the mobile communications revolution. The mobile internet naturally becomes a public source of everyday information and simple and universal mobile access to this data service will be the necessary way.

Third generation (3G) cellular networks are specifically designed to fulfill these requirements of the mobile internet. Due to the increased use and popularity of these services, factors such as low cost network capacity optimization and quality of service (QoS) will become more important to cellular network operators than is currently the case. These factors can be achieved through careful network planning and operation, improved transmission methods, and improved receiver techniques. To this end, carriers need technology that enables them to increase downstream throughput, thus providing high QoS performance and speed to compete with services provided by cable modems and/or DSL service providers.

To meet these demands, communication systems that place multiple antennas at both the transmitter and the receiver have recently attracted considerable attention because they can greatly increase the capacity of the system in a wireless fading environment. These multiple antenna structures, also known as smart antenna technology, may be used to reduce multipath and/or signal interference during signal reception. It is anticipated that smart antenna technology will be used more widely as base station facilities are deployed and the processing capacity requirements of mobile handset users in cellular systems increase for these systems. These requirements are due in part to the transition from voice-based services to next generation wireless multimedia services that will provide voice, video and data communication services.

The use of multiple transmit and/or receive antennas is designed to introduce diversity gain and to increase the degree of freedom to combat interference generated in the signal reception process. Diversity gain improves system performance by increasing the received signal-to-noise ratio and stabilizing the transmit chain. Viewed another way, more degrees of freedom allow multiple simultaneous transmissions, providing robustness against signal interference (robustness), and/or allowing higher frequency reuse for higher capacity. For example, in a communication system with multiple antenna receivers, M sets of receive antennas may be used to completely cancel (M-1) interference effects. Thus, N signals can be transmitted simultaneously within the same bandwidth using N transmit antennas, which are then divided into N individual signals by a set of N antennas configured at the receiver. A system using a plurality of transmission and reception antennas may be referred to as a Multiple Input Multiple Output (MIMO) system. An attractive aspect of multi-antenna systems in particular MIMO systems is that the capacity of the system can be increased to a large extent by using these transmission structures. For a fixed total transmit power and bandwidth, the capacity that a MIMO architecture can provide varies dynamically with increasing signal-to-noise ratio (SNR). For example, in the case of a multipath fading channel, the MIMO architecture may increase the capacity of the system to nearly M bits/cycle for every 3dB increase in SNR.

The increase in cost due to the increase in volume, the increase in complexity, and the increase in power consumption has limited the widespread use of multiple antenna systems in wireless communications. This is a concern for wireless system design and application. As a result, some studies on multi-antenna systems may be focused on a point-to-point link that can support a single user, and other studies may be focused on a multi-user scenario. Communication systems using multiple antennas can improve the capacity of the system to a large extent. MIMO technology is used in order to obtain very large performance gains, which however require providing channel information to the transmitter. Such channel data is called channel state information. In many wireless systems, the uplink and downlink operate in Frequency Division Duplex (FDD) mode, i.e., the uplink and downlink use different frequencies. When this is the case, the channel measurement results for the uplink may not be applicable for the downlink and vice versa. In these cases, the channel is measured only by the signal receiver and channel state information is fed back to the transmitter. In the case of many antennas, the amount of data to be transmitted in the uplink channel for feeding back the state information is large. Since the uplink channel bandwidth is limited, it is not desirable to transmit a large amount of channel state information in the uplink channel.

Limitations and disadvantages of conventional and traditional approaches will become apparent to one of ordinary skill in the art, through comparison of such aspects with some aspects of the present invention as set forth in the remainder of the present application with reference to the drawings.

Disclosure of Invention

A method and/or system for codebook design for MIMO precoders with limited-rate channel state information feedback, substantially as shown in and/or described in connection with at least one of the figures, as set forth more completely in the claims.

According to one aspect of the invention, a method for processing a communication signal comprises:

quantizing channel state information in the MIMO precoding system onto a codebook using a cost function, the codebook comprising one or more unitary matrices; and

iteratively updating the codebook based on at least the channel state information.

Preferably, the channel state information is a matrix V.

Preferably, the cost function f (a) is defined by the following relation:

where A is an NxN matrix, a_ijIs the element (i, j) of matrix a.

Preferably, the method further comprises generating Voronoi (Voronoi) regions from the codebook for the one or more unitary matrices.

Preferably, the method further comprises generating a set of matrices based on the voronoi region and the unitary matrix.

Preferably, the method further comprises updating the codebook by modifying the set of matrices to a new set of unitary matrices, wherein the new set of unitary matrices is integrated into the codebook.

Preferably, the method further comprises transmitting an index of an element in the codebook from a receiver to a transmitter in the MIMO precoding system, wherein the channel state information is quantized onto the codebook.

Preferably, the MIMO precoding system comprises one or more transmit antennas and one or more receive antennas.

Preferably, the method further comprises generating the matrix V using Singular Value Decomposition (SVD).

Preferably, the method further comprises generating the matrix V using Geometric Mean Decomposition (GMD).

Preferably, the method further comprises linearly transforming the vector signal at the transmitter in the MIMO precoding system using one of the unitary matrices.

According to one aspect of the invention, a system for processing a communication signal comprises:

a MIMO precoding system comprising one or more circuits that quantize channel state information onto a codebook comprising one or more unitary matrices using a cost function and iteratively update the codebook based on at least the channel state information.

Preferably, the channel state information is a matrix V.

Preferably, the cost function f (a) may be defined by the following relation:

where A is an NxN matrix, a_ijIs the element (i, j) of matrix a.

Preferably, the one or more circuits generate voronoi regions from the codebook for the one or more unitary matrices.

Preferably, the one or more circuits generate a set of matrices based on the voronoi region and the unitary matrix.

Preferably, the one or more circuits update the codebook by modifying the set of matrices to a new set of unitary matrices, wherein the new set of unitary matrices is integrated into the codebook.

Preferably, the one or more circuits are configured to transmit an index of an element in the codebook from a receiver to a transmitter in the MIMO precoding system, wherein the channel state information is quantized onto the codebook.

Preferably, the MIMO precoding system comprises one or more transmit antennas and one or more receive antennas.

Preferably, the one or more circuits generate the matrix V using Singular Value Decomposition (SVD).

Preferably, the one or more circuits generate the matrix V using Geometric Mean Decomposition (GMD).

Preferably, the one or more circuits linearly transform the vector signal at the transmitter in the MIMO precoding system using one of the unitary matrices.

These and other advantages, aspects, and novel features of the present invention, as well as specific embodiments thereof, will be more fully understood from the following description and drawings.

Drawings

FIG. 1A is a schematic diagram of cellular multipath communications between a base station and a mobile computing terminal incorporating an embodiment of the present invention;

FIG. 1B is a diagram of a MIMO communication system in accordance with an embodiment of the present invention;

fig. 2 is a block diagram of a MIMO precoding transceiver chain model in accordance with an embodiment of the present invention;

FIG. 3 is a block diagram of a MIMO precoding system with finite rate channel state information feedback in accordance with an embodiment of the present invention;

FIG. 4 is a flow chart of an implementation of a codebook algorithm according to an embodiment of the present invention;

figure 5 is a graph of the performance of a 2 x 2MIMO system with an MMSE receiver and finite rate feedback in accordance with an embodiment of the invention;

figure 6 is a graph of the performance of a 2 x 2MIMO system with a GMD-OSIC receiver and finite rate feedback in accordance with an embodiment of the present invention.

Detailed Description

Embodiments of the present invention introduce some aspects of a method and system for codebook design for MIMO precoders with finite-rate channel state information feedback. Methods and systems for codebook design for MIMO precoders with finite rate channel state information feedback include quantizing channel state information in a MIMO precoding system onto a codebook using a cost function, the codebook including one or more unitary matrices; and iteratively updating the codebook based on at least the channel state information. The channel state information comprises a matrix V, and the cost function f (a) is defined by the following relation:

where A is an NxN matrix, a_ijIs the element (i, j) of matrix a. For the unitary matrix, Voronoi regions may be generated from the codebook. A set of matrices may be generated based on the Voronoi regions and the unitary matrix. Updating the codebook may be achieved by modifying the set of matrices into a new set of unitary matrices, wherein the new set of unitary matrices is integrated into the codebook. Transmitting, by a receiver to a transmitter, an index of an element in the codebook onto which the channel state information is quantized in the MIMO precoding system. The MIMO precoding system comprises one or more transmitting antennas and receiving antennas. The matrix V may be generated using Singular Value Decomposition (SVD) or Geometric Mean Decomposition (GMD). One of the unitary matrices may be used to linearly transform vector signals at a transmitter in the MIMO precoding system.

FIG. 1A illustrates cellular multipath communications between a base station and a mobile computing terminal in accordance with an embodiment of the present invention. Referring to fig. 1A, a house 120, a mobile terminal 122, a factory 124, a base station 126, a car 128, and communication paths 130, 132, and 134 are shown.

The base station 126 and the mobile terminal 122 may comprise suitable logic, circuitry, and/or code that may enable generating and processing MIMO communication signals.

Wireless communications between base station 126 and mobile terminal 122 may occur over a wireless channel. The wireless channel includes a plurality of communication paths, e.g., communication paths 130, 132, and 134. The wireless channel may change dynamically as the mobile terminal 122 and/or the automobile 128 move. In some cases, the mobile terminal 122 is line-of-sight (LOS) with the base station 126. In other cases, where there may not be a direct line of sight between the mobile terminal 122 and the base station 126, the radio frequency signals travel between reflected communication paths between the communicating entities, as shown by exemplary communication paths 130, 132, and 134. The radio frequency signal may be reflected by man-made structures such as a house 120, a factory 124, or a car 128, or naturally occurring obstacles such as mountains. Such systems are referred to as non line-of-sight (NLOS) communication systems.

A communication system may include both LOS and NLOS signal components. If there is an LOS signal component, the LOS signal component is much stronger than the NLOS signal component. In some communication systems, NLOS signal components can create interference and degrade receiver performance. This refers to multipath interference. For example, communication paths 130, 132, and 134 may have different delays to reach mobile terminal 122. The communication paths 130, 132, and 134 may also have different attenuations. For example, in the downlink, the signal received at mobile terminal 122 may be a summation of signals of the different attenuated communication paths 130, 132, and/or 134, which may be unsynchronized and dynamically varying. Such channels are known as fading multipath channels. Fading multipath channels may introduce interference, but may also introduce diversity and degrees of freedom for the wireless channel. Communication systems having multiple antennas at the base station and/or mobile terminal, such as MIMO systems, are particularly well suited to take advantage of the characteristics of wireless channels and can achieve significant performance gains from fading multipath channels, which can result in significant performance improvements over communication systems having a single antenna at the base station 126 and mobile terminal 122, particularly NLOS communication systems.

Fig. 1B shows a schematic diagram of a MIMO communication system according to an embodiment of the invention. Referring to fig. 1B, a MIMO transmitter 102 and a MIMO receiver 104 are shown, along with antennas 106, 108, 110, 112, 114, 116. Also shown is a wireless channel comprising a communication path h₁₁、h₁₂、h₂₂、h₂₁、h_2NTX、h_1NTX、h_NRX1、h_NRX2、h_NRXNTXWherein h is_mnMay represent the channel coefficients from the transmitter antenna n to the receiver antenna m. May have N_TXA transmitter antenna and N_RXA receiver antenna. The figure also shows the transmitted symbol x₁、x₂、x_NTXAnd receiving symbols y1, y2, y_NRX。

The MIMO transmitter 102 may comprise suitable logic, circuitry and/or code that may enable generation of the transmit symbol x_i i∈{1，2，...N_TXWhich are transmitted via transmit antennas 106, 108, 110 as shown in fig. 1B. The MIMO receiver 104 may comprise suitable logic, circuitry and/or code that may be enabled to process received symbols y received via the receive antennas 112, 114, 116 shown in FIG. 1B_i i∈{1，2，...N_RX}. The input-output relationship between transmitted and received signals in a MIMO system can be expressed as follows:

y＝Hx+n

wherein y ═ y₁，y₂，…y_NRX]^TIs provided with N_RXThe column vector of the individual elements.^TDenotes the transposition of the vector, H ═ H_ij]:i∈{1，2，...N_RX}；j∈{1，2，...N_TXDenotes N_RX×N_TXDimensional channel matrix, x ═ x₁，x₂，...x_NTX]^TIs provided with N_TXA column vector of elements, N being of N_RXThe noise of a single element samples the column vector. The channel matrix H can be written as H ═ U Σ V using Singular Value Decomposition (SVD)^HWherein.^HDenotes Hamiltonian transposition (hermitian transpose), U is N_RX×N_TXIs N, sigma is_TX×N_TXAnd V is N_TX×N_TXIs used to generate the unitary matrix. Other matrix decompositions than the SVD decomposition may also be used to diagonalize or transform the matrix H. Other matrix decompositions that can convert the matrix H to a lower/upper triangular matrix are also possible if the receiver algorithm performed within the MIMO receiver 104 is, for example, sequenced interference cancellation (OSIC). Such decomposition includes Geometric Mean Decomposition (GMD), where H QRP^HWhere R may be an upper triangular matrix with the geometric mean of the singular values of H on diagonal elements, and Q and P may be unitary matrices.

Fig. 2 shows a block diagram of a MIMO precoding transceiver chain model in accordance with an embodiment of the invention. Referring to fig. 2, a MIMO precoding system 200 is shown that includes a MIMO transmitter 202, a MIMO equivalent baseband channel 203, a MIMO receiver 204, and a summing module 208. The MIMO transmitter 202 includes a Transmitter (TX) baseband processing module 210 and a transmit precoding module 214. The MIMO equivalent baseband channel 203 includes a wireless channel 206, a TX Radio Frequency (RF) processing module 212, and a Receiver (RX) RF processing module 218. The MIMO receiver 204 includes a precoding decoding module 216 and an RX baseband processing module 220. Also shown are symbol vector s, precoding vector x, noise vector n, received vector y and channel decoding vector y'.

The MIMO transmitter 202 may comprise a baseband processing module 210 comprising suitable logic, circuitry, and/or code that may enable generating MIMO baseband transmit signals. The MIMO baseband transmit signals may be transmitted to transmit precoding module 214. The baseband signals are suitably encoded within transmit pre-coding module 214 for transmission in wireless channel 206, and transmit pre-coding module 214 may comprise suitable logic, circuitry, and/or code that may enable it to perform these functions. The TX RF processing module 212 may comprise suitable logic, circuitry and/or code that may enable signals transmitted to the TX RF processing module 212 to be modulated onto a Radio Frequency (RF) for transmission over the wireless channel 206. The RX RF processing module 218 may comprise suitable logic, circuitry and/or code that may be enabled to perform radio frequency front end functions to receive signals transmitted over the wireless channel 206. The RX RF processing module 218 may comprise suitable logic, circuitry and/or code that may be operable to demodulate its input signals into baseband signals. The summing block 208 represents adding noise to the received signal in a MIMO receiver. The MIMO receiver 204 includes a precoding decoding module 216 that linearly decodes the received signals and transmits the signals to an RX baseband processing module 220. The RX baseband processing module 220 may comprise suitable logic, circuitry and/or code that may be enabled to apply further signal processing to the baseband signal.

The MIMO transmitter 202 may comprise a baseband processing module 210 comprising suitable logic, circuitry, and/or code that may enable generating MIMO baseband transmit signals. The MIMO baseband transmit signal, i.e., symbol vector s, is transmitted to transmit precoding module 214. The symbol vector s may be N_TXVector of x 1 dimension.

Transmit precoding module 214 performs a linear transformation on symbol vector s such that x is Ws, where W has dimension N_TXMultiplied by the length of s, and x ═ x₁，x₂，…x_NTX]^T. Each element of the precoding vector x is available N_TXTransmitted on a different antenna.

The transmitted precoding vector x passes through the MIMO equivalent baseband channel 203. From N_RXThe signal y received by each receiver antenna is the transmitted signal x transformed by the MIMO equivalent baseband channel 203 (represented by matrix H) and added with a noise component (represented as noise vector n). As shown in the summing block 208, the received vector y is expressed as y — HxAnd + n is HWs + n. The received vector y is transmitted to a precoding decoding module 216, where a linear decoding operation B may be performed on the received vector y to obtain a decoded vector y' ═ B^Hy＝B^HHWs+B^Hn, where B is a complex matrix of appropriate dimensions. The decoded vector y' is then passed to the RX baseband processing module 220 where further signal processing is performed on the output of the precoding decoding module 216.

If the MIMO equivalent baseband channel 203 transfer function H applied to transmit the precoding vector x is known at both the MIMO transmitter 202 and MIMO receiver 204, the channel can be diagonalized, e.g., by setting W ═ V and B ═ U, where H ∑ U ∑ V^HIs a singular value decomposition. In these cases, the channel decoding vector y' is shown as follows:

y’＝U^HU∑V^HVs+U^Hn＝∑s+U^Hn

since Σ is a diagonal matrix, no interference exists between elements of the symbol vector s in y', and thus the wireless communication system is similar to the one having up to N_TXLike a parallel single-antenna wireless communication system, the rank of the channel matrix H may be equal to or less than N for each element of s_TX。

Fig. 3 shows a block diagram of a MIMO precoding system with finite rate channel state information feedback in accordance with an embodiment of the present invention. Referring to fig. 3, a MIMO precoding system 300 is shown that includes a partial MIMO transmitter 302, a partial MIMO receiver 304, a wireless channel 306, a summing module 308, and a feedback channel 320. The partial MIMO transmitter 302 includes a transmit precoding module 314. The partial MIMO receiver 304 includes a precoding decoding module 316, a channel estimation module 322, a channel quantization module 310, a channel decomposition module 312, and a codebook processing module 318. Fig. 3 also shows a symbol vector s, a precoding vector x, a noise vector n, a received vector y, and a decoded vector y'.

The transmit precoding module 314, the wireless channel 306, the summing module 308, and the precoding decoding module 316 are substantially the same as the transmit precoding module 214, the MIMO equivalent baseband channel 203, the summing module 208, and the precoding decoding module 216 shown in fig. 2. The channel estimation module 322 may comprise suitable logic, circuitry and/or code that may enable estimating a transfer function of the wireless channel 206. The channel estimate may be communicated to a channel decomposition block 312, and the channel decomposition block 312 may comprise suitable logic, circuitry, and/or code that may be enabled to decompose the channel. In this regard, the decomposed channels are communicatively coupled to a channel quantization module 310. The channel quantization module 310 may comprise suitable logic, circuitry and/or code that may enable quantization of the channel portion onto a codebook. The codebook processing module 318 may comprise suitable logic, circuitry and/or code that may enable generation of a codebook. Feedback channel 320 is represented as a channel that transmits channel state information from partial MIMO receiver 304 to partial MIMO transmitter 302.

In many wireless systems, the channel state information, i.e., knowledge of the channel transmission matrix H, may not be known to the transmitter and receiver. However, in order to utilize a precoding system as shown in fig. 2, it may be desirable to know at least part of the channel information at the transmitter end. In the embodiment of the invention as disclosed in fig. 2, the MIMO transmitter 302 needs to have a unitary matrix V for precoding within the transmit precoding module 214 of the MIMO transmitter 202.

In a Frequency Division Duplex (FDD) system, a communication band for downlink communication from a base station to a mobile terminal is different from a communication band for uplink communication. Channel measurements on the uplink are generally not useful for the downlink due to the difference in frequency bands, and vice versa. In these cases, the measurements are only performed at the receiver end, and Channel State Information (CSI) may be transmitted back to the transmitter by way of feedback. For this reason, CSI may be fed back from the partial MIMO receiver 304 to the transmit precoding module 314 of the partial MIMO transmitter 302 through the feedback channel 320. The transmit precoding module 314, the wireless channel 306, and the summing module 308 are substantially the same as the corresponding modules 214, 203, and 208 shown in fig. 2.

At the partial MIMO receiver 304, the received signal y is modulo the channel estimateBlock 222 may be used to estimate the channel transfer function H, usingAnd (4) showing. The estimate may be further decomposed into a diagonal matrix or triangular matrix form, depending on the particular receiver implementation, as explained in fig. 2, for example. For example, the channel decomposition module 312 may perform SVD decomposition:in the case of multiple antennas, the dimensions of the matrices U, Σ, and V increase very quickly. In these cases, it may be desirable to use a matrixQuantized dimension N_TX×N_RXMatrix V of_qIn which V is_qThe set C ═ V can be derived from a previously defined finite unitary matrix_iAnd (6) selecting. The unitary matrix set C is called a codebook. By finding a proximity matrix from the codebookMatrix V of_qIn a sense, if the codebook C is known by the partial MIMO transmitter 302, it is sufficient to be able to transmit the index q from the channel quantization module 310 to the transmit precoding module 314 via the feedback channel 320. The codebook C changes much more slowly than the channel transfer function H, so that the codebook in the transmit precoding module 314 can be periodically updated from the codebook processing module 318 through the feedback channel 320. The codebook C selected may be static or adaptive. Moreover, the codebook may also be adaptively or non-adaptively selected from a set of codebooks, including adaptively and/or statically designed codebooks. In these cases, the partial MIMO receiver 304 may be in any of these casesThe point in time informs the partial MIMO transmitter 302 of the codebook used. Matrix arrayCan be quantified as V by the relationship described below_q：

Wherein A ═ a_ij]And the dimension of A is N. Thus, the matrix V can be selected_qAs matrices in a codebookWhich can maximize the function as defined above. The function f () may average the squares of the absolute values of the diagonal elements of its input matrix. By maximizing f (), in a sense, the matrix V_qIs selected to satisfy the productMaximally similar to the identity matrix. The above expression of f () may maximize the instantaneous capacity of the precoded MIMO system under some approximation. Thus, channel H may be estimated in channel estimation module 322 and decomposed in channel decomposition module 312.

Within the channel quantisation module 310, matrices, e.g.Can be quantized into a matrix V_qThe index q is fed back to the partial MIMO transmitter 302 via a feedback channel 320. The codebook C may also be chosen to be invariant over time. Moreover, the codebook C may also be adaptively or non-adaptively selected from a set of codebooks, including adaptively and/or statically designed codebooks. The codebook C from the codebook processing module 318, which varies more slowly than the index q, may be sent to the partial MIMO transmitter 302 via a feedback channel 320. When the cardinality | C | of the codebook C is less than or equal to | C | < 2^MM bits are sufficient for feeding back the index q.

The transmit precoding module 314 may perform, for example, a linear transformation of x ═ V_qAnd s. The receiver-side precoding decoding module 316 may perform linear transformationIn some cases, the rank r of the channel matrix H is less than the number of transmit antennasI.e. satisfy r.ltoreq.N_TX. In these cases, it may be desirable to map a reduced amount of spatial (spatial) flow to vector x, as depicted in FIG. 2. For example, the vector s may be chosen to satisfy x ═ Ws, where the dimension of W is N_TXThe length of x s, which is the number of spatial streams, is generally smaller than the rank r. For example, the matrix W may be formed from V_qIs constructed from the desired column(s). In another embodiment of the present invention, vector x may be represented by x ═ V, as described above_qs is generated and has a length of N_TXSome suitably chosen elements of the vector s of (a) are set to 0, such that in general the non-0 elements of the vector s are smaller than the rank r. In these cases, the element set to 0 in s corresponds to the spatial stream that is not used.

Fig. 4 shows a flowchart of an implementation of a codebook algorithm according to an embodiment of the present invention. Referring to FIG. 4, there is shown a start step 402, an end step 418, process steps 404, 406, 410, 412, 414, 416, and decision step 408.

The selection of codebook C is important for determining the performance of the feedback system as shown in fig. 3. The flow chart in fig. 4 gives an example method of performing a codebook algorithm suitable for any quantization precision, i.e. codebook C ═ { V ═_iIn V matrix_iAnd any number of receive and transmit antennas. However, the codebook can be designed to maximize the following relationship:

wherein E_HIs the period of the set of channel realizations HThe expected value. The above relationships indicate that the best codebook selection, in some cases, may be the inclusion of matrix V_qThat can maximize the expectation of the function f (.) with respect to the channel H. This can be achieved by the algorithm of fig. 4.

In step 404, a codebook is initialized. For example, codebook C ═ { V ═ V_iCan be composed of a random matrix V_iAnd (5) constructing. The matrix V_iThe initial selection of (a) may be arbitrary. In step 406, the variable W is initialized to a value of "not converged". Although a is not set to the value "converge" in decision step 408, the algorithm may calculate Voronoi regions from codebook C in step 410. In step 410, the multi-dimensional space contained in the codebook is decomposed into a plurality of regions (cells) R_iI.e. Voronoi zones, such that each zone R_iRepresenting a subspace which may in a sense be associated with a particular matrix V_iMost similar, as shown by the following relationship:

in step 412, codebook C may be based on Vor calculated in step 410The onoi area is updated. For any given region R_iA new matrix V can be generated_i＝V_i，1，V_i，2，…V_i，NTXIn which V is_i，kIs of length N_TXIs shown in the following relation:

wherein V_kIs the kth column of the matrix V. The solution to the above optimal problem is shown by the following relation:

v_i，k＝u_MAX

λ_MAX·u_MAX＝R_kku_MAX

wherein V_i，kIs selected to correspond to matrix R_KKMaximum eigenvalue λ of_MAXCharacteristic vector U of_MAX. For the above equation, the expected value E for the variable Z_z{. may be calculated by taking the average of the samples in space.

In step 414, a test for convergence may be applied to codebook C to verify whether the codebook converges to a stable set. For example, a convergence test may pass through each matrix V_qHow much the degree of change between successive steps 412 is measured. If the change is less than the threshold, convergence may be considered to have been achieved. In step 414, variable a will be set to the value "converge" if the codebook has converged, otherwise variable a is set to the value "not converge".

If, in step 414, the variable A has been set to "converge," the loop including steps 410, 512, and 414 can be skipped in step 416, and the resulting codebook C is orthogonal. Matrix { V_iIt is also desirable to be orthogonal, that is:

where I denotes an identity matrix and 0 denotes a matrix with elements all 0. To obtainCode book of (C ═ V)_iCan be obtained by pairing matrix V_iApply e.g. Gram-Schmidt (gray-Schmidt) orthogonalization but orthogonal. Orthogonalization of the codebook terminates the generation of the codebook in an end step 418.

Figure 5 is a graph of the performance of a 2 x 2MIMO system with an MMSE receiver and finite rate feedback in accordance with an embodiment of the present invention. Referring to fig. 5, a signal-to-noise ratio (SNR) axis and a spectral efficiency axis (Seff) are shown. Also shown are plot 502 of the full feedback singular value decomposition SVD (Fit-FB), plot 504 of the 2-bit feedback SVD (SVD-2B), plot 506 of SVD-3B, plot 508 of SVD-4B, plot 510 of SVD-5B, and plot 512 of SVD-6B.

As explained in fig. 3, the number of bits available for feedback determines the unitary matrixThe quantization accuracy of (2). For M bits, codebook C includes | C | ≦ 2^MElement { V_i}. Thus, a large number of feedback bits may allow for better channel quantization and may provide better performance. As shown in fig. 5, an embodiment in accordance with the invention can provide performance close to full feedback SVD (Fit-FB)502 despite using few feedback bits. Graphs SVD-2B 504, SVD-3B 506, SVD-4B 508, SVD-5B 510, and SVD-6B 512 are also close to SVD (Fit-FB) 502. A Minimum Mean Square Error (MMSE) receiver may be used in fig. 5.

Fig. 6 is a graph of the performance of a 2 x 2MIMO system with a GMD-OSIC receiver and finite rate feedback in accordance with an embodiment of the invention. Referring to fig. 6, a signal-to-noise ratio (SNR) axis and a spectral efficiency axis (Seff) are shown. Also shown are a graph of a full feedback geometric mean decomposition (GMD Fit-FB)602, a graph 604 of a 2-bit feedback GMD (GMD-2B), a graph 606 of GMD-3B, a graph 608 of GMD-4B, a graph 610 of GMD-5B, and a graph 612 of GMD-6B.

The performance graph shown in fig. 6 utilizes a sequenced successive interference cancellation (OSIC) receiver. GMD was used instead of SVD as described in connection with fig. 1B. As can be seen in FIG. 6, the performance of one embodiment of the present invention using several feedback bits, e.g., GMD-2B 604, GMD-3B 606, GMD-4B 608, GMD-5B 610, and GMD-6B 612, is very close to the performance of a system with full feedback (graph of GMD Fit-FB 602).

According to one aspect of the invention, a method and system for codebook design for MIMO precoders with finite-rate channel state information feedback includes: channel state information in MIMO precoding systems 200 and 300 is quantized using a cost function onto a codebook comprising one or more unitary matrices in a channel quantization module 310 and the codebook is repeatedly updated based on at least the channel state information in a codebook processing module 318. The channel state information obtained in the channel estimation module 322 includes the matrix V generated in the channel decomposition module 312. The cost function f (a) is defined by the following relation:

where A is an NxN matrix, a_ijIs the element (i, j) of matrix a. Voronoi regions may be generated from the codebook for the unitary matrix in codebook processing module 318. Also, in codebook processing module 318, a set of matrices is generated based on the Voronoi regions and the unitary matrix. Updating the codebook may be achieved by modifying the set of matrices to a new set of unitary matrices, wherein the new set of matricesThe unitary matrix of (a) is integrated into a codebook. In the channel quantization module 310, the index of the element in the codebook to which the channel state information is quantized is transmitted from the receiver to the transmitter in the MIMO precoding system through the feedback channel 320. MIMO precoding system 200/300 includes one or more transmit antennas and receive antennas. In the channel decomposition module 312, the matrix V is generated using Singular Value Decomposition (SVD) or Geometric Mean Decomposition (GMD). In the transmit precoding module 314, one of the unitary matrices is used to linearly transform the vector signal at the transmitter end in the MIMO precoding system.

Another embodiment of the present invention provides a machine-readable storage, having stored thereon, a computer program having at least one code section executable by a machine, thereby causing the machine to perform the steps of a method for codebook design for MIMO precoders with finite rate channel state information feedback as described herein.

Accordingly, the present invention may be realized in hardware, software, or a combination of hardware and software. The present invention can be realized in a centralized fashion in at least one computer system, or in a distributed fashion where different elements are spread across several interconnected computer systems. Any kind of computer system or other apparatus adapted for carrying out the methods described herein is suited. A typical combination of hardware and software could be a general purpose computer system with a computer program that, when being loaded and executed, controls the computer system such that it carries out the methods described herein. The method is implemented in a computer system using a processor and a memory unit.

The present invention can also be implemented by a computer program product, which comprises all the features enabling the implementation of the methods of the invention and which, when loaded in a computer system, is able to carry out these methods. The computer program in this document refers to: any expression, in any programming language, code or notation, of a set of instructions intended to cause a system having an information processing capability to perform a particular function either directly or after either or both of the following: a) conversion to other languages, codes or symbols; b) reproduced in a different format.

While the invention has been described with reference to several embodiments, it will be understood by those skilled in the art that various changes may be made and equivalents may be substituted without departing from the scope of the invention. In addition, many modifications may be made to adapt a particular situation or material to the teachings of the invention without departing from its scope. Therefore, it is intended that the invention not be limited to the particular embodiment disclosed, but that the invention will include all embodiments falling within the scope of the appended claims.

Description of the invention

This application incorporates by reference and claims priority from U.S. provisional patent application No. 60/884118 filed on.1/9/2007 in its entirety.

This application also refers to:

U.S. patent application No. 60/884118, filed on date 2007, month 1 and day 9;

U.S. patent application No. 60/884133, filed on date 2007, month 1 and day 9;

U.S. patent application No. 60/884132, filed on date 2007, month 1 and day 9;

U.S. patent application No. _________, (attorney docket No. 18183US02), filed on day 22/6/2007;

U.S. patent application No. _________, (attorney docket No. 18184US02), filed on day 22/6/2007;

U.S. patent application No. _________, (attorney docket No. 18185US02), filed on day 22/6/2007;

each of the applications identified above is hereby incorporated by reference in its entirety.

Claims (8)

1. A method for processing a communication signal, the method comprising:

quantizing channel state information in an MIMO precoding system onto a codebook by using a cost function, wherein the channel state information is a unitary matrix V used for precoding in a transmitting precoding module of an MIMO transmitter in the MIMO precoding system, and the codebook is a unitary matrix set C and comprises one or more unitary matrices;represents an estimate of the channel transfer function H, and

matrix arrayQuantified as V according to the relationship described below_q:

Wherein, V_qIs a unitary matrix from codebook C, f (A) is a relation of the cost function, A is an N × N matrix, a_ijIs the element (i, j) of matrix a; and

iteratively updating the codebook based on at least the channel state information, comprising:

generating from said codebook, for said one or more unitary matrices, voronoi regions, i.e. the decomposition of the multidimensional space comprised by the codebook into a plurality of regions R_i(ii) a The R is_iAs shown by the following relationship:

generating a set of matrices based on said voronoi region and said unitary matrix, i.e. for a given said region R_iGenerating a new matrixThe V is_i.kIs a column vector of length NTX; the NTX is the number of the transmitter antennas;

updating the codebook by modifying the set of matrices to a new set of unitary matrices, wherein the new set of unitary matrices is integrated into the codebook.

2. The method for processing a communication signal as recited in claim 1, further comprising transmitting an index of an element in the codebook from a receiver to a transmitter in the MIMO precoding system, wherein the channel state information is quantized onto the codebook.

3. The method for processing a communication signal of claim 1, further comprising generating the matrix V using singular value decomposition.

4. The method for processing a communication signal of claim 1, further comprising generating the matrix V using geometric mean decomposition.

5. The method for processing communication signals according to claim 1, further comprising linearly transforming vector signals at a transmitter in the MIMO precoding system using one of the unitary matrices.

6. A system for processing a communication signal, the system comprising a MIMO precoding system comprising a channel quantization module and a codebook processing module;

quantizing, in the channel quantization module, channel state information onto a codebook comprising one or more unitary matrices using a cost function; the channel state information is a unitary matrix V used for precoding in a transmitting precoding module of a MIMO transmitter in the MIMO precoding system, and the codebook is a unitary matrix set C and comprises one or more unitary matrices;represents an estimate of the channel transfer function H, and

matrix arrayQuantified as V according to the relationship described below_q:

Wherein, V_qIs a unitary matrix from codebook C, f (A) is a relation of the cost function, A is an N × N matrix, a_ijIs the element (i, j) of matrix a; and

in the codebook processing module, iteratively updating the codebook based on at least the channel state information, including:

generating from said codebook, for said one or more unitary matrices, voronoi regions, i.e. the decomposition of the multidimensional space comprised by the codebook into a plurality of regions R_i(ii) a The R is_iAs shown by the following relationship:

generating a set of matrices based on said voronoi region and said unitary matrix, i.e. for a given said region R_iGenerating a new matrixThe V is_i.kIs a column vector of length NTX; the NTX is the number of the transmitter antennas;

updating the codebook by modifying the set of matrices to a new set of unitary matrices, wherein the new set of unitary matrices is integrated into the codebook.

7. The system for processing a communication signal as recited in claim 6, wherein in the channel quantization module, indices of elements in the codebook to which the channel state information is quantized are transmitted from a receiver to a transmitter in the MIMO precoding system through a feedback channel.

8. The system for processing communication signals according to claim 6, wherein the MIMO precoding system further comprises a channel decomposition module;

in the channel decomposition module, the matrix V is generated using singular value decomposition or geometric mean decomposition.