Multiple input, multiple output wireless communication system, associated methods

ES 2 553 344 T3

DESCRIPTION

Wireless communication system of the multi-input-multi-output type and its associated data structures and methods

PRIORITY

This application is a continuation in part of a co-pending United States Patent Application No. TBD, filed December 5, 2005 by Lin et al. The applicant considers that the parent application fully enables the corresponding claims.

FIELD OF THE INVENTION

The embodiments of the invention are usually intended for communication systems and more particularly, for a wireless communication system of the multi-input, multi-output (MIMO) type as well as its associated data structures and methods.

BACKGROUND OF THE INVENTION

The use of multiple signaling paths, eg, through the use of multi-input, multi-output (MIMO) technology can greatly increase the effective headroom of a wireless communication channel. Despite the notable benefits gained since the introduction of MIMO technology in a communications device, implementation problems persist that have limited the wide adoption of such technologies.

As a result, MIMO technology implementations that allow for the wide adoption of such technologies have not yet been implemented. A variety of exemplary implementations of MIMO technology, as well as their associated data structures and methods, are disclosed in the following inventive idea that eliminates one or more of these conventional limitations.

The document by Mondal et al, entitled "Adaptive Feedback for MIMO Technology Beamforming Systems", Signal Processing Devices in Wireless Communications, 2004, IEEE 5th Informative Workshop in Lisbon, Portugal, dated July 11-14, Piscataway, NJ, United States, IEEE, July 11, 2004, pages 213-217, XP010806835. ISBN 0-7803-8337-0.

BRIEF DESCRIPTION OF THE DRAWINGS

The embodiments of the present invention are illustrated by way of example and not by way of limitation, in the Figures and in the attached drawings where similar numerical references refer to similar elements and where:

Figure 1 is an exemplary communications block diagram within which embodiments of the invention may be practiced;

Figure 2 provides a graphical representation of an exemplary beamforming process in accordance with one embodiment;

Figure 3 is a flow chart of an exemplary method for beamforming using a non-uniform codebook, in accordance with one embodiment;

Figure 4 is a graphic illustration of a uniform codebook;

Figure 5 is a graphical illustration of a non-uniform codebook, in accordance with one embodiment;

Figure 6 is a flow chart of an exemplary method for beamforming using a hierarchical codebook, in accordance with one embodiment;

Figure 7 is a graphical illustration of a hierarchical codebook, in accordance with one embodiment;

Figure 8 is a block diagram of an exemplary communication device within which embodiments of the invention may be practiced in accordance with one embodiment; Y

Figure 9 is a block diagram of an exemplary article of manufacture that includes content that, when executed by an access machine, causes the machine to implement one or more aspects of the embodiments. of the invention.

DETAILED DESCRIPTION OF THE INVENTION

ES 2 553 344 T3

Wireless communication system embodiments, methods, and associated data structures are often presented to implement one or more exemplary MIMO technology embodiments.

In co-pending United States patent application No. 11 / 036,906, entitled Codebook Generation System and Associated Methods, filed on January 13, 2005 by Lin et al, (hereinafter referred to as the Lin patent), a codebook generation agent (CGA) was filed with one or more dynamically generated matrix codewords from, eg, matrix codewords that are dynamically generated from, eg, Vector codebooks for vectors of 2, 3, 4, ..., N units. Those skilled in the art will appreciate that such N-dimensional unit vectors are already arranged in such communication devices to support various other features such as, eg, single data stream beamforming. Consequently, Lin's co-pending application represents a significant improvement in the implementations of MIMO technology.

With this application, the Lin reference is extended to introduce an innovative code management agent (CMA). In accordance with an embodiment, more fully described below, a codebook management agent (CMA), which can be selectively requested to dynamically generate and / or use one or more uniform codebooks, non-uniform and / or hierarchical in support of eg, an implementation of closed-loop MIMO technology (or, beamforming). The CMA agent disclosed herein may also include and / or be coupled to one or more of the CGA agents, although the scope of the invention is not thus limited.

Beamforming is a useful technique to increase the effective signal-to-noise ratio (SNR) perceived by receivers within MIMO technology systems. In an explicit operational feedback system, information associated with a beamforming matrix is ​​forwarded to the transmitter. To efficiently encode the beamforming matrix, a series of vector codebooks can be designed and / or dynamically generated, eg, as disclosed in the aforementioned co-pending patent application. In non-correlated channels, the beamforming matrix is ​​uniformly distributed in the so-called Steifel manifold (see, eg, Figure 4). In this regard, in traditional implementations a vector codebook is largely optimized to cover the full range with the following criteria:

C (m_tAQ = argmax (min (rf_c(c_F, c_Y), f <J = í, ..., Af), c₍ e C „witá || c¡ || = 1) (1)

Cl where m is the dimension of the vector codebook and N is the number of codewords in the codebook. The complex m-tuple space is indicated by C_m. The chordal distance is defined as:

= jHkAy II²(2) where a normal inner product is used.

When the beamforming vector V is quantized_m, the codeword with the smallest chordal distance is selected:

^ argmintí / ^ c ^ v ^ Xc ,. e (3) n

This kind of "uniform" codebook can also be designed and implemented using one or more of the techniques disclosed in the aforementioned co-pending patent application. The codebooks described there represent an order of magnitude improvement of a factor of 2-4 in quantization efficiency with the same or better performance compared to other systems.

In situations where consecutive data packets are sent, the time between packets is short. In static or slow fading channels, the characteristics of the channels do not usually change practically during this period of time. Consequently, only a small beam direction correction is required to correct for back drift. In the embodiments described below, this property of static or slow fading channels is reinforced in the design and use of new codebooks with an effective reduction in quantization complexity (and, like such computational complexity) by the same time as the beamforming accuracy is improved.

Reference, throughout this description, to "a single embodiment" or "one embodiment" means that a particular property, structure, or characteristic described in connection with the embodiment is included in at least one embodiment of the present invention. Thus, the occurrences of the terms "in one embodiment" or "in one embodiment" in various places throughout this description do not necessarily all refer to the same embodiment. Furthermore, the structural properties, or particular characteristics, can be combined in any suitable way in one or more embodiments.

ES 2 553 344 T3

Technical details regarding some of the operational characteristics of mobile devices and / or wireless communication networks where CMA can be implemented can be found in, eg, IEEE 802.11, 1999 edition; Information technology telecommunications and information exchange between systems - Local and metropolitan area networks - Specific requirement, part 11: WLAN network Media Access Control (MAC) and physical layer specifications (PHY), their progeny and their add-ons (eg, 802.11a, 802.11g and 802.11η). See also IEEE Std 802.16-2001, IEEE Std. 802.16-2001 for local and metropolitan area networks, part 16: An air interface for fixed broadband wireless access systems, their progeny and complements (eg, 802.16a, 802.16d and 802.16e).

Communications environment as an example

In Figure 1, a block diagram of an exemplary wireless communication environment 100 is illustrated within which environment embodiments of the invention may be suitably practiced. In accordance with the exemplary illustrated embodiment of Figure 1, an exemplary communication environment 100 is illustrated including the wireless communication device 102 in communication with another wireless communication device 106 via a wireless communications link 104. As used herein, the communications environment 100 is intended to represent any of a wide range of wireless communications networks including, without limitation, a near field communications network (NFC), a wireless local area network (WLAN). , a wireless metropolitan area network (WMAN), a wireless wide area network (WWAN), a cellular radio network, a personal communication system (PCS) network, and similar networks.

In accordance with one embodiment, for purposes of illustration and not limitation, communication network 100 will be described in the context of an 802.11x standard (where x is a, b, g, n and / or combinations and / or their progeny), although the scope of the invention is not limited in this regard. In accordance with this embodiment, by way of illustrative example, device 102 may be an access point (AP) while device 106 may be a remote station (STA), although the scope of the invention is not limited in this regard. .

In a closed-loop MIMO system, a data signal is weighted by a V beamforming matrix and then selectively transmitted over a plurality of antennas, as illustrated. According to one embodiment, the data signal may include several data streams (N, ... Ne), although the invention is not limited in this regard. The number of data streams may represent the number of spatial channels, with appropriate bit load assignments, power weighting and subcarriers, although the invention is not limited in this regard.

In accordance with an embodiment with four (4) transmitting antennas and three (3) data streams (for ease of illustration), the signal (x) transmitted through the Nt (4) of 4 antennas can be represented as x = Vx j (4) where v =

<img file="ES2553344T3_D0001.tif" />

As illustrated, s is a vector N_s of data symbols and V is the Nt by N type beamforming matrix_sdeveloped from information (eg, matrix codebooks and / or their indices fed back from a distant receiver (eg, 102). According to one embodiment, the beamforming matrix V is usually unitary and the bit loading / power can be applied on the vector s, as presented above, although the invention is not limited in this regard.

Device 106 is illustrated including a codebook management agent (CMA) 108 to dynamically generate and / or manage the use of one or more matrix codebooks whose channel status information can be characterized and fed back to a remote device, eg, 102. In accordance with one embodiment, more fully described in the previously filed co-pending patent application, and by not memorizing one or more matrix codebooks, CMA 108 can compile the matrix codebooks or a part thereof, necessary to characterize channel state information from dynamically generated matrix codewords from one or more vector codebooks for vectors of 2, 3, 4, ..., N units.

Implementing beamforming as an example

Referring back to Figure 2, a graphical representation of an exemplary beamforming process is illustrated in accordance with one embodiment. With reference to Figure 2, the process, by way of

ES 2 553 344 T3 example, starts with the access point (AP) (eg, 106) that broadcasts a request to send message (CTS) (or, polling packet) to a distant station over a communications channel (eg, 104).

The remote station (STA) (eg, 102) measures the channel and requests an instance of the channel management agent (CMA) 108 to calculate a beamforming matrix, eg, with singular value decomposition (SVD) as:

'(5) where i / y is a matrix Nt by N_t; Ui is a matrix N_r by N_rand Σι is a matrix N_r by Nt.

The agent STA then forwards the quantized beamforming matrix E which is N, by Λ / y, in a release-to-send (CTS) packet or other feedback (or response) packet.

The access point AP multiplies the message s, which is Ni by 1 with N_s nonzero fluxes, by the beamforming matrix E as E^si. Access point AP sends data packet 1 with beamforming, from where agent STA measures the new equivalent channel

The STA agent calculates the new beamforming matrix as:

H₂V_} &lt;= = U₂Z₂V «(6)

If the channel change between Hi and H2 is small and the quantization error in E is small, it is closer to the identity identity. The agent STA quantifies E and feeds back E. Since E is Ni by Λ / y, the access point AP can directly multiply the message vector Λ / y by 1 with E as E ^ 'j and send the message to the agent STA. The STA agent calculates the new / 3 beamforming matrix as:

h /₂ = E / jSj yV₂ = σ₃£₃Ι ~ / (7)

Since agent STA can decide to receive signal from the strongest N beamforming channels_s, only feed back N_s columns of the calculated Nt by Nt type beamforming matrix, corresponding to the strongest N channels_s. For this case, the probing is not direct. Access point AP needs to expand the Nt-type beamforming matrix by N_s fed back to the Nt by Nt matrix, so that all Nt channels can be probed. Expansion adds Nt - N_s columns so that the expanded matrix is ​​a unit matrix like:

É = -> ^ = (^ -. ^^, ... ^] ^ (8)

The expansion can be calculated by the so-called Householder reflection as follows, or by other methods.

<img file="ES2553344T3_D0002.tif" />

<img file="ES2553344T3_D0003.tif" />

(9) where - ^r~^2w^ is the Householder reflection matrix of the feedback Householder vector ^v; in the V,. -e, Lin's co-pending patent application; ll^v ~ M and - ·, ο]^Γ; v, is calculated from the feedback index q_s in co-pending Lin patent application. The unit difference between equation (7) and the reconstruction stage in Lin's application is that in (8) it is replaced by^ν·^ν· In Lin's patent application. For adaptive loading

In bit or power ES 2 553 344 T3, the beamforming vectors can be classified based on their corresponding singular values ​​prior to quantization and index feedback.

Λ

It should be appreciated, however, that any unit matrix G that transforms L^{1 0} ·· ° Γ in ^v¡(Within a global phase) can also be substituted for the Householder reflection described above. As an example, the Householder reflection can also be substituted with a series of Given rotations, eg,

V φ_Νι ~ φ_Ν^) G (7V ,. -1, N., 0_N¡_,) - ψ (2,) G (1,2, Θ.)

<img file="ES2553344T3_D0004.tif" />

(10) eos # -sin #

G (/ 72₉ «, #) = Sin Θ eos #

A- „J _{is a} Given rotation between row m and n with an angle Θ;

where

Ψ (/ τ?, Φ); is an identity matrix with the m-th diagonal element replaced by e^{/ &lt;p}; and Ni is the complex dimension of v,. The angles θ and r are the Given angular representation of a unit norm vector. As an example, a unit norm vector 4 can be represented by cos0, and^m without eos ^ e⁷^²sin 0t sin θ2 eos 03e⁷^ 'sin θχ sin θ₂ sin θ, β^ίφ*

Similarly, the Householder reflection can be substituted for a 180 degree rotation around the vector [10 ... 0]^T + vy. In this regard, the codebook of the gross vector can be memorized in the form of angles 6_Y and φ / ί. by Given.

It should be noted that the Householder reflection matrix F, in the co-pending Lin patent application for quantization (not reconstruction) can be substituted for any unitary matrix Ay that has ^vO^vLike the first column, where ^vY ^v/ are the vector to be quantized and the vector for quantization, respectively. As an example, the matrix Ay that converts^vto ^{00 0 0}l with quantization error can be E, where

£. · = And B, you can convert [' ⁰ - to ^andA · with quantization error.

Sample codebook data structure

Having introduced a form of operational implementation above, by way of example, several improvements will now be made to the conventional uniform codebook. In accordance with a more fully developed embodiment below, the CMA agent 108 may be requested to selectively reorder the elements of a uniform codebook, eg, such as a previously generated codebook, to effectively reduce the computational complexity of the quantization process. According to one embodiment, the code words of the code book are arranged so that said code words with the largest real component of the first element appear in the code book before the code words with smaller real components. . Consequently, there is no need to search the entire codebook. The operational feedback from the best of the five least code words is sufficient for convergence. This particular ordering of the code words reduces the complexity of quantization in the tracking mode.

In accordance with another aspect of embodiments of the invention, the CMA agent 108 can selectively add other m elements to a uniform codebook, in any other way, to generate a non-uniform codebook, where m is two to thirty-two. That is, in accordance with an embodiment, developed more fully below, the CMA agent 108 selectively adds an am codeword to the codebook to generate a non-uniform codebook, eg, when detected a stationary (or, slowly changing) channel. According to one embodiment, the additional m codewords of the non-uniform codebook are closely centered around the codeword [1,0, ..., 0]^T, as illustrated in Figure 5. Alternatively, the CMA agent 108 may dynamically select a preconfigured non-uniform codebook.

In accordance with another aspect of embodiments of the invention, the CMA agent 108 may use two sets of codewords for quantization. The first set is for coarse quantization while the second set is for fine quantization around a selected center. In accordance with a form

In an embodiment ES 2 553 344 T3, the CMA agent 108 can concatenate the codebooks. The concatenation of a thick codebook and one (or multiple) thin codebooks (with decreasing radii) effectively forms a hierarchical codebook as illustrated in Figure 7.

Although the concatenated codebook cannot provide optimal computational performance for a given feedback load, it allows for low-complexity quantization and scalable feedback resolution. Furthermore, since the beamforming matrix approximates the identity matrix after several iterations of operational feedback, the station can only use the thin codebooks without thick codebooks for quantization of some vectors in slow fading. . The foregoing reduces the operational feedback load for tracking the beamforming matrix.

It will be appreciated that for the introduction of the CMA agent 108, the device 106 is intended to represent any of a wide range of electronic devices with wireless communication capabilities. In some embodiments, the CMA agent 108 is responsible for a communicatively coupled receiver to perform the functions described herein. In accordance with some embodiments, the CMA agent 108 can be made in hardware, software, firmware, and / or any of their combinations.

CMA operation as an example

Returning to Figure 3, a flow chart of an exemplary method for utilizing non-uniform codebooks is generally presented in accordance with one embodiment. The exemplary method illustrated in Figure 3 begins with block 302 where the code management agent (CMA) 108 selects one or more codebooks. In accordance with one embodiment, the codebook can be selected from a preconstituted plurality of codebooks or dynamically generated in accordance with the teachings provided by the Lin patent application.

At block 304, CMA agent 108 can selectively reorder the previously entered uniform codebook. In accordance with one embodiment, the CMA agent 108 can arrange the codebook as follows: take any codeword co, perform a unit rotation or reflection G, so that G · co = [1,0, .. . 0]^T. An equivalent codebook can be formed by the global operation of G on each codeword of C (m, N) C '(m, N) = G · C (m, N).

According to one embodiment, the CMA agent 108 can sort the code words according to the magnitude of the first element I A-0) I. An example of C (2,12) is provided in the following table (table 1):

1.0000 + O.OOOOi

OO.OOOOi

0.8507+ O.OOOOi

-0.3599 - 0.3832Í

0.8507 - O.OOOOi

0.0659 + 0.5216Í

0.8507 -O.OOOOi

0.5164 + 0.0985Í

0.8507 + O.OOOOi

0.2532 - 0.4607Í

0.8507+ O.OOOOi

-0.4757+ 0.2239Í

0.5257 - O.OOOOi

0.5824 + 0.6200Í

0.5257-O.OOOOi

-0.1067 - 0.8439Í

0.5257 + O.OOOOi

-0.8356 - 0.1593Ϊ

0.5257-O.OOOOi

-0.4097+ 0.7455Í

0.5257 - O.OOOOi

0.7697 - 0.3623Í

0.0000 + O.OOOOi

-0.9967 -0.0818Í

A presentation of these code words is illustrated in Figure 4. In accordance with the graphical representation of Figure 4, the coordinates x, y, z are realfc, (2)), imag (Ci (2)) and abs (c, (1)), respectively. It should be noted that the so-called Steifel manifold of complex 2D space is not the same as the real 3D unit spherical surface. Therefore, the code words may not evenly occupy the spherical surface.

The vector along the positive y-axis illustrates the first code word. The next 5 code words form a "Polar Cap" near the first code word. In the aforementioned tracking mode, the beamforming matrix can be quantized by the code words in the "polar cap", thereby reducing the need to search the entire codebook. This particular ordering of the code words reduces the complexity of quantization in the tracking mode. Furthermore, the ordering of the codebook also allows a flexible microarchitectural implementation of the quantization process. In this illustrated example, the circuit only needs to search the first 6 code words to ensure beamforming convergence when the channel changes slowly, while a higher performance device can implement full search complexity for a faster convergence.

As discussed earlier, in a stationary channel condition, the ultimate beamforming accuracy only depends on the codeword density near the polar cap. Consequently, CMA agent 108 selectively enhances the codebook by introducing a denser "polar cap" for tracking purposes, block 306. By way of example only, it is provided in Table 2 below, where the

ES 2 553 344 T3 codeword C ^ u (2,12 + 4) is introduced where an extra set of m apales (p e., Where m is 4 in this example) is added near the so-called “north pole ":

1.0000 + O.OOOOi

- 0.0000Í

0.9962

0.0872

0.8507 + 0.0000Í

0.2532 - 0.4607Í

0.5257 + O.OOOOi

-0.8356 - 0.1593i

0.9962

-i- 0.0872Í

0.8507 + O.OOOOi

-0.3599 - 0.3832Í

0.8507 + O.OOOOi

-0.4757 + 0.2239Í

0.5257 - O.OOOOi

-0.4097 + 0.7455Í

0.9962

-0.0872

0.8507-O.OOOOi

0.0659 + 0.5216Í

0.5257 - O.OOOOi

0.5824 + 0.6200Í

0.5257-O.OOOOi

0.7697 -0.3623Í

0.9962

- 0.0872Í

0.8507 - O.OOOOi

0.5164 + 0.0985Í

0.5257 - O.OOOOi

-0.1067 - 0.8439Í

0.0000 + O.OOOOi

-0.9967-0.0818i

A graphical representation of such a non-uniform codebook is presented with reference to Figure 5. In accordance with the example illustrated in Figure 5, the graphical representation of the codebook shows the additional m codewords (4) and the words codes (in black) grouped near the first code words.

Once the non-uniform codebook is generated, the CMA agent 108 can perform a selective quantization of the non-uniform codebook for forwarding to a distant communication device (eg, 102), block 308, as explicit operational feedback , continuing the closed-loop MIMO technology process.

Returning to Figure 6, a flow chart of an exemplary method for generating a hierarchical codebook is presented in accordance with one embodiment. At block 602, CMA agent 108 selects a first codebook. In accordance with one embodiment, the first codebook is selected for coarse quantization. In accordance with one embodiment, the thick codebook has N_c Vectors and vectors have an M dimension. If the channel matrix inputs are independently and identically distributed, it may be desirable for the vectors to be as uniformly distributed as possible in a complex M-dimensional Steifel manifold (similar to the unit sphere) . If not, the vectors may not be uniformly distributed.

At block 604, the CMA agent 108 selects a second and / or subsequent codebooks. In accordance with one embodiment, the second codebook is selected for fine quantization, around, eg, a selected center.

At block 606, CMA agent 108 may concatenate at least subsets of each of the selected first and second (and / or subsequent) codebooks to form a hierarchy codebook. Although the concatenated codebook may not provide the best performance for a given feedback load, it allows for low complexity quantization and scalable operational feedback resolution. Furthermore, since the beamforming matrix approximates the identity matrix after several operational feedbacks, the station can only use ring codebooks without vector codebooks for quantization of some vectors in a slow fading mode. . The foregoing reduces the operational feedback load for beamforming matrix tracking.

A graphical representation of an exemplary hierarchy codebook is presented with reference to Figure 7. As illustrated, for each vector of the coarse codebook, a fine codebook is defined. The thin codebook vectors (indicated as η s) surround the selected vectors from the thick codebook (indicated as v,). The fine code book can be dynamically generated in accordance with v, of the co-pending Lin patent application to reduce the complexity of memorization. The fine vector η s can constitute a cap, or a ring, or other shape in the collector. The profile size of the fine codebook can be dynamically adjusted during tracking to reduce quantization error. By way of example, the initial operational feedback may employ a large profile size since the quantization error due to the thick codebook is large and the latter tracking feedback may be small because the beamforming matrix is ​​close to the identity matrix.

An example embodiment of a fine code book is as follows. The thin vectors η s of the thick vector v, are located in a dimensional (complex) ring M-1 (or circle), whose center is the M-dimensional vector v, (complex) and the plane containing the ring is perpendicular to the vector v, as illustrated in Figure 7. As used herein, vector v can have M complex dimensions or 2M real dimensions. The radius of the ring can be dynamically adjusted to reduce quantization error. Fine vectors can be calculated as:

ES 2 553 344 T3

<img file="ES2553344T3_D0005.tif" />

eos # sin # f_} (OR)

<img file="ES2553344T3_D0006.tif" />

<img file="ES2553344T3_D0007.tif" />

where Θ is the angle between r, and v,; · ^ · ~¹ 2h; .w _{it is} |_{to ma}t_r¡_z d_and Householder reflection of 'y

r. f 'is an M-1 dimensional unit vector for the jth vectors in the ring. Since F, is unitary and its first column is v ,, all columns except the first form the plane orthogonal to V, which contains the ring. Since the Householder matrix is ​​easy to compute, this system is computationally efficient.

In general, the matrix F can be replaced by any other unit matrix where the first column is v ,. By way of example, F can be substituted for B, in (10). This substitution can allow implementation using CORDIC algorithms. The vectors f¡ s are obtained from search as:

argmax min Ileos² #₀ + without * #₀ //

<img file="ES2553344T3_D0008.tif" />

In accordance with one embodiment, the search increases (eg, substantially maximizes) the distance

COS0_ORbetween the two closest vectors that have the general form L^without^ ° I

Since the search result is not sensitive to θ₀, which determines the radius of the ring, fine codebooks with different radii can be dynamically generated by keeping the same value of f, and changing Θ in (10). In searching for (11), θο can be determined from the quantization error statistic. For the 4x1 unit complex vector quantization, the angle θο is approximately 15 degrees and θ is approximately 20 degrees for the initial operational feedback. The radius Θ can usually be reduced for later follow-up feedbacks. For the OFDM system, the value of Θ can remain constant for all subcarriers for operational feedback.

According to one embodiment, the fine codebook consists of the central vector, v ,, and the ring vectors η, for j = 1, ..., Nf. The total number of vectors in the fine codebook is Nf +1, which can be a power of 2.

By way of another example, the fine codebook may not only have codeword vectors in the ring. Instead, the thin codeword vectors can effect polar cap scattering having the thick codebook vector as the center and the ring as the boundary. The fine vectors can be uniformly distributed in the cap or they can have a higher density the closer they are to the center. The distribution can be obtained by optimizing the quantization accuracy (that is, reducing the quantization error) for a given operational feedback load.

As previously presented, the CMA agent 108 can concatenate the first (thick) codebook with one or more subsequent (thin) codebooks, block 606, prior to quantization and forward to a distant communication device, block 608. It should be appreciated that concatenating a thick codebook with a thin codebook reduces the quantization error of the thick codebook at the cost of an additional feedback load for the thin codebook. If the resolution of the concatenated codebook is not sufficient, you can also concatenate another fine codebook, where the vector of the first concatenated codebook is the center of the second concatenated codebook.

In the tracking mode presented above, the accumulated beamforming matrix is ​​close to the identity matrix, the quantization is only needed to be around [1 0, ..., 0]¹ for each Householder vector. Therefore, the CMA agent 108 can rotate the entire thick codebook so that a codeword vector is [1 0, ..., 0]¹. This rotation reduces the quantization error for the tracking mode. However, the rotation is undesirable from the point of view of maximum power since it can put all the transmitting power in a single antenna. When the beamforming matrix observed at the receiver in co-pending Lin patent application approximates the identity matrix, the center can be set to [1 0, ..., 0]¹ and coarse quantization can be omitted. Only fine quantization and corresponding operational feedback need to be used, thereby reducing computational complexity and operational feedback burden. This simplification is referred to as localization. Since the quantization error statistic is different for different Householder vectors and different operational feedbacks, concatenation, location, and radius adjustment for each Householder vector are used for each operational feedback.

<img file="ES2553344T3_D0009.tif" />

ES 2 553 344 T3

For illustration purposes, it is assumed, by way of example, that a transmitter has 4 antennas and a receiver has 3 antennas. For initial operational feedback, the 6, 5, and 4 quantization bits are used for the coarse quantization of 4, 3, and 2 vectors (from Householder) respectively and 3 and 4 bits are used for the fine quantization for 4 and 3 vectors for improve resolution, respectively, whose radius Θ is 20 degrees for both. For the second form of operational feedback, the coarse 4-vector quantization is omitted. The radius is reduced to 15 degrees for 4 and 3 vector fine codebooks.

Exemplary realization of communication device architecture

Having presented the communications environment and operational characteristics of the CMA 108 with respect to Figures 1 through 7 inclusive above, reference is now made to Figure 8 which provides an exemplary electronic device architecture within which you can put into practice the CMA 108.

Figure 8 illustrates a flow diagram of an exemplary architecture of an electronic device within which the teachings provided by the present invention can be practiced, in accordance with one embodiment. The electronic device 800 is illustrated including one or more antennas, a radio frequency (RF) front end (RFE) 802, a baseband processor 804, one or more network interfaces 406, one or more processors 808 (eg, applications and / or general use) and a memory 810. In accordance with embodiments of the invention, device 800 may suitably implement one or more aspects of codebook management agent (CMA) 108 and / or the previously entered data structures themselves.

In some embodiments, electronic device 800 may represent a system that includes one or more than one access point, mobile station, base station, and / or subscriber unit, and may include other circuitry. By way of example, in some embodiments, electronic device 800 may be a computer, such as a personal computer, workstation, or similar device, that includes an access point or mobile station as an integral element and / or peripheral. Additionally, electronic device 800 can include a number of devices that are coupled together in a network.

In operation, device 800 can send and receive signals using one or more of the antennas, where the signals are processed by the various elements illustrated. As used herein, the antennas can be a matrix array of antennas or any type of antenna structure that supports MIMO technology processing. According to one embodiment, said antennas are practically omni-directional antennas but the scope of the invention is not limited in this regard. Device 800 may operate in partial or full compliance with the wireless network standard such as, eg, the previously introduced 802.11 or 802.16 standards.

In accordance with one embodiment, the front end of RF 802 can be selectively coupled to one or more of the antennas to interact with a wireless network. The front end of RF 802 may include circuitry to support the transmission and reception of radio frequency (RF) signals. By way of example, in some embodiments, the RF front end 802 may include an RF receiver to receive signals and perform one or more processing tasks such as low noise amplification (LNA), filtering, frequency conversion or similar operations. In addition, in some embodiments, the radio frequency RFE 802 may include transformation mechanisms and beamforming circuits to support MIMO signal processing. Additionally, by way of example, in some embodiments, RFE 802 may include circuitry to support frequency upconversion and an RF transmitter.

Baseband processors 804 can be a transmitter with hardware, software, and / or firmware (or combinations thereof) to perform baseband processing. The baseband processor can be combined or implemented through the 808 processor.

Processor 808 can read instructions and data from memory 810 and perform actions on its response. By way of example, processor 808 can access instructions from memory 810 and perform embodiments of the method of the present invention, such as, eg, method 300 (Figure 2), method 600 (Figure 6), and / or other methods described here. In this regard, processor 808 is intended to represent any type of processor including, without limitation, a microprocessor, a digital signal processor, a microcontroller, or the like.

Memory 810 represents an item that includes machine-readable media. By way of example, memory 810 represents random access memory (RAM), dynamic random access memory (DRAM), static random access memory (SRAM), read-only memory (ROM), flash memory. or any other type of item that includes 808 processor-readable media. The memory 810 may store instructions to dynamically generate and / or use the previously entered hierarchical and / or non-uniform codebooks. Alternatively, such instructions can also be integrated into and / or provided to a baseband processor 804.

Network interface 806 can provide communications between electronic device 800 and other systems. TO

By way of example, in some embodiments, electronic device 800 may be an access point that uses network interface 806 to communicate with a wired network or to communicate with other access points. In some embodiments, electronic device 800 may be a network interface card (NIC) that communicates with a computer or network using a bus connector or other type of port.

As used herein, the CMA 108 embodiments may be implemented in one or more of RFEs 802, a baseband processor 804, processors 808, and / or combinations thereof. As discussed earlier, the CMA agent 108 can be implemented in hardware, software, firmware, or combinations thereof.

Although the various elements of the device 800 are illustrated as disparate elements in Figure 8, it is considered that the embodiments may combine one or more elements or that they may contain more elements. By way of example, the circuitry of the processor 808, memory 810, network interface 806, and baseband processor 804 can be suitably integrated into a single integrated circuit. Alternatively, memory 810 may be internal memory within a baseband processor 804 or processor 808 or it may be a firmware control memory within processor 810. In some embodiments, the various elements of device 400 may be packaged. separately and mounted on a common circuit board. In another embodiment, the various elements are separate IC segments packed together, such as a multiple IC module, and in other embodiments, multiple elements are on the same IC chip.

Alternative forms of realization

Figure 9 illustrates a flowchart of an exemplary storage medium comprising content that, when requested, can cause an access machine to implement one or more operational aspects of the book management agent. codes 108 and / or associated methods 300 and / or 600 and / or associated data structures (eg, codebooks). In this regard, the storage medium 900 may include content 902 (eg, instructions, data, or any combination thereof) that, when executed, causes an access device to implement one or more aspects of the book management agent. code 108 previously described.

Machine-readable media (memory) 900 may include, without limitation, floppy disks, optical discs, CDROMs and magneto-optical discs, ROMs, RAM, EPROM, EEPROM, magnetic or optical cards, flash memory, or other readable media. by machine / multimedia suitable for memorizing electronic instructions. Furthermore, the present invention can also be downloaded as a computer program product, wherein the program can be transferred from a distant computer to a requesting computer via data signals embedded in a carrier wave or other propagation medium over a link. communications (eg, a modem, radio or network connection). As used herein, all such media are considered, in the broad sense, as storage media.

It should be understood that embodiments of the present invention can be used in a variety of applications. Although the present invention is not limited in this regard, the circuits disclosed herein can be used in numerous apparatus such as transmitters and receivers of a radio system. Radio systems intended to be included within the scope of the present invention include, by way of example only, wireless local area network (WLAN) devices and wide area network (WWAN) devices that include network interface devices. wireless and network interface cards (NICs), base stations, access points (APs), gateways, bridges, hubs, cellular radiotelephone communications systems, satellite communications systems, two-way radio communications systems, one-way paging devices, two-way paging devices, personal communication systems (PCS), personal computers (PCs), personal digital assistants (PDAs), sensor networks, personal area networks (PANs) and similar devices , although the scope of the invention is not limited in this regard. Such devices can suitably be used within any of a variety of elements.

Embodiments of the present invention can also be included in integrated circuit blocks, referred to as core memory, cache memory or other types of memory that store electronic instructions to be executed by the microprocessor or store data that can be used in arithmetic operations. In general, an embodiment using so-called multistage domino logic in accordance with the claimed subject matter can provide an operational advantage to microprocessors, in particular they can be incorporated into an address decoder for a memory device. It should be noted that the embodiments can be integrated into radio systems or portable devices, in particular when the devices depend on low power consumption. Thus, laptop computers, cellular radiotelephone communication systems, two-way radio communication systems, one-way search devices, two-way search devices, personal communication systems (PCS), personal digital assistants (PDAs), cameras and other products are available. intended to be included within the scope of protection of the present invention.

The present invention includes various operations. The operations of the present invention can be performed

ES 2 553 344 T3 using hardware components, or they can be embodied in machine-executable content (eg, instructions) that can be used to make a general-purpose or special-purpose processor or logic circuits programmed with the instructions to perform the intended operations . Alternatively, the operations can be performed using a combination of hardware and software. Furthermore, although the invention has been described within the context of a computing device, those skilled in the art will appreciate that such functionality may suitably be embodied in any of several alternative embodiments such as, by way of example, integrated within a computing device. communications (eg, a mobile phone).

In the foregoing description, for the purpose of explanation, numerous specific details are set forth in order to provide a thorough understanding of the present invention. It will be apparent, however, to those skilled in this art, that the present invention can be practiced without some of these particular details. In other operational instances, well known structures and devices are illustrated in block diagram form. Any number of variants of the inventive concept are envisaged within the scope of the present invention. In this regard, the particular exemplary illustrated embodiments are not provided to limit the invention but merely to illustrate it. Thus, the scope of the present invention is not to be determined by the particular exemplary embodiments disclosed above but only by the content of the following claims.