RU2437220C2 - Calibration of channel for communication system with duplex communication and time division channelling - Google Patents

Abstract

FIELD: information technologies.

SUBSTANCE: in one of realisation versions, pilot signals are transferred along a downlink and an upperlink, and are used to produce estimate responses of the downlink and the upperlink, accordingly. Then two sets of correction factors are identified, being based on estimate responses of the downlink and the upperlink. The calibrated downlink is formed using the first set of correction factors for the downlink, and the calibrated upperlink is formed using the second set of correction factors for the upperlink. The first and the second sets of correction factors may be determined using calculation of matrices ratio or calculation with a minimum mean square error (MMSE). Calibration may be carried out on a real time basis based on their transfer along a radio channel.

EFFECT: account of differences in frequency responses of transmitting and receiving circuits in a point of access and a user terminal.

38 cl, 6 dwg

Description

Priority claim by 35 U.S.C. § 119.

This patent application claims priority for provisional patent application US No. 60/421462, entitled “Channel Calibration for a Time Division Duplexed Communication System”, and provisional patent application US No. 60/421309, entitled “MIMO WLAN System”, both filed October 25, 2002, and the rights to which belong to the copyright holder of this patent application, and which are incorporated into this description in its entirety by reference.

FIELD OF THE INVENTION

The present invention relates generally to communications, and more specifically to methods for calibrating downlink and uplink responses in a full duplex and time division multiplexing (TDD) communication system.

BACKGROUND

In a wireless communication system, data transmission between the access point and the user terminal occurs wirelessly. Depending on the design of the system, the same or different frequency bands can be used for the downlink and the uplink. A downlink (or straight line) refers to transmission from an access point to a user terminal, and an uplink (or reverse line) refers to transmission from a user terminal to an access point. If two frequency bands are available, then the downlink and uplink can be transmitted in separate frequency bands using frequency division duplex (FDD). If only one frequency band is available, then the downlink and the uplink can share the same frequency band using time division duplex (TDD).

To achieve high performance, it is often necessary to know the frequency response of a wireless channel. For example, a downlink response may be needed by an access point to perform spatial processing (described below) for downlink data to a user terminal. The downlink response can be estimated by the user terminal based on the pilot transmitted by the access point. The user terminal may then send the channel estimate back to the access point for future use. For such a channel estimation scheme, it is necessary to transmit a pilot signal in a downlink and sending the channel estimate to the access point causes additional delays and requires additional resources.

представляет матрицу отклика канала от антенной решетки A до антенной решетки B, то взаимно-обратный канал подразумевает, что соединение от решетки B к решетке A дается

, где

обозначает транспонированную матрицу

. Таким образом, для TDD системы отклик канала для одной линии может быть оценен, основываясь на пилот-сигнале, посланном по другой линии. Например, отклик восходящего канала может быть оценен, основываясь на пилот-сигнале восходящей линии, и транспонированный отклик восходящего канала может быть использован в качестве оценки отклика нисходящего канала. For TDD systems with a common frequency band, it can be assumed that the responses of the downlink and uplink are mutually inverse. That is, if

represents the response matrix of the channel from the antenna array A to the antenna array B, then the reciprocal channel implies that the connection from the array B to the array A is given

where

denotes the transposed matrix

. Thus, for a TDD system, the channel response for one line can be estimated based on the pilot sent on the other line. For example, the uplink channel response can be estimated based on the uplink pilot, and the transposed uplink response can be used as an estimate of the downlink response.

However, the frequency responses of the transmitting and receiving circuits at the access point usually differ from the frequency responses of the transmitting and receiving circuits at the user terminal. In particular, the frequency responses of the transmit / receive chains used for uplink transmission may be different from the frequency responses of the transmit / receive chains used for downlink transmission. The “effective” response of the downstream channel (that is, including the transmit / receive chains) may be different from the response of the reciprocal effective downstream channel due to differences in the transmit / receive chains (ie, the responses of the effective channels are not mutually reverse). If the reciprocal inverse channel response estimate obtained for one line is used for spatial processing in another line, then the difference in the frequency responses of the transmit / receive circuits is an error that, if not determined and taken into account, can cause degradation in performance.

Thus, in the art there is a need for methods for calibrating downlink and uplink in a TDD communication system.

SUMMARY OF THE INVENTION

Methods for calibrating the downstream and upstream channels to account for differences in the frequency responses of the transmitter and receiver chains at the access point and user terminal are described herein. After calibration, the channel response estimate obtained for one line can be used to obtain the channel response estimate for the other line. This simplifies channel estimation and spatial processing.

In one embodiment, a method for calibrating downlink and uplink channels in a multi-input multi-output wireless TDD communication system (MIMO) is provided. According to the method, a pilot signal is transmitted on the uplink and is used to derive an estimate of the response of the uplink. The pilot is also transmitted in the downlink and is used to derive an estimate of the response of the downlink. Then, two sets of correction factors are determined based on the estimates of the responses of the downstream and upstream channels. A calibrated downlink is formed by using the first set of correction factors for the downlink, and a calibrated uplink is formed by using the second set of correction factors for the uplink. Corresponding correction factors are used in the respective transmitters for the downstream and upstream channels. The responses of the calibrated downstream and upstream channels are approximately mutually inverse due to two sets of correction factors. The first and second sets of correction factors can be determined using matrix ratio calculation or minimum mean square error (MMSE) calculation, as described below.

Calibration can be performed in real time based on radio transmission. Each user terminal in the system can output a second set of correction factors for use in it. The first set of correction factors for the access point may be output by a plurality of terminals. For an orthogonal frequency division multiplexing (OFDM) system, calibration can be performed for a first set of subbands to obtain two sets of correction factors for each subband in the set. Correction factors for other “uncalibrated” subbands can be interpolated based on correction factors obtained for the “calibrated” subbands.

Various aspects of embodiments of the present invention are described in more detail below.

BRIEF DESCRIPTION OF THE DRAWINGS

Distinctive features, the essence and advantage of the present invention will become more apparent from the following detailed description, taken in conjunction with the drawings, in which the same reference position denotes the same elements.

Figure 1 shows the transmit and receive chains at the access point and user terminal of the MIMO system;

Figure 2 illustrates the use of correction factors to account for differences in the transmit / receive chains of the access point and user terminal;

figure 3 shows the calibration process of the responses of the downward and upward channels in the TDD MIMO-OFDM system;

figure 4 shows the process of deriving estimates of correction vectors from estimates of responses of the downward and upward channels;

5 is a block diagram of an access point and user terminal; and

FIG. 6 is a block diagram of a TX spatial processor.

DETAILED DESCRIPTION OF THE INVENTION

The calibration methods described herein can be used for various communication systems. In particular, these methods can be used in systems with one input and one output (SISO), systems with multiple inputs and one output (MISO), systems with one input and many outputs (SIMO), and systems with many inputs and multiple outputs ( MIMO).

A MIMO system uses multiple (N _T ) transmit antennas and multiple (N _R ) receive antennas for data transmission. The MIMO channel formed by N _T transmit and N _R receive antennas can be divided into N _S independent channels, with N _S ≤min {N _T , N _R }. Each of the N _S independent channels is also called a spatial channel or its own MIMO channel mode and corresponds to a dimension. A MIMO system can provide improved performance (i.e., increased transmission capacity) if additional dimensions created by multiple transmit and receive antennas are used. This usually requires an accurate estimate of the channel response between the transmitter and receiver.

In FIG. 1 shows a block diagram of a transmitter and receiver chain at an access point 102 and a user terminal 104 in a MIMO system. For this system, the uplink and downlink share the same frequency range in a time division duplex manner.

“передачи”) обрабатываются передающей цепью (TMTR) 114 и передаются через N_ap антенны 116 по беспроводному каналу. В пользовательском терминале 104 сигналы нисходящей линии принимаются N_ut антеннами 152 и обрабатываются приемной цепью (RCVR) 154 для обеспечения принятых символов (обозначаемых вектором

“приема”). Обработка в передающей цепи 114, как правило, включает в себя цифроаналоговое преобразование, усиление, фильтрацию, повышение частоты и т.д. Обработка в приемной цепи 154, как правило, включает в себя понижение частоты, усиление, фильтрацию, аналого-цифровое преобразование и т.д.For a downlink at access point 102, symbols (defined by a vector

“Transmissions”) are processed by a transmit chain (TMTR) 114 and transmitted via an N _ap antenna 116 over a wireless channel. At user terminal 104, downlink signals are received by N _ut antennas 152 and processed by a receive circuit (RCVR) 154 to provide received symbols (denoted by a vector

“Reception”). Processing in the transmit chain 114 typically includes digital-to-analog conversion, amplification, filtering, frequency boost, etc. Processing in receive chain 154 typically includes down-frequency, gain, filtering, analog-to-digital conversion, etc.

передачи) обрабатываются в передающей цепи 164 и передаются через N_ut антенны 152 по беспроводному каналу. В точке 102 доступа сигналы восходящей линии принимаются N_ap антеннами 116 и обрабатываются в приемной цепи 124 для обеспечения принятых символов (обозначаемых вектором

приема).In the case of an uplink in the user terminal 104 characters (denoted by a vector

transmissions) are processed in the transmit chain 164 and transmitted via N _ut antennas 152 over a wireless channel. At access point 102, the uplink signals are received by N _ap antennas 116 and processed in the receive chain 124 to provide received symbols (denoted by a vector

reception).

In the case of a downlink, the receive vector in the user terminal may be expressed as

,

, equation (1)

представляет собой вектор передачи с N_ap элементами для символов, передаваемых через N_ap антенны точки доступа;Where

represents a transmission vector with N _ap elements for symbols transmitted through the N _ap antennas of the access point;

представляет собой вектор приема с N_ut элементами для символов, принятых N_ut антеннами в пользовательском терминале;

represents a receive vector with N _ut elements for symbols received by N _ut antennas in a user terminal;

представляет собой N_ap×N_ap диагональную матрицу с элементами, представляющими собой комплексные усиления, связанные с передающей цепью для N_ap антенн в точке доступа;

represents an N _ap × N _ap diagonal matrix with elements representing complex amplifications associated with the transmit chain for the N _ap antennas at the access point;

представляет собой N_ut×N_ut диагональную матрицу с элементами, представляющими собой комплексные усиления, связанные с приемной цепью для N_ut антенн в пользовательском терминале;

represents a N _ut × N _ut diagonal matrix with elements representing complex amplifications associated with the receive circuit for N _ut antennas in a user terminal;

представляет собой N_ut×N_ap матрицу откликов каналов для нисходящей линии.

represents the N _ut × N _ap channel response matrix for the downlink.

The responses of the transmit / receive circuits and the wireless channel are typically a function of frequency. For simplicity, it is assumed that the channel is a channel with amplitude fading (i.e., with a uniform frequency response).

In the case of an uplink, the receive vector at the access point can be expressed as

equation (2)

представляет собой вектор передачи для символов, передаваемых через N_ut антенны пользовательского терминала;Where

represents a transmission vector for symbols transmitted through the N _ut antennas of a user terminal;

представляет собой вектор приема для символов, принятых N_ap антеннами в точке доступа;

represents a receive vector for the symbols received by the N _ap antennas at the access point;

представляет собой N_ut×N_ut диагональную матрицу с элементами в виде комплексных усилений, связанных с передающей цепью для N_ut антенн в пользовательском терминале;

represents a N _ut × N _ut diagonal matrix with elements in the form of complex amplifications associated with the transmission chain for N _ut antennas in the user terminal;

представляет собой N_ap×N_ap диагональную матрицу с элементами в виде комплексных усилений, связанных с приемной цепью для N_ap антенн в точке доступа; и

represents an N _ap × N _ap diagonal matrix with elements in the form of complex amplifications associated with the receive circuit for N _ap antennas at the access point; and

представляет собой N_ap×N_ut матрицу откликов каналов для восходящей линии.

represents the N _ap × N _ut channel response matrix for the uplink.

,

, соответственно, как показано в уравнениях (1) и (2). Однако отклики передающих/приемных цепей в точке доступа, как правило, не совпадают с откликами передающих/приемных цепей в пользовательском терминале. Указанные различия приводят к неравенству

.For a TDD system, since the downlink and uplink share the same frequency range, there is usually a strong correlation between the responses of the downlink and uplink. Thus, the response matrices of the descending and ascending channels can be considered as mutually inverse (i.e., transposed) with respect to each other, and denoted as

,

, respectively, as shown in equations (1) and (2). However, the responses of the transmit / receive chains at the access point, as a rule, do not coincide with the responses of the transmit / receive chains in the user terminal. These differences lead to inequality

.

и

, которые включают в себя отклики, используемых передающих и приемных цепей, могут быть выражены какFrom equations (1) and (2), the “effective” responses of the downward and upward channels,

and

, which include the responses used by the transmit and receive chains, can be expressed as

и

and

equation (3)

Combining these two equations and equation (3), the following relation can be obtained:

equation (4)

Transforming equation (4), we obtain the following:

или

or

equation (5)

и

.Where

and

.

Equation (5) can also be represented as:

.

. equation (6)

и

, в эффективных откликах нисходящего и восходящего каналов, как показано в уравнении (6), позволяет выразить отклики калиброванных каналов для нисходящей линии и восходящей линии как результат транспонирования друг друга. (N_ap×N_ap) диагональная матрица

для точки доступа представляет собой отношение отклика

приемной цепи к отклику

передающей цепи (т.е.

), причем отношение является поэлементным отношением. Аналогично (N_ut×N_ut) диагональная матрица

для пользовательского терминала представляет собой отношение отклика

приемной цепи и отклика

передающей цепи.The left side of equation (6) represents the response of the calibrated uplink channel, and the right side represents the transposed response of the calibrated downlink channel. Using diagonal matrices,

and

, in the effective responses of the descending and ascending channels, as shown in equation (6), allows expressing the responses of the calibrated channels for the descending line and the ascending line as a result of transposing each other. (N _ap × N _ap ) diagonal matrix

for an access point is the response ratio

receive chain to response

transmission chain (i.e.

), and the relation is an elementwise relation. Similarly (N _ut × N _ut ) diagonal matrix

for the user terminal is the response ratio

receive circuit and response

transmission chain.

и

включают в себя значение, учитывающие различия в передающих/приемных цепях точки доступа и пользовательского терминала. Это позволяет отклик канала для одной линии выразить через отклик канала для другой линии, как показано в уравнении (6).Matrices

and

include a value that takes into account the differences in the transmit / receive chains of the access point and user terminal. This allows the channel response for one line to be expressed through the channel response for another line, as shown in equation (6).

и

может быть выполнена калибровка. Как правило, истинный отклик

канала и отклики передающей/приемной цепи не являются известными, а также не могут быть легко и точно получены. Напротив, эффективные отклики нисходящего и восходящего каналов,

и

, могут быть оценены, основываясь на пилот-сигналах, передаваемых по нисходящей линии и восходящей линии, соответственно, как описано ниже. Затем оценки матриц

и

, которые называются поправочными матрицами

и

, могут быть выведены, основываясь на оценках откликов нисходящего и восходящего каналов,

и

, как описано ниже. Матрицы

и

включают в себя поправочные факторы, которые позволяют учитывать различия в передающих/приемных цепях точки доступа и пользовательского терминала. To determine the matrices

and

calibration can be performed. Usually true response

The channel and responses of the transmit / receive chain are not known, nor can they be easily and accurately received. On the contrary, the effective responses of the downward and upward channels,

and

can be estimated based on the pilot signals transmitted on the downlink and uplink, respectively, as described below. Then matrix scores

and

called correction matrices

and

can be derived based on the estimates of the responses of the downstream and upstream channels,

and

as described below. Matrices

and

include correction factors that allow for differences in the transmit / receive chains of the access point and user terminal.

и

для учета различий в передающих/приемных цепях точки доступа и пользовательского терминала. В случае нисходящей линии вектор

передачи сначала умножают на матрицу

в блоке 112. Последующая обработка в передающей цепи 114 и приемной цепи 154 для нисходящей линии является такой же, как показано на Фиг. 1. Аналогично, в случае восходящей линии вектор

передачи сначала умножают на матрицу

в блоке 162. Опять же последующая обработка в передающей цепи 164 и приемной цепи 124 для восходящей линии является такой же, как показано на Фиг. 1.FIG. 2 illustrates the use of correction matrices

and

to account for differences in the transmit / receive chains of the access point and user terminal. In the case of a downward line, the vector

gears are first multiplied by the matrix

at block 112. Subsequent processing in the transmit chain 114 and the receive chain 154 for the downlink is the same as shown in FIG. 1. Similarly, in the case of an ascending line, the vector

gears are first multiplied by the matrix

at block 162. Again, subsequent processing in the transmit chain 164 and the receive chain 124 for the uplink is the same as shown in FIG. one.

The “calibrated” responses of the downstream and upstream channels, visible in the user terminal and access point, respectively, can be expressed as

и

and

equation (7)

и

представляют собой выражения для оценки “истинных” откликов калиброванных каналов в уравнении (6). Комбинируя два уравнения набора уравнений (7) с использованием выражения из уравнения (6), можно показать, что

. Точность отношения

зависит от точности матриц

и

, которая в свою очередь, как правило, зависит от качества оценок откликов нисходящего и восходящего каналов,

и

.Where

and

are expressions for estimating the “true” responses of calibrated channels in equation (6). Combining the two equations in the set of equations (7) using the expression from equation (6), we can show that

. Relationship accuracy

depends on the accuracy of the matrices

and

, which, in turn, usually depends on the quality of the responses of the downstream and upstream channels,

and

.

), может быть использована для определения оценки отклика калиброванного канала для другой линии (например,

).As shown above, calibration can be performed in a TDD system to determine the differences in the responses of the transmit / receive chains at the access point and user terminal, and to account for these differences. After calibrating the transmit / receive circuits, an estimate of the response of the calibrated channel obtained for one line (for example,

), can be used to determine the estimated response of a calibrated channel for another line (for example,

)

The calibration methods described herein can also be used for wireless communication systems that use OFDM. With OFDM, the entire system frequency range is effectively divided into several (N _F ) orthogonal subbands, which are also called frequency bins or subchannels. In the case of OFDM, each subband is associated with a corresponding subcarrier that can be modulated with data. For a MIMO system that uses OFDM (i.e., a MIMO-OFDM system), each subband of each eigenmode can be considered an independent transmission channel.

Calibration can be performed in various ways. For clarity, a specific calibration scheme is described below for a TDD MIMO-OFDM system. For such a system, each subband of the wireless line can be considered mutually inverse.

In FIG. 3 is a flowchart of a process 300 for calibrating downlink and uplink responses in a TDD MIMO-OFDM system. First, the user terminal receives the timing and frequency of the access point using the acquisition procedures specific to the system (step 310). The user terminal may then send a message to initiate calibration by the access point, or calibration may be initiated by the access point. Calibration can be performed in parallel with the registration / authentication of the user terminal by the access point (for example, during call setup) and can also be performed as needed at any time.

Calibration can be performed for all subbands that can be used to transmit data (which are called “data” subbands). Subbands not used for data transmission (i.e., guard subbands) generally do not require calibration. However, since the frequency responses of the transmitter / receiver chains at the access point and user terminal are usually uniform across most frequency bands of interest, and since adjacent subbands are likely to be correlated, calibration can only be performed on a subset of the data subbands. If not all data subbands are calibrated, then information about the subbands intended for calibration (which are referred to as “assigned” subbands) can be sent to the access point (for example, in a message sent to initiate calibration).

For calibration, the user terminal transmits a MIMO pilot on the assigned subbands to the access point (step 312). MIMO pilot generation is described in more detail below. The duration of the uplink transmission of the MIMO pilot may depend on the number of assigned subbands. For example, 8 OFDM symbols may be sufficient if calibration is performed for four subbands, and for a larger number of subbands, a larger number (e.g. 20) OFDM symbols may be required. Typically, the total transmit power is fixed, so if a MIMO pilot is transmitted over a small number of subbands, then a higher level of transmit power can be used for each of these subbands and the SNR for each subband will be high. In contrast, if the MIMO pilot is transmitted over a large number of subbands, then a lower transmit power level will be used for each subband, and the SNR for each subband will be worse. If the SNR for each subband is not high enough, then a larger number of OFDM symbols can be sent for the MIMO pilot, which are integrated at the receiver to obtain a higher overall SNR for that subband.

, для каждого из назначенных поддиапазонов, где k представляет собой индекс поддиапазона. Оценка канала, основанная на MIMO пилот-сигнале, описана ниже. Оценки откликов восходящих каналов оцифровывают и посылают в пользовательский терминал (этап 314). Элементы каждой матрицы

представляют собой комплексные усиления каналов между N_ut передающими и N_ap приемными антеннами для восходящей линии для k-го поддиапазона. Усиление каналов для всех матриц могут быть масштабированы на конкретный фактор масштабирования, который является общим для всех назначенных поддиапазонов, для получения требуемого динамического диапазона. Например, усиления каналов в каждой матрице

могут быть единообразно масштабированы на наибольшее усиление канала для всех матриц

для назначенных поддиапазонов таким образом, чтобы величина наибольшего усиления канала составляла единицу. Поскольку задачей калибровки является нормализация различий в усилении/фазе между нисходящей линией и восходящей линией, абсолютные усиления каналов не являются важными. Если 12-битные комплексные значения (т.е. с 12-битными синфазными (I) и 12-битными квадратурными (Q) компонентами) используют для представления усилений канала, тогда оценки откликов нисходящих каналов могут быть отправлены в пользовательский терминал в 3·N_ut·N_ap·N_sb в байтах, где “3” возникает вследствие того, что для представления I и Q компонентов используют 24 бита и N_sb представляет собой количество назначенных поддиапазонов.The access point receives the MIMO pilot on the uplink and outputs an estimate of the uplink response,

, for each of the assigned subbands, where k is a subband index. Channel estimation based on a MIMO pilot is described below. The uplink channel response estimates are digitized and sent to the user terminal (block 314). Elements of each matrix

are the complex channel gains between the N _ut transmit and N _ap receive antennas for the uplink for the kth subband. The channel gain for all matrices can be scaled to a specific scaling factor that is common to all assigned subbands to obtain the desired dynamic range. For example, channel gains in each matrix

can be uniformly scaled to the highest channel gain for all matrices

for the assigned subbands so that the magnitude of the maximum channel gain is one. Since the purpose of calibration is to normalize the differences in gain / phase between the downlink and the uplink, the absolute channel gains are not important. If 12-bit complex values (i.e. with 12-bit in-phase (I) and 12-bit quadrature (Q) components) are used to represent channel gains, then downlink channel response estimates can be sent to the user terminal at 3 · N _ut · N _ap · N _sb in bytes, where “3” is due to the fact that 24 bits are used to represent the I and Q components and N _sb represents the number of assigned subbands.

, для каждого из назначенных поддиапазонов, основываясь на принятом пилот-сигнале (этап 318). Затем пользовательский терминал определяет поправочные факторы,

и

, для каждого из назначенных поддиапазонов, основываясь на оценках откликов восходящего и нисходящего каналов,

и

(этап 320).The user terminal also receives a downlink MIMO pilot transmitted by the access point (block 316) and outputs a downlink response estimate,

, for each of the assigned subbands, based on the received pilot (block 318). The user terminal then determines the correction factors,

and

, for each of the assigned subbands, based on the estimates of the responses of the uplink and downlink,

and

(step 320).

To derive correction factors, it is assumed that the responses of the downstream and upstream channels for each subband are mutually inverse, with corrections for gain / phase to account for differences in the transmit / receive chains of the access point and user terminal, which are given as

, для k∈K,

, for k∈K, equation (8)

where K is the set of all data subbands. Since, during calibration, only estimates of the responses of the effective downstream and upstream channels for the assigned subbands are available, equation (8) can be rewritten as

, для k∈K',

, for k∈K ', equation (9)

может быть определен как включающий в себя только N_ut диагональных элементов

. Аналогично поправочный вектор

может быть определен как включающий в себя только N_ap диагональных элементов

.where K 'is the set of all assigned subbands. Correction vector

can be defined as including only N _ut diagonal elements

. Similarly, the correction vector

can be defined as including only N _ap diagonal elements

.

и

могут быть выведены из оценок каналов

и

различными способами, в том числе при помощи вычисления отношения матриц и вычисления с минимальной среднеквадратичной ошибкой (MMSE). Оба указанных способа вычисления более подробно описаны ниже. Также могут использоваться другие способы вычисления, и это находится в пределах объема настоящего изобретения. Correction Factors

and

can be derived from channel ratings

and

in various ways, including by calculating the ratio of the matrices and computing with the minimum mean square error (MMSE). Both of these calculation methods are described in more detail below. Other calculation methods may also be used, and this is within the scope of the present invention.

A. Calculation of the ratio of matrices

и

из оценок откликов нисходящего и восходящего каналов

и

, используя вычисление отношения матриц. Процесс 320а может быть использован в качестве этапа 320 по Фиг. 3.FIG. 4 is a flowchart of an embodiment of a process 320a for deriving correction vectors

and

from downstream and upstream channel response estimates

and

using matrix relation calculation. Process 320a may be used as step 320 of FIG. 3.

(этап 412), следующим образом:First, for each assigned subband, a (N _ut × N _ap ) matrix is computed

(step 412), as follows:

, для k∈K'

, for k∈K ' equation (10)

таким образом может быть вычислен какwhere the ratio is calculated elementwise. Every item

thus can be calculated as

, для i={1…N_ut} и j={1…N_ap}

, for i = {1 ... N _ut } and j = {1 ... N _ap } equation (11)

и

представляют собой ((i,j)-й (строка, столбец) элемент

и

, соответственно,

представляет собой (i,j)-й элемент

.Where

and

represent the ((i, j) th (row, column) element

and

accordingly

represents the (i, j) th element

.

, определяют как равный среднему нормированных строк

и выводят на этапах блока 420. Каждая строка

сначала нормируется посредством масштабирования каждого из N_ap элементов в строке на первый элемент в этой строке (этап 422). Таким образом, если

представляет собой

i-ю строку

, то нормированная строка

может быть выражена какIn one embodiment, the correction vector for the access point,

, defined as equal to the average of normalized rows

and output at block 420. Each line

first normalized by scaling each of the N _ap elements in the row by the first element in this row (step 422). So if

represents

i-th line

, then the normalized string

can be expressed as

, уравнение (12)

equation (12)

определяют как равный указанному среднему (этап 426), что может быть выражено какThen, the average value of the normalized strings is determined as the sum of N _ut normalized strings divided by N _ut (step 424). Correction vector

determined as equal to the specified average (step 426), which can be expressed as

, для k∈K',

, for k∈K ', equation (13)

является единичным.Due to normalization, the first element

is single.

, определен как равный среднему обратных значений нормированных столбцов

, и определяется на этапах блока 430. Сначала j-й столбец

нормируют путем масштабирования каждого элемента в столбце на j-й элемент вектора

, который обозначен как

(этап 432). Таким образом, если

представляет собой j-й столбец

, то нормированный столбец

может быть выражен какIn one embodiment, a correction vector for a user terminal,

, defined as equal to the average of the inverse of normalized columns

, and is determined in steps of block 430. First, the jth column

normalize by scaling each element in the column on the j-th element of the vector

which is designated as

(step 432). So if

represents the jth column

then the normalized column

can be expressed as

equation (14)

определяют как равный указанному среднему (этап 436), что может быть выражено какThen, the average of the inverse values of the normalized columns is determined as the sum of the inverse values N _{ap of the} normalized columns divided by N _ap (step 434). Correction vector

determined as equal to the specified average (step 436), which can be expressed as

, для k∈K,

, for k∈K, equation (15)

, получают на поэлементной основе.where the inverse of the normalized columns,

receive on an element-by-element basis.

B. MMSE calculation

и

выводят из оценок откликов нисходящего и восходящего каналов

и

таким образом, что среднеквадратичная ошибка (MSE) между откликом калиброванного нисходящего канала и откликом калиброванного восходящего канала является минимальной. Это условие может быть выражено какFor MMSE calculations, correction factors

and

derive from the estimates of the responses of the downward and upward channels

and

so that the mean square error (MSE) between the response of the calibrated downlink and the response of the calibrated uplink is minimal. This condition can be expressed as

, для k∈K,

, for k∈K, equation (16)

which can also be written as

, для k∈K,

, for k∈K,

, поскольку

является диагональной матрицей.Where

, insofar as

is a diagonal matrix.

определен как равный единице (т.е.

). Без такого ограничения будет получено тривиальное решение, в котором все элементы матриц

и

равны нулю. В уравнении (16) матрицу

сначала получают как

. Затем получают квадрат абсолютного значения для каждого из N_ap·N_ut элементов матрицы

. Среднеквадратичная ошибка (или квадратичная ошибка, если не производится деление на N_ap·N_ut) при этом равна сумме всех квадратов N_ap·N_ut значений.Equation (16) is constrained by the fact that the first element

defined as equal to unity (i.e.

) Without such a restriction, a trivial solution will be obtained in which all elements of the matrices

and

equal to zero. In equation (16), the matrix

first get how

. Then get the square of the absolute value for each of the N _ap · N _ut elements of the matrix

. The root-mean-square error (or quadratic error if division by N _ap · N _{ut is} not performed) is equal to the sum of all squares of N _ap · N _ut values.

и

для этого поддиапазона. MMSE вычисление для одного поддиапазона описано ниже. Для простоты индекс поддиапазона, k, в нижеследующем описании опущен. Также для простоты элементы оценки

отклика нисходящего канала обозначены

, элементы оценки

отклика восходящего канала обозначены как

, диагональные элементы матрицы

обозначены как

и диагональные элементы матрицы

обозначены как

, где i={1…N_ap} и j={1…N_ut}.MMSE calculation is performed for each assigned subband to obtain correction factors

and

for this subband. MMSE calculation for one subband is described below. For simplicity, the subband index, k, is omitted in the following description. Also for simplicity, evaluation elements

downlink response marked

, assessment items

uplink response labeled

, diagonal elements of the matrix

marked as

and diagonal matrix elements

marked as

, where i = {1 ... N _ap } and j = {1 ... N _ut }.

The root mean square error can be rewritten based on equation (16) as follows:

,

, equation (17)

. Минимальная среднеквадратичная ошибка может быть получена путем вычисления частных производных уравнения (17) по

и ν и приравнивания частных производных нулю. Результатом этих операций являются следующие наборы уравнений:and again subject to limitations

. The minimum mean square error can be obtained by calculating the partial derivatives of equation (17) with

and ν and equating partial derivatives to zero. The result of these operations are the following sets of equations:

, для i∈{2…N_ap}, и

, for i∈ {2 ... N _ap }, and equation (18a)

, for j∈ {1 ... N _ut } equation (18b)

, поэтому для этого случая частная производная отсутствует, и индекс i меняется от 2 до N_ap.In equation (18a)

, therefore, for this case, the partial derivative is absent, and the index i varies from 2 to N _ap .

The set of (N _ap + N _ut -1) equations in the sets of equations (18a) and (18b) can be expressed with great convenience in matrix form as follows:

equation (19)

Where

и

.

and

.

включает в себя (N_ap+N_ut-1) строк, причем первые N_ap-1 строк соответствуют N_ap-1 уравнений из набора уравнений (18а), и последние N_ut строк соответствуют N_ut уравнений из набора уравнений (18b). В частности, первая строка матрицы

образована из набора уравнений (18а) при i=2, вторая строка образована при i=3 и т.д. N_ap-я строка матрицы

образована из набора уравнений (18b) при j=1 и т.д. И последняя строка образована при j=N_ut. Как показано выше, элементы матрицы

и элементы вектора

могут быть получены, основываясь на элементах матриц

и

.Matrix

includes (N _ap + N _ut -1) lines, the first N _ap -1 lines corresponding to N _ap -1 equations from the set of equations (18a), and the last N _ut lines corresponding to N _ut equations from the set of equations (18b). In particular, the first row of the matrix

formed from the set of equations (18a) at i = 2, the second line is formed at i = 3, etc. N _ap -th row of the matrix

formed from the set of equations (18b) with j = 1, etc. And the last line is formed with j = N _ut . As shown above, matrix elements

and vector elements

can be obtained based on matrix elements

and

.

, который может быть получен какCorrection factors included in the vector.

which can be obtained as

equation (20)

и

, которые минимизируют среднеквадратичную ошибку откликов калиброванных нисходящего и восходящего каналов, как показано в уравнении (16). Поскольку матрицы

и

получены, основываясь на оценках откликов нисходящего и восходящего каналов,

и

, качества поправочных матриц

и

таким образом зависит от качества оценок каналов

и

. MIMO пилот-сигнал может быть усреднен в приемнике для получения более точных оценок для

и

.The result of MMSE calculation is a correction matrix

and

that minimize the standard error of the responses of the calibrated downstream and upstream channels, as shown in equation (16). Since matrices

and

obtained based on the estimates of the responses of the downward and upward channels,

and

, quality correction matrices

and

thus depends on the quality of the channel estimates

and

. MIMO pilot can be averaged at the receiver to obtain more accurate estimates for

and

.

и

, полученные основываясь на MMSE вычислении, являются в общем случае лучшими, чем поправочные матрицы, полученные, основываясь на вычислении отношения матриц, особенно когда некоторые из усилений каналов являются малыми, и измеренный шум может привести к сильной деградации усилений каналов.Correction Matrices

and

obtained based on the MMSE calculation are generally better than correction matrices obtained based on the calculation of the matrix ratio, especially when some of the channel gains are small and the measured noise can lead to severe degradation of the channel gains.

C. Additional calculations

, для всех назначенных поддиапазонов. Если для каждого поправочного фактора в

используются 12-битные комплексные значения, тогда поправочные вектора

для всех назначенных поддиапазонов могут быть отправлены в точку доступа в 3·(N_ap-1)N_sb байтах, где “3” возникает вследствие того, что для I и Q компонентов в сумме используется 24 бита и (N_ap-1) является результатом того, что первый элемент в каждом векторе

равен единице и, следовательно, его не требуется передавать. Если первому элементу присвоено значение 2⁹-1=+511, то при этом доступен диапазон 12 дБ (поскольку максимальная положительная 12-битная величина со знаком представляет собой 12¹¹-1=+2047), что дает возможность, пользуясь 12-битными значениями, регулировать несоответствия до 12 дБ в усилениях между нисходящей линией и восходящей линией. Если нисходящая линия и восходящая линия соответствуют друг другу в пределах 12 дБ, и первый элемент нормирован на величину 511, тогда другие элементы не должны превышать 511·4=2044 по абсолютной величине, и могут быть представлены при помощи 12 битов.Regardless of the specific calculation method selected for use, after completion of the calculation of the correction matrices, the user terminal sends correction vectors to the access point for the access point,

, for all assigned subbands. If for each correction factor in

12-bit complex values are used, then correction vectors

for all assigned subbands, they can be sent to the access point in 3 · (N _ap -1) N _sb bytes, where “3” arises due to the fact that 24 bits are used for I and Q components in total and (N _ap -1) is the result of the first element in each vector

equal to one and, therefore, it does not need to be transferred. If the first element is assigned the value 2 ⁹ -1 = + 511, then the range is 12 dB (since the maximum positive 12-bit value with a sign is 12 ¹¹ -1 = + 2047), which makes it possible, using 12-bit values , adjust the mismatches to 12 dB in amplifications between the downlink and the uplink. If the descending line and the ascending line correspond to each other within 12 dB, and the first element is normalized to 511, then the other elements should not exceed 511 · 4 = 2044 in absolute value, and can be represented using 12 bits.

и

получают для каждого назначенного поддиапазона. Если калибровку выполняют не для всех поддиапазонов данных, тогда поправочные факторы для “некалиброванных” поддиапазонов могут быть получены интерполяцией поправочных факторов, полученных для назначенных поддиапазонов. Интерполяция может быть выполнена в точке доступа для получения поправочных векторов

, для k∈K. Аналогично, интерполяция может быть выполнена в пользовательском терминале для получения поправочных векторов

для k∈K.A couple of correction vectors

and

receive for each assigned subband. If calibration is not performed for all data subbands, then the correction factors for the “uncalibrated” subbands can be obtained by interpolating the correction factors obtained for the assigned subbands. Interpolation can be performed at the access point to obtain correction vectors

, for k∈K. Similarly, interpolation can be performed in a user terminal to obtain correction vectors

for k∈K.

и

или соответствующие поправочные матрицы

и

, для k∈K, для масштабирования символов модуляции перед передачей по беспроводному каналу, как описано ниже. При этом эффективный нисходящий канал, видимый со стороны пользовательского терминала, представляет собой

.Subsequently, the access point and user terminal use their respective correction vectors

and

or corresponding correction matrices

and

, for k∈K, to scale the modulation symbols before transmitting over the wireless channel, as described below. In this case, the effective downward channel visible from the user terminal is

.

The calibration scheme described above, by which a vector of correction factors is obtained for both the access point and the user terminal, makes it possible to derive “compatible” correction vectors for the access point when calibration is performed by various user terminals. If the calibration at the access point has already been performed (for example, by one or more user terminals), then the current correction vectors can be updated using the newly derived correction vectors.

For example, if two user terminals are simultaneously performing a calibration procedure, then the calibration results from these user terminals can be averaged to improve performance. However, as a rule, calibration is performed for one user terminal at a time. Thus, the second user terminal sees a downlink for which the correction vector for the first user terminal has already been used. In this case, the product of the second correction vector and the old correction vector can be used as a new correction vector, or “weighted averaging” (described below) can also be used. Typically, an access point uses one correction vector for all user terminals, rather than different correction vectors for different user terminals (although this option can also be implemented). Updates from multiple user terminals or consecutive updates from a single user terminal can be handled in the same way. updated vectors can be applied directly (using the multiplication operation). Alternatively, if some averaging is required to reduce the measurement noise, a weighted averaging can be used, as described below.

для передачи MIMO пилот-сигнала, из которого пользовательский терминал определяет новые поправочные вектора

, то обновленные поправочные вектора

представляют собой результат умножения текущего и нового поправочных векторов. Поправочные вектора

и

могут быть выведены в одном или в разных пользовательских терминалах.Thus, if the access point uses correction vectors

for transmitting a MIMO pilot from which the user terminal determines new correction vectors

then updated correction vectors

represent the result of multiplying the current and new correction vectors. Correction vectors

and

can be displayed in one or in different user terminals.

, причем умножение выполняется поэлементно. В другом варианте осуществления обновленные поправочные вектора могут быть переопределены как

, где α представляет собой фактор, используемый для обеспечения взвешенного усреднения (т.е. 0<α<1). Если обновление калибровки происходит редко, то тогда лучше работает α со значением, близким к единице. Если обновления калибровки являются частыми, но зашумленными, то предпочтительными являются меньшие значения α. Затем обновленные поправочные вектора

могут быть использованы точкой доступа до их следующего обновления.In one embodiment, the updated correction vectors are defined as

, and multiplication is performed element by element. In another embodiment, the updated correction vectors may be redefined as

where α is the factor used to provide weighted averaging (i.e., 0 <α <1). If calibration updates are rare, then α works better with a value close to one. If calibration updates are frequent but noisy, lower α values are preferred. Then updated correction vectors

can be used by the access point until their next update.

As indicated above, calibration may not be performed for all data subbands. For example, calibration can be performed for each n-th subband, where n can be determined from the expected response of the transmit / receive chains (for example, n may be 2, 4, 8, 16, etc.). Calibration can also be performed for unevenly distributed subbands. For example, since at the borders of the frequency range, the filter characteristic may have a larger drop, which may create more mismatch in the transmit / receive chains, then more sub-bands can be calibrated at the borders of the frequency band. In general, any number of subbands distributed in any way can be calibrated, and this is within the scope of the present invention.

и

для k∈K' выводятся пользовательским терминалом, и вектора

отправляются в точку доступа. Эта схема преимущественно распределяет обработку калибровки между пользовательскими терминалами в случае системы с множественным доступом. Однако поправочные вектора

и

также могут быть выведены в точке доступа, которая затем отправляет вектора

в пользовательский терминал, и это находится в пределах объема настоящего изобретения.In the above description, correction vectors

and

for k∈K 'are displayed by the user terminal, and the vectors

sent to the access point. This scheme advantageously distributes the calibration processing between user terminals in the case of a multiple access system. However, the correction vectors

and

can also be output at an access point that then sends vectors

to a user terminal, and this is within the scope of the present invention.

The calibration scheme described above allows each user terminal to calibrate its transmit / receive circuits in real time during transmission over the air. This allows user terminals with different frequency responses to provide high performance without stringent requirements for frequency response or calibration during manufacturing. The access point can be calibrated by a variety of user terminals to provide improved accuracy.

D. Reinforcement

Calibration can be performed based on the normalized gains for the downstream and upstream channels, which are gains relative to the noise level at the receiver. After calibrating the downlink and uplink, the use of normalized gains allows you to obtain the characteristics of one line (including channel gains and SNR for each eigenmode), based on measurements of gains for the other line.

The access point and the user terminal can first balance the input levels of their receivers so that the noise levels in the receive circuits of the access point and the user terminal are approximately the same. Balancing can be done by estimating the noise level, that is, determining a section of a received TDD frame (i.e., a downlink / uplink transmission unit) that has a minimum average power over a specific time period (for example, one or two symbol periods). In the general case, the time interval immediately before the start of each TDD frame is transmission free, since any uplink data must be received by the access point, and then a transmit / receive switch must take some time before the access point begins to transmit on the downlink. Depending on the interference environment, the noise level can be determined based on several TDD frames. Then measure the responses of the downward and upward channel relative to this noise level. More specifically, the channel gain for a given subband of a given pair of transmit / receive antennas can be obtained, for example, as the ratio of the received pilot symbols to transmitted pilot symbols for this subband of a given pair of transmit / receive antennas. In this case, the normalized gain is the measured gain divided by the noise level.

установлен в 1.The big difference between the normalized amplifications for the access point and the normalized amplifications for the user terminal can lead to the fact that the correction factors for the user terminal can be very different from unity. Correction factors for the access point are close to unity, since the first element of the matrix

set to 1.

If the correction factors for the user terminal are very different from unity, then the user terminal may not be able to use the calculated correction factors. This may be due to the fact that the user terminal has a limit on its maximum transmit power and may not be able to increase the transmit power for large correction factors. In addition, a decrease in transmit power for small correction factors is generally not desirable, as this can reduce the data rate.

Thus, the user terminal can transmit using a scaled version of the calculated correction factors. Scaled calibration factors can be obtained by scaling the calculated correction factors to a specific scale, which can be set equal to the difference in gains (in the form of a difference or ratio) between the responses of the downstream and upstream channels. Such a difference in gains can be calculated as the average of the differences (or differences) between the normalized gains for the descending line and the ascending line. The scale (or gain difference) used for the correction factors in the user terminal can be sent to the access point along with the calculated correction factors for the access point.

If there are correction factors and the scale or difference in amplifications, the characteristics of the downlink can be determined from the measured response of the uplink and vice versa. If the noise level at either the access point or the user terminal changes, then the gain difference can be updated, and the updated gain difference can be sent in the message of another entity.

In the above description, calibration led to two sets (vectors or matrices) of correction factors for each subband, with one set being used at the access point for downlink data, and the other set being used at the user terminal for uplink data. Calibration can also be performed in such a way that two sets of correction factors are provided for each subband, with one set being used at the access point to receive data on the uplink, and the second set is used at the user terminal for receiving data on the downlink. Calibration can also be performed in such a way that for each subband one set of correction factors is obtained, and this set can be used either at the access point or in the user terminal. In general, calibration is performed in such a way that the responses of the calibrated downstream and upstream channels are mutually inverse, regardless of where the correction factors are applied.

2. MIMO pilot

For calibration, the MIMO pilot is transmitted by the user terminal in an uplink to enable the access point to evaluate the uplink response, and the MIMO pilot is transmitted by the access point in a downlink to allow the user terminal to evaluate the downlink. For the downlink and the uplink, the same or different MIMO pilots can be used, and the used MIMO pilots are known at both the access point and the user terminal.

In one embodiment, the MIMO pilot contains an OFDM symbol (denoted by “P”), which is transmitted through each of the N _T transmit antennas, where N _T = N _ap for the downlink and N _T = N _ut for the uplink. For each transmit antenna, the same OFDM symbol P is transmitted in each symbol period assigned to transmit the MIMO pilot. However, OFDM symbols P for each antenna are covered by different Walsh sequences with N chips assigned to this antenna, where N≥N _ap for the downlink and N≥N _ut for the uplink. Walsh coverage maintains orthogonality between N _T transmit antennas and allows the receiver to distinguish between individual transmit antennas.

The OFDM symbol P includes one modulation symbol for each of the N _sb assigned subbands. The OFDM symbol P thus contains a certain “word” of N _sb modulation symbols that can be selected to facilitate channel estimation by the receiver. This word can also be defined to minimize peak-to-average ratios when transmitting a MIMO pilot. This reduces the amount of distortion and non-linearity generated by the transmit / receive circuits, which in turn leads to improved channel estimation accuracy.

For clarity, a specific MIMO pilot is described below for a specific MIMO-OFDM system. For this system, both the access point and the user terminal are equipped with four transmit / receive antennas. The system frequency band is divided into 64 orthogonal subbands (i.e., N _F = 64), which are assigned indices from +31 to -32. Of these 64 sub-bands, 48 sub-bands (for example, with indices ± {1, ..., 6, 8, ..., 20, 22, ..., 26}) are used for data, 4 sub-bands (for example, ± {7, 21}) are used for of the pilot signal and possibly for signaling, the DC subband (with index 0) is not used, and the remaining subbands are also not used and serve as guard subbands. This OFDM subband structure is described in more detail in IEEE 802.11a, entitled “Part 11: Wireless LAN Medium Access Control (MAC) and Physical Layer (PHY) specifications: High-speed Physical Layer in the 5 GHz Band,” September 1999 , which is publicly available and incorporated herein by reference in its entirety.

The OFDM symbol P includes a set of 52 QPSK modulation symbols for 48 data subbands and 4 pilot subbands. The indicated OFDM symbol P may have the following form:

P (real) = g · {0,0,0,0,0,0,0, -1, -1, -1, -1,1,1,1, -1, -1,1, -1,1 , 1,1,1, -1, -1,1, -1,1, -1, -1, -1, -1,1, -1,

0,1, -1, -1, -1, -1,1, -1, -1, -1, -1,1,1, -1, -1,1, -1, -1,1, 1, -1,1, -1,1, -1,1, -1,0,0,0,0,0,0},

P (imaginary) = g · {0,0,0,0,0,0,0, -1,1,1,1, -1, -1,1, -1,1,1,1, -1,1 , -1, -1, -1, -1, -1, -1,1,1, -1,1,1, -1,1,

0, -1, -1, -1, -1,1,1, -1,1, -1, -1,1, -1,1, -1,1,1,1, -1,1, 1,1,1,1,1,1 -1 -1,0,0,0,0,0,0}

where g is the gain for the pilot signal. The values inside the brackets {} are given for subbands with indices from -32 to -1 (for the first line) and from 0 to +31 (for the second line). Thus, the first line for P (real) and P (imaginary) indicates that the character (-1-j) is transmitted in the subband -26, the character (-1 + j) is transmitted in the subband -25, etc. The second line for P (real) and P (imaginary) indicates that the character (1-j) is transmitted in subband 1, the character (-1-j) is transmitted in subband 2, etc. Other OFDM symbols may also be used for the MIMO pilot.

In one embodiment, the four transmit antennas for the MIMO pilot are assigned Walsh sequences W ₁ = 1111, W ₂ = 1010, W ₃ = 1100 W ₄ = 1001. For this Walsh sequence, a value of “1” indicates that an OFDM symbol P is being transmitted, and a value of “0” indicates that an OFDM symbol -P is being transmitted (i.e., each of the 52 modulation symbols in P is inverted).

Table 1 lists the OFDM symbols transmitted through each of the four transmit antennas when transmitting a MIMO pilot with a duration of 4 symbol periods.

Table 1 OFDM Symbol Antenna 1 Antenna 2 Antenna 3 Antenna 4 one + P + P + P + P 2 + P -R + P -R 3 + P + P -R -R four + P -R -R + P

For longer MIMO transmissions of the pilot, the Walsh sequences for each transmit antenna are repeated. For such a set of Walsh sequences, pilot MIMO transmission is performed over a number of symbol periods a multiple of 4 symbol periods to ensure orthogonality between the four transmit antennas.

для k=±{1, …, 26}, что представляет собой оценку отклика эффективного канала от передающей антенны i до приемной антенны j (т.е. включающего в себя отклики передающих/приемных цепей) для 52 поддиапазонов данных и пилот-сигналов.The receiver may derive a channel response estimate based on the received MIMO pilot by performing complementary processing. In particular, to reconstruct the pilot signal transmitted through the transmitting antenna i and received by the receiving antenna j, the pilot signal received by the receiving antenna j is first processed with the Walsh sequence assigned to the transmitting antenna i in a manner complementary to the Walsh coating performed in the transmitter. The uncoated OFDM symbols for all N _ps symbol periods for the MIMO pilot are then summed, the summation being performed individually for each of the 52 subbands used to transmit the MIMO pilot. The result of the summation is

for k = ± {1, ..., 26}, which is an estimate of the response of the effective channel from the transmitting antenna i to the receiving antenna j (that is, including the responses of the transmitting / receiving circuits) for 52 data subbands and pilot signals.

и

, для каждого из 52 поддиапазонов.The same processing can be performed to reconstruct the pilot signal from each transmit antenna to each receive antenna. The pilot processing provides N _ap · N _ut values, which are elements of the effective channel response estimate,

and

, for each of the 52 subbands.

, и оценки отклика эффективного нисходящего канала,

, соответственно, которая затем используется для вывода поправочных факторов, как описано выше.The channel estimation described above can be performed by both the access point and the user terminal during calibration to obtain an estimate of the response of the effective uplink channel,

, and estimates of the effective downlink response,

, respectively, which is then used to derive correction factors, as described above.

3. Spatial processing

To simplify channel estimation and spatial processing at the access point and user terminal for TDD MIMO and MIMO-OFDM systems, a correlation between downlink and uplink responses can be used. This simplification is possible after performing calibration to account for differences in the transmit / receive chains. As indicated above, the responses of calibrated channels are:

,
для нисходящей линии и

,
for the downlink and equation (21a)

,
for the ascending line. equation (21b)

.The last equality in equation (21b) appears due to the relationship between the responses of the effective downward and upward channels,

.

отклика канала для каждого поддиапазона может быть “диагонализирована” для получения N_S собственных мод для этого поддиапазона. Это может быть достигнуто либо при помощи разложения по сингулярным значениям матрицы

отклика канала, либо разложения по собственным векторам корреляционной матрицы для

, которая представляет собой

.Matrix

The channel response for each subband can be “diagonalized” to obtain N _S eigenmodes for that subband. This can be achieved either by expanding the singular values of the matrix

channel response, or expansion in eigenvectors of the correlation matrix for

which is

.

может быть выражено какSingular value decomposition of the response matrix of a calibrated uplink

may be expressed as

, для k∈K,

, for k∈K, equation (22)

представляет собой (N_ut×N_ut) унитарную матрицу левых собственных векторов для

;Where

is a (N _ut × N _ut ) unitary matrix of left eigenvectors for

;

представляет собой (N_ut×N_ap) диагональную матрицу сингулярных значений для

; и

represents (N _ut × N _ap ) the diagonal matrix of singular values for

; and

представляет собой (N_ар×N_ар) унитарную матрицу правых собственных векторов для

;

is a (N _ar × N _ar ) unitary matrix of right eigenvectors for

;

, где

представляет собой единичную матрицу. Соответственно, разложение по сингулярным значениям матрицы откликов калиброванного нисходящего канала,

, может быть выражено какThe unitary matrix M is characterized by the property

where

is a unit matrix. Accordingly, the expansion in the singular values of the response matrix of the calibrated downlink,

may be expressed as

, для k∈K

, for k∈K equation (23)

и

также являются матрицами левых и правых собственных векторов, соответственно, для

. Матрицы

,

,

и

представляют собой различные формы матрицы

, и матрицы

,

,

и

также являются различными формами матрицы

. Для простоты ссылки на матрицы

и

в нижеследующем описании также могут представлять собой ссылки на их различные формы. Матрицы

и

используются в точке доступа и пользовательском терминале, соответственно, для пространственной обработки, и определены как таковые их нижними индексами.So matrices

and

are also matrices of left and right eigenvectors, respectively, for

. Matrices

,

,

and

represent various forms of matrix

, and matrices

,

,

and

are also different forms of matrix

. For simplicity, matrix references

and

in the following description may also be references to their various forms. Matrices

and

are used at the access point and user terminal, respectively, for spatial processing, and are defined as such by their subscripts.

Singular value decomposition is described in more detail in Gilbert Strang, entitled “Linear Algebra and Its Applications”, second edition, Academic Press, 1980.

, для k∈K, для получения диагональных матриц

и матриц

левых собственных векторов для

. Такое разложение по сингулярным значениям может быть описано, как

, где знак (“ˆ”) над каждой матрицей указывает, что она является оценкой реальной матрицы.The user terminal may evaluate the response of the calibrated downlink based on the MIMO pilot transmitted by the access point. After that, the user terminal can perform a decomposition into the singular values of the response estimate of the calibrated downlink,

, for k∈K, to obtain diagonal matrices

and matrices

left eigenvectors for

. Such a decomposition in singular values can be described as

, where the sign (“ˆ") above each matrix indicates that it is an estimate of the real matrix.

, для k∈K, для получения диагональных матриц

и матриц

левых собственных векторов для

, для k∈K. Такое разложение по сингулярным значениям может быть описано, как

.Similarly, the access point can perform a calibrated uplink response estimate based on the MIMO pilot transmitted by the user terminal. Then, the access point can perform a decomposition into the singular values of the response estimate of the calibrated uplink,

, for k∈K, to obtain diagonal matrices

and matrices

left eigenvectors for

, for k∈K. Such a decomposition in singular values can be described as

.

, так и матриц

. В случае выполнения в пользовательском терминале матрицы

используются для пространственной обработки в пользовательском терминале, а матрицы

могут быть переданы в точку доступа.Due to the fact that the channel and calibration are mutually inverse, decomposition into singular values can be performed either only in the user terminal, or only at the access point to receive as matrices

and matrices

. If the matrix is executed in the user terminal

are used for spatial processing in the user terminal, and matrices

can be transferred to an access point.

и

, основываясь на направленном опорном сигнале, передаваемом пользовательским терминалом. Аналогично, пользовательский терминал также может иметь возможность получить матрицы

и

, основываясь на направленном опорном сигнале, передаваемом точкой доступа. Направленный опорный сигнал подробно описан в вышеупомянутой предварительной заявке на патент США № 60/421309.An access point may also be able to obtain matrices.

and

based on the directional reference signal transmitted by the user terminal. Similarly, a user terminal may also be able to obtain matrices

and

based on the directional reference signal transmitted by the access point. The directional reference signal is described in detail in the aforementioned provisional application for US patent No. 60/421309.

и

могут быть использованы для передачи независимых потоков данных по N_S собственным модам MIMO канала, где N_S≤min{N_ap,N_ut}. Пространственная обработка для передачи множества потоков данных по нисходящей линии и восходящей линии описана ниже.Matrices

and

can be used to transmit independent data streams over N _S eigenmodes of the MIMO channel, where N _S ≤min {N _ap , N _ut }. Spatial processing for transmitting multiple data streams in the downlink and uplink is described below.

A. Spatial processing for the uplink

Spatial processing in the user terminal for uplink transmission can be expressed as

, для k∈K,

, for k∈K, equation (24)

представляет собой вектор передачи для восходящей линии для k-го поддиапазона; иWhere

represents the transmission vector for the uplink for the k-th subband; and

представляет собой вектор “данных” с ненулевыми элементами, количеством до N_S, для символов модуляции, предназначенных для передачи по N_S собственным модам k-го поддиапазона.

represents a vector of “data” with nonzero elements, up to N _S , for modulation symbols intended for transmission on N _S eigenmodes of the k-th subband.

Also, before transmission, additional processing of modulation symbols may be performed. For example, for data subbands (for example, for each eigenmode), channel inversion can be applied such that the reception SNR is approximately the same for all data subbands. In this case, the spatial processing can be expressed as

, для k∈K,

, for k∈K, equation (25)

представляет собой матрицу весов для (необязательной) инверсии восходящего канала.Where

represents a matrix of weights for the (optional) inversion of the uplink.

включает в себя коэффициенты инверсии канала, и матрица

в уравнении (25) может быть опущена. В нижеследующем описании использование матрицы

указывает на то, что коэффициенты инверсии канала не включены в вектор

. Отсутствие матрицы

в уравнении может указывать на то, что (1) либо инверсия канала не производится, (2) либо инверсия канала производится и учтена в векторе

.Channel inversion can also be performed by assigning transmit power to each subband before performing the modulation, in which case the vector

includes channel inversion coefficients, and matrix

in equation (25) can be omitted. In the following description, the use of the matrix

indicates that channel inversion coefficients are not included in the vector

. Lack of matrix

in the equation may indicate that (1) either the channel inversion is not performed, (2) either the channel inversion is performed and taken into account in the vector

.

Channel inversion can be performed as described in the aforementioned provisional application for US patent No. 60/421309, and in patent application US No. 10/229209, entitled "Coded MIMO Systems with Selective Channel Inversion Applied Per Eigenmode", filed August 27, 2002 , the rights to which belong to the copyright holder of this patent application and which is incorporated into this description in its entirety by reference.

The received uplink transmission at the access point may be expressed as

, для k∈K,

, for k∈K, equation (26)

представляет собой принятый вектор для восходящей линии для k-го поддиапазона;Where

represents the received vector for the ascending line for the k-th subband;

представляет собой аддитивный белый Гауссовский шум (AWGN) для k-го поддиапазона; и

represents additive white Gaussian noise (AWGN) for the k-th subband; and

дается уравнением (24).

is given by equation (24).

The spatial processing (or matched filtering) at the access point for a received uplink transmission can be expressed as

, для k∈K,

, for k∈K,

equation (27)

представляет собой оценку вектора

, переданного пользовательским терминалом по восходящей линии, и

представляет собой шум после обработки. В уравнении (27) предполагается, что инверсия канала не выполняется в передатчике и что принятый вектор

имеет вид, представленный уравнением (26).Where

is an estimate of the vector

transmitted by the user terminal in an uplink, and

represents the noise after processing. Equation (27) assumes that channel inversion is not performed at the transmitter and that the received vector

has the form represented by equation (26).

B. Spatial downlink processing

Spatial processing at the access point for downlink transmission can be represented as

, для k∈K,

, for k∈K, equation (28)

представляет собой вектор передачи, и

представляет собой вектор данных для нисходящей линии. Where

represents the transmission vector, and

is a data vector for a downlink.

Again, additional processing (e.g., channel inversion) of modulation symbols may be performed before transmission. In this case, the spatial processing can be expressed as

, для k∈K,

, for k∈K, equation (29)

представляет собой матрицу весов для (необязательной) инверсии нисходящего канала.Where

is a weight matrix for the (optional) downlink inversion.

The received downlink transmission in the user terminal may be expressed as

, для k∈K,

, for k∈K,

equation (30)

представляет собой вектор передачи, представленный уравнением (28).Where

represents the transmission vector represented by equation (28).

Spatial processing (or matched filtering) in a user terminal for a received downlink transmission can be expressed as

, для k∈K

, for k∈K

equation (31)

имеет вид, представленный уравнением (30).Equation (31) assumes that channel inversion is not performed at the transmitter and that the received vector

has the form represented by equation (30).

выполняется в передатчике. Однако если такая дополнительная обработка не выполняется, то

может рассматриваться в качестве единичной матрицы.Table 2 shows the spatial processing at the access point and user terminal for transmitting and receiving data. Table 2 assumes that additional processing

performed in the transmitter. However, if such additional processing is not performed, then

can be considered as a unit matrix.

table 2 Ascending line Descending line User terminal

Broadcast:

Reception:

Access point Reception:

Broadcast:

и

используются для пространственной обработки при передаче в точке доступа и пользовательском терминале, соответственно. Это может упростить общую пространственную обработку, поскольку в любом случае (например, для инверсии канала) может требоваться масштабирование символов модуляции, и поправочные матрицы

и

могут комбинироваться с весовыми матрицами

и

для получения матриц

и

усиления, где

и

. Обработка также может выполняться таким образом, что поправочные матрицы используются для пространственной обработки при приеме (вместо пространственной обработки при передаче).In the above description and as shown in table 2, correction matrices

and

are used for spatial processing during transmission at the access point and user terminal, respectively. This can simplify the overall spatial processing, since in any case (for example, for channel inversion) scaling of modulation symbols and correction matrices may be required

and

can be combined with weight matrices

and

for matrixes

and

gain where

and

. Processing can also be performed in such a way that the correction matrices are used for spatial processing at reception (instead of spatial processing at transmission).

4. MIMO-OFDM system

5 is a block diagram of an embodiment of an access point 502 and a user terminal 504 in a MIMO-OFDM TDD system. For simplicity, the following description assumes that the access point and user terminal are equipped with four transmit / receive antennas.

In the case of a downlink at access point 502, a transmitting (TX) data processor 510 receives traffic data (i.e., information bits) from a data source 508 and signaling and other information from a controller 530. TX data processor 510 formats, codes, interleaves, and interleaves and modulation (i.e., symbol mapping) of the data to provide a modulation symbol stream for each eigenmode used to transmit the data. TX spatial processor 520 receives modulation symbol streams from TX data processor 510 and performs spatial processing to provide four transmit symbol streams, one stream for each antenna. TX spatial processor 520 also performs additional pilot symbol multiplexing as required (eg, for calibration).

Each modulator (MOD) 522 receives and processes a respective transmit symbol stream to provide a corresponding OFDM symbol stream. Each OFDM symbol stream is further processed in a transmit chain in modulator 522 to provide a corresponding downlink modulated signal. Then, four modulated signals from modulator 522a through 522d are transmitted through four antennas 524a through 524d, respectively.

, основываясь на MIMO пилот-сигнале, передаваемом точкой доступа.At user terminal 504, antennas 522 receive transmitted downlink modulated signals and each antenna provides a received signal to a corresponding demodulator (DEMOD) 554. Each demodulator 554 (which includes a receive chain) performs processing complementary to that performed in modulator 522 and provides received symbols . The receive (RX) spatial processor 560 then performs spatial processing from all demodulators 554 to provide recovered symbols, which are estimates of the modulation symbols transmitted by the access point. During RX calibration, the spatial processor 560 provides an estimate of the calibrated downlink,

based on the MIMO pilot transmitted by the access point.

, который выведен в точке доступа и передан по нисходящей линии.The RX data processor 570 processes (e.g., performs symbol demapping, deinterleaving, and decoding) the recovered symbols to provide decoded data. The decoded data may include recovered traffic data, signaling, etc. and which are provided to the consumer 572 data for storage and / or to the controller 580 for further processing. During RX calibration, the data processor 570 provides an estimate of the calibrated uplink,

that is displayed at the access point and transmitted in a downlink.

и

откликов каналов, выводить корреляционные матрицы

и

, предоставлять матрицы

в TX пространственный процессор 592 для передачи по восходящей линии и предоставлять матрицы

в TX процессор 590 данных для передачи в точку доступа. Запоминающие устройства 532 и 582 хранят данные и коды программ, используемые контроллерами 530 и 580, соответственно.Controllers 530 and 580 control the operation of various processing units at the access point and user terminal, respectively. During calibration, controller 580 may receive estimates

and

channel responses, output correlation matrices

and

provide matrices

in TX spatial processor 592 for uplink transmission and provide matrices

in TX data processor 590 for transmission to the access point. Storage devices 532 and 582 store data and program codes used by controllers 530 and 580, respectively.

The processing in the case of an uplink may be the same as the processing in the case of a downlink or different from it. Data and signaling are processed (eg, coding, interleaving and modulation) in TX data processor 590, followed by spatial processing in TX spatial processor 592, which also performs additional pilot symbol multiplexing during calibration. Pilot and modulation symbols are further processed in modulators 554 to generate modulated uplink signals, which are then transmitted via antennas 552 to the access point.

калиброванного восходящего канала, основываясь на MIMO пилот-сигнале, передаваемом пользовательским терминалом. Матрицы

принимаются контроллером 530 и затем предоставляются в TX процессор 510 данных для передачи в пользовательский терминал.At access point 110, the uplink modulated signals are received by antennas 524, demodulated in demodulators 522, and processed in the RX spatial processor 540 and RX data processor 542 in a manner complementary to that performed at the user terminal. During RX calibration, the spatial processor 560 also provides an estimate

calibrated uplink based on the MIMO pilot transmitted by the user terminal. Matrices

are received by the controller 530 and then provided to the TX data processor 510 for transmission to the user terminal.

6 is a block diagram of a TX spatial processor 520a that can be used as TX spatial processors 520 and 592 of FIG. 5. For simplicity, the following description assumes that all four eigenmodes have been selected for use.

In processor 520a, demultiplexer 632 receives four modulation symbol streams (denoted s ₁ (n) -s ₄ (n)) for transmission on four eigenmodes, demultiplexes each stream into N _D subflows for N _D data subbands, and provides four modulation symbol substreams for each subband to a corresponding TX subband spatial processor 640. Each processor 640 performs the processing described by equation (24), (25), (28) or (29) for one subband.

, где

или

. В случае восходящей линии усиление представляет собой диагональные элементы матрицы

, где

или

. Каждый умножитель 642 выполняет масштабирование его символов модуляции на соответствующее усиление g_m(k) для предоставления масштабированных символов модуляции. Умножители 642а-642d предоставляют четыре потока масштабированных символов модуляции в четыре формирователя 650а-650d лучей, соответственно.In each TX subband spatial processor 640, four sub-streams (denoted by s ₁ (k) -s ₄ (k)) of modulation symbols are provided in four multipliers 642a-642d, which also receive g ₁ (k), g ₂ (k), g ₃ (k) and g ₄ (k) for four eigenmodes of the corresponding subband. In the case of a downlink, the four gains for each data subband are diagonal elements of the corresponding matrix

where

or

. In the case of an ascending line, the gain is the diagonal elements of the matrix

where

or

. Each multiplier 642 scales its modulation symbols with a corresponding gain g _m (k) to provide scaled modulation symbols. Multipliers 642a through 642d provide four streams of scaled modulation symbols to four beamformers 650a through 650d, respectively.

для соответствующей собственной моды. В каждом формирователе 650 пучка масштабированные символы модуляции предоставляются в четыре умножителя 652а-652d, которые также принимают четыре элемента

,

,

и

, собственного вектора

для соответствующей собственной моды. Собственный вектор

представляет собой m-й столбец матрицы

для нисходящей линии, и представляет собой m-й столбец матрицы

для восходящей линии. Каждый умножитель 652 затем выполняет умножение масштабированных символов модуляции на соответствующее значение

собственного вектора для предоставления “обработанных для формирования луча” символов. Умножители 652a-652d предоставляют четыре подпотока обработанных для формирования луча символов (которые предназначены для передачи через четыре антенны) в сумматоры 660a-660d, соответственно.Each beamformer 650 performs beamforming to transmit one symbol substream in one eigenmode of one subband. Each beam former 650 receives one substream s _m (k) of scaled symbols and performs beam formation using its own vector

for the corresponding own fashion. In each beam former 650, scaled modulation symbols are provided in four multipliers 652a through 652d, which also accept four elements

,

,

and

, eigenvector

for the corresponding own fashion. Eigenvector

represents the mth column of the matrix

for the descending line, and is the mth column of the matrix

for the ascending line. Each multiplier 652 then multiplies the scaled modulation symbols by a corresponding value

eigenvector to provide “beam-processed” characters. Multipliers 652a through 652d provide four substreams of beam-processed symbols (which are intended to be transmitted through four antennas) to adders 660a through 660d, respectively.

Each adder 660 receives and summarizes four beamformed symbols for four eigenmodes for each symbol period to provide preprocessed symbol for a respective transmit antenna. Adders 660a through 660d provide four pre-processed symbol substreams for the four transmit antennas to buffers / multiplexers 670a through 670d, respectively.

Each buffer / multiplexer 670 receives pilot symbols and preprocessed symbols from TX subband spatial processors 640 for N _D data subbands. Each buffer / multiplexer 670 then multiplexes the pilot symbols, preprocessed symbols and zeros for the pilot subbands, data subbands and unused subbands, respectively, to form a sequence of N _F transmission symbols for a given symbol period. During calibration, pilot symbols are transmitted over the assigned subbands. Multipliers 668a-668d cover the pilot symbols for four antennas with Walsh sequences W ₁ -W ₄ respectively assigned to four antennas, as described above and shown in Table 1. Each buffer / multiplexer 670 provides a stream of transmit symbols x _i (n) for one a transmit antenna, wherein the transmit symbol stream comprises series-connected sequences of N _F transmit symbols.

Spatial processing and OFDM modulation are described in more detail in the aforementioned provisional application for US patent No. 60/421309.

In various embodiments of the present invention set forth herein, peer-to-peer communication between different user terminals (UT or STA) within one basic service area (BSS) or different BSSs can be implemented, as described below. UTs or STAs that calibrate with a single access point (AP) are members of a Basic Service Area (BSS). One access point is a common node for all UTs in the BSS. The calibration methods described above facilitate the following types of communication:

(i) The UT in the BSS can use directional transmission (TX) for direct communication with the uplink (UL) AP, and the AP can use directional transmission (TX) for communication with the UT in downlink (DL).

(ii) A UT in a BSS can directly communicate with another UT in the same BSS using directional communication. In this case, such a peer-to-peer communication should be pre-initiated, since none of the UT has channel information between them. In various embodiments, the preliminary initiation procedure works as follows:

- the peer initiator is an AP source (DAP) and the other UT is a UT receiver (DUT).

- The DAP sends a MIMO pilot to the DUT along with a link setup request that contains the BSS ID and DAP ID. The request should be sent in general mode (i.e., diversity in transmission).

- The DUT responds by sending a pilot MIMO pilot and an acknowledgment that contains the DUT ID, its BSS ID and some indicator of the transmission rate for use in the DAP.

- Then the DAP can use directional transmission over DL and the DUT can use directional transmission over UL. Speed control and tracking can be done by splitting the gears into DL and UL segments with sufficient time intervals between them to complete the processing.

(iii) UTs that belong to one BSS (e.g., BSS1) can use directional transmission to UTs that belong to another BSS (e.g., BSS2), even if each of them performed a calibration with a different AP. However, in this case there is uncertainty in the phase angle (for each subband). This is because the calibration procedure described above establishes a relationship that is unique to the AP with which the calibration was performed. The indicated ratio is a complex constant,

where k is the subband index and j is the AP index, and 0 is the index of the reference antenna (e.g., antenna 0) used in the AP. In one embodiment, this constant is common to all UTs in a given BSS, but independent of different BSSs.

As a result, when a UT from BSS1 exchanges data with a UT in BSS2, directional communication without correction or compensation for this constant can result in a phase shift and amplitude scaling of the entire native system. The phase shift can be determined using a pilot signal (directional or non-directional) and removed at the receivers of each respective UT. In one embodiment, the amplitude correction or compensation may be an ordinary SNR scaling and may be removed by estimating the noise level at each receiver, which may affect the choice of transmission rate.

In various embodiments, peer-to-peer exchanges between UTs that belong to different BSSs may be performed as follows:

- the peer initiator (for example, UT BSS1) is the AP source (DAP), and the other UT (for example, UT BSS2) is the UT receiver (DUT).

- The DAP sends a MIMO pilot to the DUT along with a link setup request that contains the BSS ID and DAP ID. The request should be sent in general mode (i.e., diversity in transmission).

- The DUT responds by sending a pilot MIMO pilot and an acknowledgment that contains the DUT ID, its BSS ID and some indicator of the transmission rate for use in the DAP.

- The DAP receiver (Rx) can perform an uplink (UL) phase shift estimate and apply a correction constant for each subband. The DAP can then use downlink directional transmission (DL), but must include the directional reference preamble in at least the first directional packet to enable the DUT receiver (Rx) to perform DL phase correction or compensation for each subrange. Subsequent DL transmissions may not require a preamble with a directional reference signal. Speed control and tracking can be done by splitting the gears into DL and UL segments with sufficient time intervals between them to complete the processing.

The calibration methods described herein can be implemented using various means. For example, these methods may be implemented in hardware, software, or a combination thereof. In the case of hardware implementation, the methods can be implemented at the access point and user terminal in one or more application-oriented integrated circuits (ASIC), digital signal processors (DSP), digital signal processor devices (DSPD), programmable logic devices (PLD ), in-circuit programmable gate arrays (FPGAs), processors, controllers, microcontrollers, and other electronic units configured to perform the functions described in this description or combinations thereof.

In the case of implementation in the form of software, calibration methods can be implemented using modules (for example, procedures, functions, etc.) that perform the functions described in the present description. Software codes may be stored in a storage device (e.g., storage devices 532 and 582 of FIG. 5) and executed by a processor (e.g., controllers 530 and 580, respectively). The storage device can be performed in the processor or as external to the processor, in which case it can be connected with the possibility of exchanging data with the processor using various means known in the art.

Headings are included herein for reference and to aid in the search for specific sections. These headings should not be construed as limiting the scope of concepts in their sections, and these concepts may be applied in other sections throughout the description.

The above description of the disclosed embodiments is provided to enable any person skilled in the art to use the present invention. Various modifications with respect to these embodiments should be apparent to those skilled in the art, and the general principles set forth herein are applicable to other embodiments without departing from the spirit and scope of the present invention. Thus, the present invention should not be limited to the embodiments disclosed in the present description, but on the contrary, corresponds to the widest scope compatible with the principles and new features disclosed in the present description.

Claims (38)

1. A method for establishing peer-to-peer communications by a first user terminal (UT) in a wireless communication system, comprising the steps of:
receiving a downlink response estimate;
receiving an uplink channel response estimate;
determining the first and second sets of correction factors based on estimates of the responses of the downward and upward channels;
calibrating the downlink and uplink based on the first and second sets of correction factors, respectively, to form a calibrated downlink and a calibrated uplink; and
establish direct peer-to-peer communication with the second user terminal without additional calibration between the first and second user terminals. 2. The method according to claim 1, in which the first set of correction factors are used to scale the symbols before transmission on the downlink, and the second set of correction factors are used to scale the symbols before transmission on the upward channel. 3. The method according to claim 1, in which the first set of correction factors are used to scale the symbols received on the downward channel, and the second set of correction factors
used to scale the symbols received on the uplink. 4. The method according to claim 1, in which the first and second sets of correction factors are determined based on the following equation:

Where

is a downlink response matrix

is an uplink channel response estimation matrix,

is a matrix of the first set of correction factors,

is a matrix of a second set of correction factors, and
^T stands for transposition. 5. The method according to claim 4, in which the determination of the first and second sets of correction factors includes calculating the matrix C in the form of an element-wise relation of the matrix

and matrices

and matrix output

and

based on the matrix C. 6. The method according to claim 5, in which the output matrix

includes the normalization of each of the many rows of the matrix C , and determining the average for the set of normalized rows of the matrix C , and the matrix

form, based on the specified average for many normalized rows. 7. The method according to claim 5, in which the output matrix

includes the normalization of each of the many columns of the matrix C and determining the average for the inverse values of the set of normalized columns of the matrix C , and the matrix

form, based on the specified average for the inverse values of the set of normalized columns. 8. The method according to claim 4, in which the matrix

and

derive based on a calculation with a minimum mean square error (MMSE). 9. The method of claim 8, in which when calculating MMSE minimize the standard error (MSE), expressed as

10. The method according to claim 1, further comprising determining a scale value corresponding to the average difference between the downlink channel response estimate and the uplink channel response estimate. 11. The method according to claim 1, in which the estimates of the responses of the downward and upward channels are normalized to take into account the noise level in the receiver. 12. The method according to claim 1, in which the determination is performed in the user terminal. 13. The method according to claim 4, in which the first set of matrices of correction factors for the downward channel is determined for the first set of subbands, the method further comprising interpolating the first set of matrices to obtain a second set of matrices of correction factors for the downward channel for the second set subbands. 14. The method according to claim 1, in which the estimates of the responses of the downward and upward channels are obtained based on the pilot signal transmitted through multiple antennas and orthogonalized using multiple orthogonal sequences. 15. The method according to claim 1, in which an estimate of the response of the uplink channel is obtained based on the pilot signal transmitted on the uplink channel, and in which an estimate of the response of the downlink channel is obtained based on the pilot signal transmitted on the downlink channel. 16. The method according to claim 1, in which the TDD system is a system with multiple inputs and multiple outputs (MIMO). 17. The method according to claim 1, wherein the TDD system uses orthogonal frequency division multiplexing (OFDM). 18. A method for establishing peer-to-peer communications by means of a first user terminal (UT) in a time division duplex (TDD) wireless communication system with multiple inputs and multiple outputs (MIMO), comprising the steps of transmitting the pilot signal on the uplink;
receiving an uplink channel response estimate derived based on the pilot transmitted on the uplink;
receive the pilot signal in the downward channel;
obtaining a downlink response estimate derived based on the pilot transmitted on the downlink; and
determining the first and second sets of correction factors based on the estimates of the responses of the downlink and uplink channels, wherein a calibrated downlink channel is generated using the first set of correction factors for the downlink and a calibrated uplink channel is generated using the first set of correction factors for the uplink; and
establish direct peer-to-peer communication with the second user terminal without additional calibration between the first and second user terminals. 19. The method of claim 18, wherein the first and second sets of correction factors are determined based on a calculation with a minimum mean square error (MMSE). 20. The method according to p, in which the first and second sets of correction factors are determined based on the calculation of the ratio of the matrices. 21. The method of claim 18, wherein the first set of correction factors is updated based on calibration performed with a plurality of user terminals. 22. The method of claim 18, further comprising scaling the symbols with a first set of correction factors before downlink transmission. 23. The method of claim 18, further comprising scaling the symbols with a second set of correction factors before transmitting on the uplink. 24. A first user terminal (UT) for determining correction factors in a time division duplex (TDD) wireless communication system with multiple inputs and multiple outputs (MIMO), comprising
means for obtaining an estimate of the response of the downward channel;
means for obtaining an estimate of the response of the upward channel;
means for determining the first and second sets of correction factors based on estimates of the responses of the downward and upward channels, wherein a calibrated downlink channel is formed using the first set of correction factors for the downlink channel and a calibrated upward channel is formed using the second set of correction factors for the uplink channel; and
means for establishing direct peer-to-peer communication with the second user terminal without additional calibration between the first and second user terminals. 25. A user terminal in a wireless time division duplex (TDD) communication system comprising
TX spatial processor, configured to transmit the first pilot signal on the uplink;
An RX spatial processor, configured to receive a second downlink pilot and derive a downlink response estimate based on the received second pilot and receive an uplink response estimate derived based on the transmitted first pilot; and
a controller configured to determine the first and second sets of correction factors based on estimates of the responses of the downstream and upstream channels, wherein the calibrated downlink channel is formed using the first set of correction factors for the downlink and the calibrated uplink channel is formed using the second set of correction factors for the uplink , and the determination of the first and second sets of correction factors, based on a calculation with a minimum mean square error by hand (MMSE). 26. The user terminal according A.25, in which the controller is additionally configured to determine the first and second sets of correction factors, based on the calculation of the ratio of the matrices. 27. A communication method in a wireless system, comprising the steps of:
calibrate one or more communication lines between a plurality of user stations and one or more access points based on one or more sets of correction factors derived from channel response estimates associated with one or more communication lines, the many user stations including a first user station and a second user station; and
establish communication between the first and second user stations using directional communication, without performing calibration between the first and second user stations, and based on said one or more sets of correction factors. 28. The method according to item 27, in which the establishment of communication between the first and second user stations comprises the steps of
sending a pilot signal and a request for establishing a communication line with the second user station from the first user station;
sending a directed pilot signal and acknowledgment from the second user station in response to receiving the pilot signal and the request from the first user station;
transmit information between the first and second user stations using directional communication based on the directional pilot signal. 29. The method of claim 28, wherein the communication request comprises an identifier of a base service area to which the first user station belongs and an identifier of the first user station. 30. The method according to p. 28, in which the confirmation contains the identifier of the second user station, the identifier of the base service area to which the second user station belongs, and a data rate indicator. 31. The method according to item 27, in which one or more access points include a first access point associated with the first basic service area (BSS), and a second access point associated with the second BSS, the first user station calibrated with respect to the first access points, and the second user station is calibrated with respect to the second access point, and establishing a connection between the first and second user stations comprises the steps of
sending a pilot signal and a request for establishing a communication line with the second user station from the first user station; sending a directed pilot signal and acknowledgment from the second user station in response to receiving the pilot signal and the request from the first user station; and
transmit information between the first and second user stations using directional communication, which is configured to compensate for the phase shift due to the calibration of the first and second user stations with respect to different access points. 32. The method according to p, in which the phase shift is determined based on the directional pilot signal received from the second user station. 33. A device for communicating in a wireless system, comprising
means for calibrating one or more communication lines between a plurality of user stations and one or more access points based on one or more sets of correction factors derived from channel response estimates associated with one or more communication lines, the many user stations including the first a user station and a second user station; and
means for establishing communication between the first and second user stations using directional communication, without performing a calibration between the first and second user stations, and based on said one or more sets of correction factors. 34. The device according to p, in which the establishment of communication between the first and second user stations contains
means for sending from the first user station a pilot signal and a request for establishing a communication line with the second user station;
means for sending a directed pilot signal from the second user station and acknowledging in response to receiving the pilot signal and the request from the first user station;
means for transmitting information between the first and second user stations using directional communication based on the directional pilot signal. 35. The device according to clause 34, in which the request for establishing a connection contains the identifier of the basic service area to which the first user station belongs, and the identifier of the first user station. 36. The device according to clause 34, in which the confirmation contains the identifier of the second user station, the identifier of the base service area to which the second user station belongs, and a data rate indicator. 37. The device according to p. 33, in which one or more access points include a first access point associated with the first basic service area (BSS), and a second access point associated with the second BSS, the first user station calibrated with respect to the first access points, and the second user station is calibrated with respect to the second access point, and establishing a connection between the first and second user stations comprises the steps of
means for sending from the first user station a pilot signal and a request for establishing a communication line with the second user station;
means for sending a pilot signal from the second user station and acknowledging in response to receiving the pilot signal and the request from the first user station;
means for transmitting information between the first and second user stations using directional communication, which is configured to compensate for the phase shift, due to the calibration of the first and second user stations with respect to different access points. 38. The device according to clause 37, in which the phase shift is determined based on the directional pilot signal received from the second user station.

Images