JP3423275B2 - Doppler frequency estimation circuit and wireless device using Doppler frequency estimation circuit - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a configuration of a wireless device capable of changing antenna directivity in real time, and more particularly to a configuration of a wireless device used in an adaptive array wireless base station.

[0002]

2. Description of the Related Art In recent years, various transmission channel allocation methods have been proposed in the mobile communication system in order to effectively use frequencies, and some of them have been put into practical use.

FIG. 13 shows a frequency division multiple access (Frequency).
FIG. 3 is a layout diagram of channels in various communication systems of Division Multiple Access (FDMA), Time Division Multiple Access (TDMA), and PDMA (Path Division Multiple Acess).

First, referring to FIG. 13, FDMA, TD
MA and PDMA will be briefly described. FIG.
(A) is a figure which shows FDMA, Comprising: Different frequency f1
The analog signals of the users 1 to 4 are frequency-divided and transmitted by the radio waves of to f4, and the signals of the users 1 to 4 are separated by the frequency filter.

In the TDMA shown in FIG. 13B,
The digitized signals of each user have different frequencies f1
The signal of each user is separated by the frequency filter and the time synchronization between the base station and each user mobile terminal device.

On the other hand, recently, the PDMA system has been proposed in order to increase the frequency utilization efficiency of radio waves due to the widespread use of portable telephones. This PDMA system is shown in FIG.
As shown in (1), one time slot in the same frequency is spatially divided and data of a plurality of users is transmitted. In this PDMA, the signal of each user is separated by using a frequency filter, time synchronization between the base station and each user mobile terminal device, and a mutual interference canceling device such as an adaptive array.

The operating principle of such an adaptive array radio base station is described in the following document, for example.

B. Widrow, et al .: “Adaptive Antenna
Systems, "Proc. IEEE, vol.55, No.12, pp.2143-2159
(Dec. 1967). SP Applebaum: “Adaptive Arrays”, IEEE Trans.
Antennas & Propag., Vol.AP-24, No.5, pp.585-598
(Sept. 1976). OL Frost, III: “Adaptive Least Squares Optimiz
ation Subject to Linear Equality Constraints, "SEL
-70-055, Technical Report, No.6796-2, Information
System Lab., Stanford Univ. (Aug. 1970). B. Widrow and SD Stearns: “Adaptive Signal Pr
ocessing, "Prentice-Hall, Englewood Cliffs (198
Five). RA Monzingo and TW Miller: “Introduction t
o Adaptive Arrays, "John Wiley & Sons, New York (1
980). JE Hudson: “Adaptive Array Principles,” Peter P
eregrinus Ltd., London (1981). RT Compton, Jr .: “Adaptive Antennas − Concept
s and Performance, "Prentice-Hall, Englewood Cliffs
(1988). E. Nicolau and D. Zaharia: “Adaptive Arrays,” Els
evier, Amsterdam (1989). FIG. 14 is a schematic diagram conceptually showing the operation principle of such an adaptive array radio base station. In FIG. 14, one adaptive array radio base station 1 is provided with an array antenna 2 including n antennas # 1, # 2, # 3, ..., #n, and its radio wave reaches the first range. It is shown as a shaded area 3. On the other hand, the range in which the radio waves of other adjacent radio base stations 6 reach is represented as a second shaded area 7.

In the area 3, a radio signal is transmitted and received between the mobile phone 4 which is the terminal of the user A and the adaptive array radio base station 1 (arrow 5). On the other hand, in the area 7, a radio signal is transmitted and received between the mobile phone 8 which is the terminal of another user B and the radio base station 6 (arrow 9).

[0010] Here, the mobile phone 4 of the user A happens to occur.
When the frequency of the radio wave signal of B is equal to the frequency of the radio wave signal of the mobile phone 8 of the user B, the radio wave signal from the mobile phone 8 of the user B becomes an unnecessary interference signal in the area 3 depending on the position of the user B. It will be mixed in the radio signal between the mobile phone 4 of the user A and the adaptive array radio base station 1.

As described above, in the adaptive array radio base station 1 which receives the mixed radio wave signals from both the users A and B, if the user A does not perform any processing.
A signal in which signals from both B and B are mixed is output, and the call of the user A who should originally call is hindered.

[Construction and Operation of Conventional Adaptive Array Antenna] In the adaptive array radio base station 1,
In order to remove the signal from the user B from the output signal, the following processing is performed. FIG. 15 is a schematic block diagram showing the configuration of the adaptive array radio base station 1.

First, assuming that the signal from the user A is A (t) and the signal from the user B is B (t), the received signal x1 (at the first antenna # 1 constituting the array antenna 2 of FIG. t) is expressed by the following equation: x1 (t) = a1 * A (t) + b1 * B (t) where a1 and b1 are coefficients that change in real time as described later.

Next, the received signal x at the second antenna # 2
2 (t) is expressed by the following equation: x2 (t) = a2 * A (t) + b2 * B (t) where a2 and b2 are coefficients that similarly change in real time.

Next, the received signal x at the third antenna # 3
3 (t) is expressed by the following equation: x3 (t) = a3 * A (t) + b3 * B (t) where a3 and b3 are similarly coefficients that change in real time.

Similarly, the received signal xn (t) at the nth antenna #n is expressed by the following equation: xn (t) = an × A (t) + bn × B (t) where: Similarly, an and bn are coefficients that change in real time.

The above coefficients a1, a2, a3, ..., An
Is the array antenna 2 for the radio signal from the user A.
The antennas # 1, # 2, # 3, ..., #n configuring the antennas have different relative positions (for example, the antennas are spaced from each other by 5 times the wavelength of the radio signal, that is, about 1 meter. Are arranged), and there is a difference in the reception intensity at each antenna.

Further, the above coefficients b1, b2, b3, ...
Similarly, bn also indicates that a radio signal from the user B has a difference in reception intensity at each of the antennas # 1, # 2, # 3, ..., #n. As each user is moving, these coefficients change in real time.

Signal x1 received by each antenna
(T), x2 (t), x3 (t), ..., xn (t) are
Corresponding switches 10-1, 10-2, 10-3, ...
The weight vector control unit 11 enters the receiving unit 1R configuring the adaptive array radio base station 1 via 10-n.
To the corresponding multipliers 12-1 and 12
, -2, 12-3, ..., 12-n are respectively applied to one input.

Weights w1, w2, w3, ..., Wn for received signals at the respective antennas are applied from the weight vector control unit 11 to the other inputs of these multipliers. These weights are calculated in real time by the weight vector control unit 11 as described later.

Therefore, the received signal x at antenna # 1
1 (t) passes through the multiplier 12-1 and then w1 × (a1A
(T) + b1B (t)), and the received signal x2 (t) at the antenna # 2 passes through the multiplier 12-2 and becomes w2 × (a
2A (t) + b2B (t)), and the received signal x3 (t) at the antenna # 3 passes through the multiplier 12-3 and becomes w3 ×.
(A3A (t) + b3B (t)), and the received signal xn (t) at the antenna #n becomes wn × (anA (t) + bnB (t)) via the multiplier 12-n.

These multipliers 12-1, 12-2, 12
The outputs of −3, ..., 12-n are added by the adder 13,
The output is as follows: w1 (a1A (t) + b1B (t)) + w2 (a2A
(T) + b2B (t)) + w3 (a3A (t) + b3B
(T)) + ... + wn (anA (t) + bnB (t)) This is divided into a term related to the signal A (t) and a term related to the signal B (t) as follows: (w1a1 + w2a2 + w3a3 + ... + wnan) A
(T) + (w1b1 + w2b2 + w3b3 + ... + wnb
n) B (t) Here, as will be described later, the adaptive array radio base station 1 identifies the users A and B, and the weights w1, w2, w3 are set so that only the signals from the desired users can be extracted.
..., calculate wn. For example, in the example of FIG. 15, the weight vector control unit 11 extracts coefficients A1, a in order to extract only the signal A (t) from the user A who should originally make a call.
, A, b1, b2, b3, ..., bn are regarded as constants so that the coefficient of the signal A (t) is 1 as a whole and the coefficient of the signal B (t) is 0 as a whole. Weight w
1, w2, w3, ..., Wn are calculated.

That is, the weight vector control unit 11
Solves the signal A by solving the following system of linear equations
Calculate the weights w1, w2, w3, ..., Wn in which the coefficient of (t) is 1 and the coefficient of the signal B (t) is 0: w1a1 + w2a2 + w3a3 + ... + wnan = 1 w1b1 + w2b2 + w3b3 + ... + wnbn = 0 This simultaneous linear equation Although the explanation of the solution is omitted, it is well known as described in the documents listed above, and is already put into practical use in the adaptive array radio base station.

Thus, the weights w1, w2, w3, ..., W
By setting n, the output signal of the adder 13 becomes as follows: Output signal = 1 × A (t) + 0 × B (t) = A (t) [User identification, training signal] The identification of the users A and B is performed as follows.

FIG. 16 is a schematic diagram showing the frame structure of a radio signal of a mobile phone. The radio signal of the mobile phone is mainly composed of a preamble composed of a signal sequence known to the radio base station and data (voice etc.) composed of a signal sequence unknown to the radio base station.

The signal sequence of the preamble includes a signal sequence of information for distinguishing whether the user is a desired user to talk to the radio base station. Weight vector control unit 11 of adaptive array radio base station 1
15 (FIG. 15) compares the training signal corresponding to the user A extracted from the memory 14 with the received signal sequence, and extracts the signal that seems to include the signal sequence corresponding to the user A. Performs vector control (determination of weight). The signal of the user A extracted in this way is externally output from the adaptive array radio base station 1 as an output signal SRX (t).

On the other hand, in FIG. 15, an input signal STX (t) from the outside enters the transmission unit 1T which constitutes the adaptive array radio base station 1, and is fed to the multipliers 15-1, 15-2,
15-3, ..., 15-n are given to one input. The weights w1, w2, w3, ..., Wn previously calculated by the weight vector control unit 11 based on the received signal are copied and applied to the other inputs of these multipliers.

The input signals weighted by these multipliers are sent to the corresponding switches 10-1, 10-2, 10
−3, ..., 10-n, the corresponding antenna # 1,
# 2, # 3, ..., #n and are transmitted in the area 3 of FIG.

Here, the signal transmitted using the same array antenna 2 as at the time of reception is the same as that of the received signal by the user A.
Since the target radio wave is weighted, the transmitted radio signal is received by the mobile phone 4 of the user A as if it had directivity to the user A. FIG. 17 is a diagram showing an image of transmission and reception of a radio signal between the user A and the adaptive array radio base station 1 as described above. In contrast to the area 3 of FIG. 14 which shows the range in which radio waves actually reach, as shown in the virtual area 3a of FIG. 17, the directivity is set from the adaptive array radio base station 1 to the mobile phone 4 of the user A as a target. Along with this, the image of the radio signal being radiated is visualized.

[0030]

As mentioned above, the PDMA is used.
The scheme requires a technique to remove co-channel interference. In this respect, the adaptive array that adaptively directs the null to the interference wave is an effective means because it can effectively suppress the interference wave even when the level of the interference wave is higher than the level of the desired wave.

By the way, when an adaptive array is used for the base station, it is possible not only to eliminate interference at the time of reception but also to reduce unnecessary radiation at the time of transmission. At this time, as the array pattern at the time of transmission, an array pattern at the time of reception may be used or a method of newly generating from the result of the DOA estimation or the like. The latter is FDD (Frequency Divi)
Although it can be applied regardless of whether it is a sion duplex or TDD (Time Division Duplex), complicated processing is required. On the other hand, when the former is used for FDD, the array pattern of transmission and reception is different, and therefore, the array arrangement and the weight must be corrected. Therefore, in general, application in TDD is premised, and good characteristics are obtained in an environment where external slots are continuous.

As described above, in the TDD / PDMA system using the adaptive array in the base station, when the array pattern (weight vector pattern) obtained in the uplink is used in the downlink, there is a motion with angular spread. When a typical Rayleigh propagation factor is assumed, the transmission directivity may deteriorate in the downlink due to the time difference between the uplink and the downlink.

That is, there is a time interval from the transmission of radio waves from the user terminal to the base station on the uplink (uplink) to the emission of the radio waves from the base station to the user terminal on the downlink (downlink). Therefore, when the moving speed of the user terminal cannot be ignored, the transmission directivity deteriorates due to an error between the emission direction of the radio wave from the base station and the actual direction in which the user terminal exists.

As a method of estimating the downlink weight in consideration of such channel fluctuations, a method of estimating the downlink transmission response vector by extrapolation using the reception response vector obtained in the uplink is proposed. Has been done.

However, if there is an estimation error in the reception response vector estimated in the uplink due to noise in the received signal, sampling error, etc., an error will occur in the result of the extrapolation processing depending on the degree of fading in the propagation path, and It is impossible to accurately estimate the transmission response vector of the line, and it becomes impossible to perform good transmission directivity control. Therefore, in order to prevent the occurrence of extrapolation error, it is necessary to know the degree of fading in the propagation path, that is, the Doppler frequency.

A method for estimating the degree of fading by obtaining a correlation value of a reference signal included in a received signal that is temporally before and after has been proposed, for example, Japanese Patent Laid-Open No. 7-16.
It is disclosed in Japanese Patent No. 2360. However, in such a conventional method, since the correlation value is calculated using the reference signal included in the received signal itself, there is a problem in that the interference value is large and accurate estimation is difficult.

Further, since the timing of the reference signal is fixed, the correlation value cannot be calculated at any timing, and there is a problem in that the flexibility of arithmetic processing is lacking.

On the other hand, there is a problem that a method for estimating the Doppler frequency of the propagation path for each user terminal separated by the adaptive array processing has not been developed yet.

The present invention has been made to solve the above problems, and is a user terminal which is not affected by an interference component in a received signal and has increased flexibility in arithmetic processing. It is an object of the present invention to provide a Doppler frequency estimation circuit for estimating the Doppler frequency for each of the above, and a wireless device using such a Doppler frequency estimation circuit.

[0040]

DISCLOSURE OF THE INVENTION The present invention relates to a radio apparatus in which antenna directivity is changed in real time and signals are transmitted and received to and from a plurality of terminals in a time division manner. Is a Doppler frequency estimation circuit for estimating the Doppler frequency of the received signal, based on signals received by a plurality of discretely arranged antennas, a received signal separation for separating a signal from a specific terminal among a plurality of terminals. Means, and a reception channel estimation means for estimating a reception response vector of a channel from a specific terminal based on signals received by a plurality of antennas, and receptions temporally before and after estimated by the reception channel estimation means. Correlation calculation means for calculating the vector correlation value based on the response vector, based on the correspondence relationship between the vector correlation value and the Doppler frequency that are empirically determined in advance, Seki and a estimating means for estimating a Doppler frequency corresponding to the vector correlation value calculated by the calculating means.

According to the present invention, by obtaining the correlation value between the reception response vectors instead of the received signal itself, the Doppler frequency of the propagation path of each separated specific terminal can be accurately measured without being affected by the interference component. Can be estimated.

Preferably, the correlation calculation means includes calculation means for calculating an instantaneous correlation value between reception response vectors that are temporally preceding and following each other and outputting it as a vector correlation value.

According to the present invention, since the reception response vectors can be correlated with each other at arbitrary timing, the instantaneous Doppler frequency of the propagation path can be accurately estimated.

More preferably, the correlation calculating means calculates the instantaneous correlation value between the reception response vectors that are temporally preceding and following, and the past correlation value and the current correlation value calculated by the calculating means. Averaging means for outputting an average value obtained by weighting and averaging as a vector correlation value.

According to the present invention, even if an error occurs in the instantaneous Doppler frequency due to sudden fading, the correlation value is averaged so that the Doppler frequency can be accurately measured without being affected by such an error. Can be estimated.

More preferably, the predetermined weight coefficient is set so that the weight for the past correlation value is large and the weight for the current correlation value is small.

According to the present invention, by increasing the weighting of the past correlation values in averaging, even if an error occurs in the instantaneous Doppler frequency due to sudden fading, such error is affected. Without, the Doppler frequency can be estimated more accurately.

More preferably, the correlation calculating means calculates the vector correlation value based on the reception response vector in the slot of the current frame and the reception response vector in the slot of the immediately preceding frame.

According to the present invention, the reception response vector is
Unlike the reference signal in the conventional technique, mutual correlation can be obtained at any timing, so that the flexibility of calculation for calculating the correlation value can be increased.

More preferably, the correlation calculating means calculates the vector correlation value based on the reception response vector in the slot of the current frame and the reception response vector in the most recent slot in which there was no reception error in the slots of the past frames. To calculate.

According to the present invention, the reception response vector is
Unlike the reference signal in the prior art, mutual correlation can be obtained at any timing, so that the flexibility of the calculation for calculating the correlation value can be increased, and in particular, the influence of the reception error can be eliminated. be able to.

More preferably, the correlation calculating means calculates the vector correlation value based on the reception response vector in the first half of the same slot and the reception response vector in the latter half.

According to the present invention, the reception response vector is
Unlike the reference signal in the conventional technique, mutual correlation can be obtained at any timing, so that the flexibility of calculation for calculating the correlation value can be increased.

According to another aspect of the present invention, there is provided a wireless device which changes the antenna directivity in real time and transmits / receives signals to / from a plurality of terminals in a time division manner. , And a transmission circuit and a reception circuit that share a plurality of antennas when transmitting and receiving signals,
The reception circuit, when receiving the reception signal, based on the signals from the plurality of antennas, the reception signal separation means for separating a signal from a specific terminal among the plurality of terminals, and a plurality of reception signal separation means when receiving the reception signal. Based on the signal from the antenna, including the reception propagation path estimation means for estimating the reception response vector of the propagation path from the specific terminal, the transmission circuit, based on the estimation result of the reception propagation path estimation means, of the transmission signal A transmission channel estimating means for estimating the transmission response vector of the channel during transmission,
Transmission directivity control means for updating the antenna directivity at the time of transmitting the transmission signal based on the estimation result of the transmission channel estimation means. The transmission channel estimation means determines the transmission response vector of the downlink slot to the specific terminal by extrapolation processing based on the plurality of reception response vectors of the uplink slot from the specific terminal estimated by the reception channel estimation means. Extrapolation means for calculating, Doppler frequency estimation means for estimating the Doppler frequency of the propagation path, predetermined in accordance with the Doppler frequency of the propagation path, storage means for holding a plurality of parameters used in the extrapolation processing, holding And a selecting means for selecting a parameter corresponding to the estimated Doppler frequency from the plurality of parameters thus selected and applying it to the extrapolation processing by the extrapolating means. Doppler frequency estimation means
Correlation calculation means for calculating a vector correlation value based on the reception response vectors estimated in time by the reception propagation path estimation means, and based on the correspondence relationship between the vector correlation value and the Doppler frequency determined empirically in advance. And an estimating means for estimating the Doppler frequency corresponding to the vector correlation value calculated by the correlation calculating means.

According to the present invention, by obtaining the correlation value between the reception response vectors instead of the received signal itself, the Doppler frequency of the propagation path of each separated specific terminal is accurately obtained without being affected by the interference component. Therefore, even if there is an estimation error in the reception response vector estimated in the uplink, it is possible to accurately estimate the transmission response vector in the downlink, which in turn realizes good transmission directivity control. can do.

Preferably, the correlation calculating means includes a calculating means for calculating an instantaneous correlation value between reception response vectors that are temporally before and after and outputting it as a vector correlation value.

According to the present invention, since the reception response vectors can be correlated with each other at any timing, it is possible to accurately estimate the instantaneous Doppler frequency of the propagation path, and thus the more accurate transmission response vector. Can be estimated.

More preferably, the correlation calculating means calculates the instantaneous correlation value between the reception response vectors that are temporally preceding and following, and the past correlation value and the current correlation value calculated by the calculating means. Averaging means for outputting an average value obtained by weighting and averaging as a vector correlation value.

According to the present invention, even when an error occurs in the instantaneous Doppler frequency due to abrupt fading, the correlation value is averaged so that the Doppler frequency can be accurately measured without being affected by such error. The transmission response vector can be estimated more accurately.

More preferably, the predetermined weight coefficient is set so that the weight for the past correlation value is large and the weight for the current correlation value is small.

According to the present invention, by increasing the weighting of the past correlation values in averaging, even if an error occurs in the instantaneous Doppler frequency due to sudden fading, such error is affected. It is possible to more accurately estimate the Doppler frequency, and thus more accurately estimate the transmission response vector.

More preferably, the correlation calculating means calculates the vector correlation value based on the reception response vector in the slot of the current frame and the reception response vector in the slot of the immediately preceding frame.

According to the present invention, the reception response vector is
Unlike the reference signal in the conventional technique, mutual correlation can be obtained at any timing, so that the flexibility of calculation for calculating the correlation value can be increased.

More preferably, the correlation calculating means calculates the vector correlation value based on the reception response vector in the slot of the current frame and the reception response vector in the most recent slot in which there was no reception error in the slots of the past frames. To calculate.

According to the present invention, the reception response vector is
Unlike the reference signal in the prior art, mutual correlation can be obtained at any timing, so that the flexibility of the calculation for calculating the correlation value can be increased, and in particular, the influence of the reception error can be eliminated. be able to.

More preferably, the correlation calculating means calculates the vector correlation value based on the reception response vector in the first half and the reception response vector in the latter half of the same slot.

According to the present invention, the reception response vector is
Unlike the reference signal in the conventional technique, mutual correlation can be obtained at any timing, so that the flexibility of calculation for calculating the correlation value can be increased.

[0068]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS [First Embodiment] FIG. 1 is a schematic block diagram showing a configuration of a radio apparatus (radio base station) 1000 of a PDMA base station according to an embodiment of the present invention.

In the configuration shown in FIG. 1, the user PS
4 antennas # 1 to # 1 in order to distinguish 1 from PS2.
# 4 is provided. However, the number of antennas may be more generally N (N: natural number).

In transmission / reception system 1000 shown in FIG. 1, signals are received from antennas # 1 to # 4 and received by corresponding user, for example, receiving section SR1 and user PS1 for separating signals from user PS1. A transmitter ST1 for transmitting a signal is provided. Antenna #
The connections between 1 to # 4 and the receiving section SR1 and the transmitting section ST1 are selectively switched by the switches 10-1 to 10-4.

That is, received signals RX1 (t), RX2 (t), RX3 (t), received by the respective antennas,
RX4 (t) is the corresponding switch 10-1, 10-
2, 10-3, 10-4 via the receiving section SR1,
It is given to the reception weight vector calculator 20 and the reception response vector calculator 22, and the corresponding multiplier 12-
It is given to one input of 1, 12-2, 12-3, 12-4, respectively.

To the other inputs of these multipliers, the weighting factors wrx11, wrx21, wr from the reception weight vector calculator 20 for the reception signals at the respective antennas are input.
x31 and wrx41 are applied. These weighting factors are calculated by the reception weight vector calculator 20 as in the conventional example.
Is calculated in real time.

The transmitting section ST1 receives the reception response vector calculated by the reception response vector calculator 22, and estimates the propagation path at the time of transmission, that is, the virtual reception response at the time of transmission, as will be described later. A transmission response vector estimator 32 that obtains a transmission response vector by estimating a vector, a memory 34 that transmits and receives data between the transmission response vector estimator 32, and stores and holds the data, and a transmission response vector estimator 32. Transmission weight vector calculator 3 for calculating a transmission weight vector based on the estimation result
0, the transmission signal is received at one input, and the weight coefficient wtx from the transmission weight vector calculator 30 is received at the other input.
Multipliers 15-1, 15-2, 15-3 and 15-4 to which 11, wtx21, wtx31 and wtx41 are applied are included. Multipliers 15-1, 15-2, 15-3, 15-4
The output from is via the switches 10-1 to 10-4,
Antennas # 1 to # 4 are provided.

Although not shown in FIG. 1, the same configuration as the receiving section SR1 and the transmitting section ST1 is provided for each user.

[Principle of Operation of Adaptive Array] The operation of the receiver SR1 will be briefly described as follows.

Received signal RX1 received by the antenna
(T), RX2 (t), RX3 (t), RX4 (t) are
It is expressed by the following formula.

[0077]

[Equation 1]

Here, the signal RXj (t) is the j-th (j
= 1, 2, 3, 4, 4), and the signal Srxi (t) represents the signal transmitted by the i-th (i = 1, 2) user.

Further, the coefficient hji represents the complex coefficient of the signal from the i-th user received by the j-th antenna, and nj (t) represents the noise included in the j-th received signal. .

The above equations (1) to (4) are expressed in vector form as follows.

[0081]

[Equation 2]

In equations (6) to (8), [...] T
Indicates the transposition of [...]. Here, X (t) is the input signal vector, Hi is the reception response vector of the i-th user,
N (t) indicates a noise vector, respectively.

As shown in FIG. 1, the adaptive array antenna multiplies the input signals from the respective antennas by the weighting factors wrx1i to wrx4i, and outputs the combined signal as the reception signal SRX (t).

With the above preparations, the operation of the adaptive array when extracting the signal Srx1 (t) transmitted by the first user is as follows.

Output signal y1 of adaptive array 100
(T) can be expressed by the following equation by multiplying the vector of the input signal vector X (t) and the weight vector W1.

[0086]

[Equation 3]

That is, the weight vector W1 is a weighting factor w multiplied by the j-th input signal RXj (t).
It is a vector whose elements are rxj1 (j = 1, 2, 3, 4).

Here, y1 represented by the equation (9)
Substituting the input signal vector X (t) expressed by the equation (5) into (t) gives the following.

[0089]

[Equation 4]

When the adaptive array 100 operates ideally, the weight vector W1 is sequentially controlled by the well-known method so that the weight vector control unit 11 satisfies the following simultaneous equations.

[0091]

[Equation 5]

When the weight vector W1 is completely controlled so as to satisfy the expressions (12) and (13), the output signal y1 (t) from the adaptive array 100 is finally expressed by the following expression.

[0093]

[Equation 6]

That is, the output signal y1 (t) contains the signal Sr transmitted by the first user of the two users.
x1 (t) will be obtained.

[Outline of Operation of Radio Equipment 1000] FIG.
3 is a flowchart for explaining an outline of a basic operation of the wireless device 1000 which is a premise of the present invention.

In radio apparatus 1000, attention is paid to the fact that the weight vector (weighting coefficient vector) of the adaptive array can be uniquely expressed by the reception response vector of each antenna element, and the time variation of the reception response vector is estimated to indirectly To estimate the weight.

First, the receiver SR1 estimates the propagation path of the received signal based on the received signal (step S).
100). The estimation of the propagation path corresponds to obtaining the impulse response of the signal sent from the user in the equations (1) to (4).

In other words, in equations (1) to (4), for example, if the reception response vector H1 can be estimated,
It is possible to estimate the transmission path when the signal from the user PS1 is received.

Next, the transmission response vector estimator 32
Predicts the propagation path at the time of transmission, that is, the reception response vector at the time of transmission from the reception response vector at the time of reception (step S102). This predicted reception response vector corresponds to the transmission response vector at the time of transmission.

Further, the transmission weight vector calculator 30
Calculates the transmission weight vector based on the predicted transmission response vector, and the multipliers 15-1 to 15-4
(Step S104).

[Operation of Reception Response Vector Calculator 22] Next, the basic operation of the reception response vector calculator 22 shown in FIG.

First, assuming that the number of antenna elements is four and the number of users communicating at the same time is two, the signal output from the receiving circuit via each antenna is expressed by the above equations (1) to (4).
It is represented by.

At this time, if the equations in which the received signals of the antennas represented by the equations (1) to (4) are expressed by vectors are to be described again, the following equations (5) to (8) are obtained.

[0104]

[Equation 7]

Here, when the adaptive array is operating well, the signals from each user are separated and extracted, so that the signals Srxi (t) (i = 1, 2) are all known values. Become.

At this time, by utilizing the fact that the signal Srxi (t) is a known signal, the reception response vector H1 = [h
11, h21, h31, h41] and H2 = [h12, h22, h
32, h42] can be derived as described below.

That is, the received signal and the known user signal, for example, the signal Srx1 (t) from the first user.
When the ensemble average (time average) is calculated by multiplying by, the result is as follows.

[0108]

[Equation 8]

In the equation (16), E [...] Indicates the time average, and S * (t) is the conjugate complex of S (t).
If the time to take the average is long enough, the average value is as follows.

[0110]

[Equation 9]

Here, the value of the equation (18) becomes 0 because
This is because the signal Srx1 (t) and the signal Srx2 (t) have no correlation with each other. The value of the equation (19) becomes 0 because there is no correlation between the signal Srx1 (t) and the noise signal N (t).

Therefore, the ensemble average of equation (16) results in the received response vector H as shown below.
Is equal to 1.

[0113]

[Equation 10]

By the above procedure, the reception response vector H1 of the signal transmitted from the first user PS1 can be estimated.

Similarly, the reception response vector H2 of the signal transmitted from the second user PS2 can be estimated by performing the ensemble averaging operation of the input signal vector X (t) and the signal Srx2 (t). Is.

The ensemble averaging as described above is performed, for example, on a predetermined number of data symbol strings at the beginning and a predetermined number of data symbol strings at the end in one time slot at the time of reception.

[Estimation of Transmission Response Vector] FIG. 3 is a conceptual diagram for explaining the basic operation of the transmission response vector estimator 32 which is the premise of the present invention. Consider an 8-slot configuration in which 4 users are assigned to each of the upper and lower lines as a PDMA burst. The slot configuration is such that, for example, the first 31 symbols are the first training symbol sequence, the subsequent 68 symbols are the data symbol sequence, and the last 31 symbols are the second training symbol sequence.

As described above, the training symbol sequence is provided at the beginning and the end of the uplink slot, and both reception response vectors are calculated using the algorithm of the reception response vector calculator 22 described above.

Then, the reception response vector for the downlink is estimated by extrapolation processing (linear extrapolation).

That is, if the value at any one time t of the elements of the reception response vector is f (t), the value f (t0) at the time t0 of the head training symbol sequence of the uplink slot and the uplink slot Based on the value f (t1) of the last training symbol sequence at time t1, the value f (t) at time t of the downlink slot is
It can be predicted as follows.

F (t) = [f (t1) -f (t0)] /
(T1−t0) × (t−t0) + f (t0) In the above description, a training symbol string is provided at the beginning and the end of the uplink slot, and primary extrapolation is performed. A training symbol string is also provided in the center of the slot, and the value f (t) at time t is calculated as 2 from the value of the reception response vector at three points in the uplink slot.
It is also possible to adopt a configuration in which the extrapolation is used for estimation. Alternatively, higher-order extrapolation can be performed by increasing the positions of the training symbol strings in the uplink slots.

The present invention relates to a Doppler frequency estimating circuit necessary for improving the downlink (reception) response vector estimating method by such extrapolation processing, the details of which will be described later. Now, the determination of the transmission weight vector will be described first.

[Determination of Transmission Weight Vector] When the estimated value of the reception response vector at the time of transmission is obtained as described above, the transmission weight vector can be obtained by any of the following three methods.

I) Method by orthogonalization Weight vector W (1) (i) = [w of user PS1 at time t = iT (i: natural number, T: unit time interval)
tx11, wtx12, wtx13, wtx14]. In order to direct the null to the user PS2, the following conditions may be satisfied.

The propagation path (reception response vector) predicted for the user PS2 is V (2) (i) = [h1 '(2) (i),
h2 '(2) (i), h3' (2) (i), h4 '(2) (i)]. Here, hp '(q) (i) is p of the qth user.
It is a predicted value for the time i of the reception response vector for the th antenna. Similarly, it is assumed that the propagation path V (1) (i) has been predicted for the user PS1.

At this time, W (1) (i) TV (2) (i) = 0
W (1) (i) is determined so that The following conditions c1) and c2) are imposed as restraint conditions.

C1) W (1) (i) TV (1) (i) = g (constant value) c2) ‖W (1) (i) ‖ is minimized.

The condition c2) corresponds to minimizing the transmission power. ii) Method Using Pseudo-Correlation Matrix Here, as described above, the adaptive array is composed of several antenna elements and a part for controlling each element weight value. Generally, when the input vector of the antenna is represented by X (t) and the weight vector is represented by W, the output Y (t) = WTX.
When the weight vector is controlled so as to minimize the mean square difference between (t) and the reference signal d (t) (MMSE criterion: least square error method criterion), the optimum weight Wopt is expressed by the following equation (Wiener solution). Given.

[0129]

[Equation 11]

However,

[0131]

[Equation 12]

It is necessary to satisfy. Here, YT represents the transpose of Y, Y * represents the complex region of Y, and E [Y] represents the ensemble average. This weight value causes the adaptive array to generate an array pattern so as to suppress unnecessary interference waves.

By the way, in the method using the pseudo correlation matrix, the above equation (21) is calculated by the pseudo correlation matrix described below.

That is, the weight vector W (k) (i) for the user k is calculated using the estimated complex received signal coefficient h '(k) n (i). Assuming that the array response vector of the kth user is V (k) (i), it can be obtained as follows, as described above.

[0135]

[Equation 13]

At this time, the autocorrelation matrix Rxx (i) of the virtual received signal at t = iT is expressed by the following equation using V (k) (i).

[0137]

[Equation 14]

However, N is a virtual noise term added because Rxx (i) is an integer. In the calculation in the present invention, for example, N = 1.0 × 10 −5.

Correlation vector rxd between received signal and reference signal
(I) is expressed by the following equation.

[0140]

[Equation 15]

Therefore, the weights for the downlink at time t = iT can be obtained from the equations (21), (25) and (26).

The inverse matrix operation of the equation (25) can be optimally calculated for the user k by the inverse matrix lemma.
In particular, in the case of two users, the weight is calculated by the following simple formula.

[0143]

[Equation 16]

Given an autocorrelation matrix in this way,
For the method of calculating the weight vector, see, for example, T. Ohgane, Y. Ogawa, and
K. Itoh, Proc. VTC'97, vol.
2, pp. 725-729, May 1997, or literature: Tanaka, Ohgane, Ogawa, Ito, IEICE Technical Report, vo
l. RCS98-117, pp. 103-108, Oc
t. 1998.

Iii) Method of directing beam to user PS1 Focusing only on the point of directing the beam to the user PS1, the following formula may be satisfied.

W (1) (i) = V (1) (i) * If the weight vector at the time of transmission is determined and transmitted by any one of the methods described above, dynamic Rayleigh such as angular spread can be obtained. When a propagation path is assumed, it is possible to suppress the deterioration of transmission directivity in the downlink caused by the time difference between the uplink and the downlink even in the TDD / PDMA system.

Next, FIG. 4 is a conceptual diagram for explaining a transmission response vector estimating principle which is a premise of the Doppler frequency estimating circuit according to the present invention. The "ideal state" shown in the upper part of FIG. 4 is basically a further simplification of the conceptual diagram shown in FIG.

That is, based on the reception response vector 1 and the reception response vector 2 which are the reception response vectors of the two points in the same slot of the uplink calculated by the reception response vector calculator 22 of FIG. 1 in step S100 of FIG. By performing linear extrapolation up to the original transmission timing of the corresponding slot of the downlink, the correct transmission response vector of the downlink can be estimated.

Here, the "ideal state" in FIG. 4 is based on the assumption that the reception response vectors 1 and 2 have no estimation error.

However, as shown in the case "including the response vector estimation error" in the lower part of FIG. 4, for example, the reception response vector 2 has an amplitude like the reception response vector 2'because of the estimation error due to noise or sampling error. In the case of deviation in the direction, if the linear extrapolation is performed based on these reception response vectors 1 and 2 '(with the same extrapolation distance) as in the "ideal state", the transmission response vector at the transmission timing becomes larger in the amplitude direction. Therefore, the wrong transmission response vector is estimated.

Therefore, if the transmission weight vector calculator 30 of FIG. 1 performs the transmission weight determination process (step S104 of FIG. 2) based on such an incorrect transmission response vector, the obtained transmission weight is also incorrect. Therefore, a directivity error in the downlink, that is, a transmission error is caused. In particular, since the radio base station and the terminal are at a long distance, a slight directional error causes a large transmission error.

Therefore, the application of the Doppler frequency estimating circuit of the present invention is to accurately estimate the Doppler frequency indicating the degree of fading in the propagation path, assuming that there is an estimation error in the uplink reception response vector. This is to attempt to realize a correct transmission directivity by estimating appropriate transmission response vectors in the downlink by adjusting appropriate parameters for the insertion processing, especially extrapolation distances.

FIG. 5 is a conceptual diagram for explaining the principle of extrapolation distance determination, which is the premise of the Doppler frequency estimating circuit of the present invention.

The propagation environment of the propagation path is represented by the fluctuation of the reception coefficient of the propagation path, that is, the degree of fading. As described above, the degree of fading is expressed by a so-called Doppler frequency (FD) as a physical quantity.

The principle of Doppler frequency estimation according to the present invention will be described later. First, the principle of determining the extrapolation distance will be described by focusing on the Doppler frequency representing the degree of fading.

As described above, when the reception response vector 2 is deviated by the estimation error like the reception response vector 2 ', the extrapolation error becomes larger as the extrapolation distance becomes longer, and it depends on the original transmission response vector. It will be wrong.

Generally, the smaller the fading, that is, the lower the Doppler frequency FD, the smaller the fluctuation of the reception coefficient of the propagation path. Therefore, in such a case, the extrapolation distance is shortened to prevent extrapolation beyond the actual fluctuation amount. More specifically, when the Doppler frequency FD is low, a short distance extrapolation from the reception response vector 2'to the point a marked with X is performed as in the case of FIG. The vector is estimated to be regarded as the correct transmission response vector at the point b of the X mark.

On the other hand, the greater the fading, the more
That is, the higher the Doppler frequency FD, the larger the fluctuation of the reception coefficient of the propagation path. Therefore, in such a case, a sufficient extrapolation is performed by increasing the extrapolation distance. More specifically, when the Doppler frequency FD is high, extrapolation is performed for a relatively long distance from the reception response vector 2'to the point c of the mark X as in the case of FIG. The transmission response vector is estimated and regarded as the correct transmission response vector at point d of the X mark.

Such processing is mainly executed by the transmission response vector estimator 32 shown in FIG.

Next, FIG. 6 is a schematic block diagram showing the configuration of the Doppler frequency estimating circuit according to the embodiment of the present invention. The operating principle of the Doppler frequency estimating circuit according to the embodiment of the present invention will be described below with reference to FIG.

In FIG. 6, the response vector estimation circuit 10
1 corresponds to the receiving unit SR1 in FIG. 1, and particularly corresponds to the reception response vector calculator 22 in FIG.

Here, a total of 8 slots consisting of 4 slots for the upper and lower lines as shown in FIG. 3 to FIG.
It is called a frame. Then, such frames are continuously transmitted in time series, and communication between the upper and lower lines is alternately performed.

The response vector estimation circuit 101 estimates the reception response vector in the slot of the current frame by applying the above-described ensemble averaging method to the received signal, and the correlation calculation and Doppler frequency estimation circuit 103. , And the memory 102.

The correlation calculation and Doppler frequency estimation circuit 103 uses the reception response vector in the slot of the current frame estimated by the response vector estimation circuit 101,
The correlation value with the reception response vector in the corresponding slot of the previous frame stored in the memory 102 is calculated.

The correlation value α of the reception response vectors of two frames that are temporally before and after is defined by the following equation.

Α = | h ₁ h ₂ ^H | / | h ₁ || h ₂ | where h ₂ ^H is the complex conjugate of each component of h _{2 and} is the transpose thereof.

Further, h _i (i = 1,2) represents the reception response vector (h _i1 , h _i2 , h _i3 , h _i4 ) having the phase amplitude information for each antenna element in the frame i as an element. .

Although it is difficult to find an accurate correspondence between the correlation value calculated in this way and the Doppler frequency, it is possible to experimentally find an approximate correspondence by experiment. For example, correlation values from 1 to 0.95
, The Doppler frequency FD is FD = 0.
It is estimated to be Hz. Further, if the correlation value is within the range of 0.95 to 0.80, it is estimated that FD = 10 Hz, and so on.

The approximate correlation between the reception response vector correlation value and the Doppler frequency FD obtained empirically in this way is the correlation calculation and the Doppler frequency estimation circuit 10.
3, the corresponding Doppler frequency estimated value is selected from the correlation values of the vectors calculated by the above-described calculation formula, and is output from the circuit 103.

The processing shown in FIG. 6 is usually executed by software using, for example, a digital signal processor (DSP). FIG. 7 is a flow chart showing processing by the circuit configuration shown in FIG. In the process shown in FIG. 7, a vector correlation value between the reception response vector in the slot of the current frame and the reception response vector in the corresponding slot of the immediately preceding frame is obtained.

First, in step S1, the reception response vector of the slot of the current frame is estimated.

Next, in step S2, step S
It is determined whether the reception response vector estimated in 1 is the reception response vector estimated first, and if it is the reception response vector estimated first, it is stored in the memory (memory 102 of FIG. 6) in step S5. Remembered.

On the other hand, if it is not the reception response vector estimated first, in step S3, the reception response vector of the corresponding slot of the immediately preceding frame held in the memory and the current frame estimated in step S1 Correlation value COR with the reception response vector of the corresponding slot
R is calculated.

Then, in step S4, the Doppler frequency F corresponding to the calculated correlation value CORR is calculated based on the correspondence between the vector correlation value and the Doppler frequency FD which are experimentally obtained and held in advance as described above.
D is estimated and output.

On the other hand, in step S5, the reception response vector of the slot of the current frame estimated in step S1 is stored in the memory (memory 102 in FIG. 6).

By repeatedly executing the above steps S1 to S5, it is possible to continuously obtain the instantaneous vector correlation value between the reception response vectors of the corresponding slots between two frames which are temporally preceding or following each other. You can

Next, FIG. 8 is a flow chart showing a modified example of the processing shown in FIG. In FIG. 8, if no reception error is detected in step S11, the subsequent operation is basically the same as the example of FIG. 7, and the description will not be repeated.
On the other hand, if the reception error is detected in step S11, the process cannot proceed to the next step S1 until it is not detected.

If the reception error is no longer detected, the operations of steps S1 to S5 are executed.
What is different from the example is the process of step S3 ′. That is, since the response vector estimation of the slot including the reception error is eliminated in step S11, in step S3 ′, the most recent slot having no reception error among the slots of the past frames held in the memory. Received response vector in step S
The correlation value CORR with the reception response vector of the slot of the current frame estimated by 1 is estimated. Subsequent processing is the same as the example of FIG. 7.

In the example of FIG. 7, the influence of the reception error can be eliminated, and the Doppler frequency can be estimated more accurately.

Next, FIG. 9 is a flow chart showing another process by the circuit configuration shown in FIG. In the process shown in FIG. 9, the vector correlation value between the reception response vector in the first half and the reception response vector in the latter half of the same slot is obtained.

First, the presence or absence of a reception error is determined in step S11 (step S11 may be omitted), and if no reception error is detected, step S12.
In step S13, the reception response vector of the leading edge of the current slot is estimated, and then in step S13, the reception response vector of the trailing edge of the current slot is estimated.

Next, in step S14, the correlation value CORR between the reception response vectors at each of the leading edge portion and the trailing edge portion is calculated.

Then, in step S15, based on the correspondence between the vector correlation value and the Doppler frequency FD, which are experimentally obtained and held in advance as described above,
The Doppler frequency FD corresponding to the calculated correlation value CORR is estimated and output.

As described above, in the embodiments shown in FIGS. 7 to 9, the reception response vectors that are temporally preceding and succeeding are different from the reference signal which is the target of the correlation value calculation in the prior art. Since the correlation value can be obtained even at such timing, the flexibility of the calculation for calculating the correlation value can be increased.

Next, FIG. 10 is a schematic block diagram showing the structure of a Doppler frequency estimating circuit according to another embodiment of the present invention. The circuit configuration shown in FIG. 10 differs from the circuit configuration shown in FIG. 6 in the following points.

That is, in the example of FIG. 6, the correlation calculation and Doppler frequency estimation circuit 103 calculates the instantaneous correlation value between two reception response vectors that are temporally preceding and following, whereas the example of FIG. The correlation calculation and Doppler frequency estimation circuit 105 in 1 further attempts to estimate the Doppler frequency more accurately by taking the average of the calculated correlation values.

More specifically, in FIG. 10, the correlation calculation and Doppler frequency estimation circuit 105 holds the reception response vector in the slot of the current frame estimated by the response vector estimation circuit 101 and the memory 102. The correlation value with the reception response vector in the corresponding slot of the previous frame is calculated, and the average of the correlation value just calculated and the past average correlation value read from the memory 104 and stored. Take a value.

In particular, in this embodiment, the average value weighted by a predetermined weighting coefficient is obtained. For example, in the circuit 105, one obtained by multiplying the past average correlation value stored in the memory 104 by the first weighting coefficient and one obtained by multiplying the just calculated current correlation value by the second weighting coefficient are used. The average value of is calculated.

Here, as the first coefficient, for example, 0.
A large coefficient such as 97 is set, and a second coefficient is set to a small coefficient such as 0.03. The frequency is estimated and output, and the obtained current average correlation value is stored in the memory 104.
Remember.

FIG. 11 is a flow chart showing the processing by the circuit configuration shown in FIG. In the processing shown in FIG. 11, a vector correlation value between the reception response vector in the slot of the current frame and the reception response vector in the slot of the immediately preceding frame is obtained, and the weighting between the past correlation value and the current correlation value is performed. The average is made.

First, in step S21, the reception response vector of the slot of the current frame is estimated.

Next, in step S22, it is determined whether or not the reception response vector estimated in step S21 is the first estimated reception response vector, and if it is the first estimated reception response vector, the step It is stored in the memory (memory 102 in FIG. 10) in S29.

On the other hand, if it is not the reception response vector estimated first, in step S23, the reception response vector of the slot of the immediately preceding frame held in the memory and the slot of the current frame estimated in step S21 are stored. The correlation value CORR NOW with the reception response vector is calculated.

Next, in step S24, it is determined whether or not the correlation value calculated in step S23 is the first calculated correlation value. If it is the first calculated correlation value, in step S25, Correlation value CORR NOW
An averaging operation is performed, where is the average value CORR AVE.

On the other hand, if it is not the correlation value calculated first, in step S26, the past average correlation value CORR OLD read from the memory and the current correlation value C are read.
A weighted averaging process for multiplying ORRNOW by the weighting factors α and 1-α is executed.

Then, in step S27, the Doppler corresponding to the calculated average correlation value CORR AVE is calculated based on the correspondence between the average correlation value and the Doppler frequency FD which are experimentally obtained and held in advance as described above. The frequency FD is estimated and output.

Next, at step S28, the current average correlation value COR calculated at step S25 or 26.
R AVE is stored in the memory (memory 104 in FIG. 10) as the past average correlation value CORR OLD.

Further, in step S29, the reception response vector of the slot of the current frame estimated in step S21 is stored in the memory (memory 102 in FIG. 10).
Memorized in.

By repeating the above steps S21 to S29, the Doppler frequency based on the averaged vector correlation value can be estimated. By averaging the vector correlation values in this way, even if an instantaneous Doppler frequency error occurs due to sudden fading, it is possible to accurately estimate the Doppler frequency without being affected by such an error. it can.
In particular, by setting the weighting coefficient so that the weight for the past correlation value becomes larger, even if an error occurs in the instantaneous Doppler frequency due to sudden fading, it is not affected by such error. , Can estimate the Doppler frequency more accurately.

In the examples shown in FIGS. 10 and 11, the averaging process is performed based on the correlation between the reception response vectors of the corresponding slots of consecutive frames, but the examples shown in FIGS. As described above, the averaging process based on the correlation between the reception response vector in the slot of the current frame and the reception response vector in the most recent slot in which there was no reception error in the slots of the past frame may be performed. , Averaging processing may be performed based on the correlation between reception response vectors in the first half and the second half of the same slot.

FIG. 12 is a flow chart showing an extrapolation process focusing on the Doppler frequency FD thus estimated.

Referring to FIG. 12, in step S31, first, the reception response vector calculator 22 of FIG. 1 estimates the propagation path, specifically, the reception response vectors 1 and 2'of the uplink are estimated.

Next, at step S32, the Doppler frequency estimating circuit according to the present invention described above estimates the Doppler frequency FD.

Then, in step S33, the transmission response vector estimator 32 of FIG. 1 uses the method described with reference to FIG. 5 to determine the optimum extrapolation parameter, that is, the extrapolation distance according to the estimated Doppler frequency FD. Is made. For this purpose, it is assumed that the optimum extrapolation distance determined by the preliminary measurement according to the level of the Doppler frequency FD is stored in the memory 34 of FIG. 1 in advance.

Next, in step S34, the extrapolation parameter (extrapolation distance) determined in step S33 described above.
Extrapolation processing is performed using
That is, the transmission response vector is estimated.

Finally, in step S35, the transmission weight vector calculator 30 of FIG. 1 estimates the transmission weight based on the downlink transmission response vector determined in step S34.

As described above, when the Doppler frequency estimation circuit according to the present invention is used, the optimum extrapolation distance can be selected depending on the level of the Doppler frequency FD accurately estimated for each terminal, and the uplink reception response can be selected. Even if there is an estimation error in the vector, the correct transmission response vector can be estimated.

The embodiments disclosed this time are to be considered as illustrative in all points and not restrictive. The scope of the present invention is shown not by the above description but by the scope of the claims, and is intended to include meanings equivalent to the scope of the claims and all modifications within the scope.

[0209]

As described above, according to the present invention, the Doppler frequency is estimated based on the correlation value between the reception response vectors, not on the reception signal itself, so that it is not affected by the interference component in the reception signal. Further, it becomes possible to more accurately estimate the Doppler frequency while increasing the flexibility of the arithmetic processing.

[Brief description of drawings]

FIG. 1 is a schematic block diagram showing a configuration of a radio device (radio base station) 1000 of a PDMA base station to which the present invention is applied.

FIG. 2 is a flowchart for explaining an outline of an operation of a wireless device (wireless base station) 1000.

FIG. 3 is a conceptual diagram for explaining a basic operation of a transmission response vector estimator 32.

FIG. 4 is a conceptual diagram for explaining a transmission response vector estimation principle that is a premise of the present invention.

FIG. 5 is a conceptual diagram for explaining the principle of extrapolation distance determination that is the premise of the present invention.

FIG. 6 is a schematic block diagram of a Doppler frequency estimation circuit according to an embodiment of the present invention.

FIG. 7 is a flowchart for explaining an example of Doppler frequency estimation operation according to the embodiment of the present invention.

FIG. 8 is a flowchart for explaining another example of the Doppler frequency estimation operation according to the embodiment of the present invention.

FIG. 9 is a flowchart for explaining yet another example of the Doppler frequency estimation operation according to the embodiment of the present invention.

FIG. 10 is a schematic block diagram of a Doppler frequency estimation circuit according to another embodiment of the present invention.

FIG. 11 is a flowchart illustrating an example of a Doppler frequency estimation operation according to another embodiment of the present invention.

FIG. 12 is a flowchart for explaining an outline of extrapolation processing using the Doppler frequency estimation circuit according to the present invention.

FIG. 13 is an arrangement diagram of channels in various communication systems of frequency division multiple access, time division multiple access, and PDMA.

FIG. 14 is a schematic diagram conceptually showing the basic operation of an adaptive array radio base station.

FIG. 15 is a schematic block diagram showing the configuration of an adaptive array radio base station.

FIG. 16 is a schematic diagram showing a frame configuration of a radio signal of a mobile phone.

FIG. 17 is a schematic diagram showing an image of transmission and reception of a radio signal between an adaptive array radio base station and a user.

[Explanation of symbols]

SR1 receiver, ST1 transmitter, # 1 to # 4 antennas, 10-1 to 10-4 switch circuits, 12-1 to 1
2-4 Multiplier, 13 Adder, 15-1 to 15-4
Multiplier, 20 reception weight vector calculator, 22 reception response vector calculator, 30 transmission weight vector calculator, 32 transmission response vector estimator, 34, 102,
104 memory, 101 response vector estimation circuit, 10
3,105 Correlation calculation and Doppler frequency estimation circuit, 1000 Radio equipment (radio base station).

─────────────────────────────────────────────────── ─── Continuation of front page (58) Fields investigated (Int.Cl. ⁷ , DB name) H04B ⁷ /02-7/12 H04L 1/02-1/06 H01Q 3/00-3/46 H01Q 21 / 00-25/04

Claims (14)

(57) [Claims] 1. A Doppler frequency for estimating a Doppler frequency of a propagation path to a specific terminal in a radio apparatus which changes the antenna directivity in real time and time-divisionally transmits and receives signals to and from a plurality of terminals. An estimation circuit, based on a signal received by a plurality of antennas discretely arranged, a reception signal separation means for separating a signal from the specific terminal among the plurality of terminals, the plurality of Based on the signal received by the antenna, the reception propagation path estimating means for estimating the reception response vector of the propagation path from the specific terminal, and the reception response vector estimated in time by the reception propagation path estimating means. Correlation calculation means for calculating a vector correlation value based on the above, based on the correspondence relationship between the vector correlation value and the Doppler frequency determined empirically in advance, And a estimation means for estimating a Doppler frequency corresponding to the vector correlation value calculated by the function calculating means, a Doppler frequency estimating circuit. 2. The Doppler according to claim 1, wherein the correlation calculating unit includes a calculating unit that calculates an instantaneous correlation value between the reception response vectors that are temporally preceding and following each other and outputs the instantaneous correlation value as the vector correlation value. Frequency estimation circuit. 3. The correlation calculation means calculates a momentary correlation value between reception response vectors that are temporally preceding and following each other, and a past correlation value and a current correlation value calculated by the calculation means. The Doppler frequency estimating circuit according to claim 1, further comprising: an averaging unit that outputs an average value obtained by performing weighted averaging of and as a vector correlation value. 4. The Doppler frequency estimation circuit according to claim 3, wherein the predetermined weight coefficient is set so that a weight for a past correlation value is large and a weight for a current correlation value is small. 5. The correlation calculation means calculates a vector correlation value based on the reception response vector in the slot of the current frame and the reception response vector in the slot of the immediately preceding frame. The Doppler frequency estimation circuit described in. 6. The correlation calculation means calculates a vector correlation value based on the reception response vector in the slot of the current frame and the reception response vector in the most recent slot in which there was no reception error in the slots of the past frames. The Doppler frequency estimation circuit according to claim 1, which is calculated. 7. The Doppler frequency according to claim 1, wherein the correlation calculating means calculates a vector correlation value based on a reception response vector in the first half of the same slot and a reception response vector in the latter half of the same slot. Estimation circuit. 8. A wireless device which changes the antenna directivity in real time and transmits / receives signals to / from a plurality of terminals in a time-division manner, wherein a plurality of discretely arranged antennas are used when transmitting / receiving signals. A receiving circuit, the transmitting circuit and the receiving circuit sharing the plurality of antennas, the receiving circuit, when receiving a reception signal, based on a signal from the plurality of antennas, a signal from a specific terminal among the plurality of terminals A received signal separation means for separating the received signal, based on signals from the plurality of antennas at the time of receiving the received signal, a reception propagation path estimation means for estimating a reception response vector of a propagation path from the specific terminal, Wherein, the transmission circuit, based on the estimation result of the reception propagation path estimation means, a transmission propagation path estimation means for estimating a transmission response vector of a propagation path at the time of transmission of a transmission signal, and Based on the estimation result of the transmission channel estimation means, including a transmission directivity control means for updating the antenna directivity at the time of transmission of the transmission signal, the transmission channel estimation means, by the reception channel estimation means Extrapolation means for calculating the transmission response vector of the downlink slot to the specific terminal by extrapolation processing based on the plurality of reception response vectors of the estimated uplink slot from the specific terminal, and the propagation Doppler frequency estimating means for estimating the Doppler frequency of the path, storage means for holding a plurality of parameters used in the extrapolation process, which is predetermined according to the Doppler frequency of the propagation path, and the held plurality of parameters Of the parameters, the parameter corresponding to the estimated Doppler frequency is selected and applied to the extrapolation processing by the extrapolation means. Means for calculating the vector correlation value based on the reception response vectors temporally preceding and following estimated by the reception channel estimating means, and the Doppler frequency estimating means is empirically determined in advance. And a estimating unit that estimates the Doppler frequency corresponding to the vector correlation value calculated by the correlation calculating unit based on the correspondence relationship between the vector correlation value and the Doppler frequency. 9. The radio according to claim 8, wherein the correlation calculation means includes calculation means for calculating an instantaneous correlation value between the reception response vectors that are temporally preceding and following each other and outputting the instantaneous correlation value as the vector correlation value. apparatus. 10. The correlation calculating means calculates a momentary correlation value between the reception response vectors that are temporally preceding and following each other, and a past correlation value and a current correlation value calculated by the calculating means. 9. The wireless apparatus according to claim 8, further comprising: an averaging unit configured to output an average value obtained by performing weighted averaging of and as a vector correlation value. 11. The radio apparatus according to claim 10, wherein the predetermined weighting coefficient is set so that a weight for a past correlation value is large and a weight for a current correlation value is small. 12. The correlation calculation means calculates a vector correlation value based on the reception response vector in the slot of the current frame and the reception response vector in the slot of the immediately preceding frame. The wireless device according to 1. 13. The correlation calculation means calculates a vector correlation value based on the reception response vector in the slot of the current frame and the reception response vector in the most recent slot in which there was no reception error in the slots of the past frames. The wireless device according to claim 8, which is calculated. 14. The wireless device according to claim 8, wherein the correlation calculation means calculates a vector correlation value based on a reception response vector in the first half of the same slot and a reception response vector in the second half. .