JPH07336278A - Propagation path estimation unit - Google Patents

Abstract

PURPOSE:To output a high-precision propagation path estimated value which is not delayed behind input data, containing a noise, as a desired signal on which propagation path characteristics are reflected and has a noise component suppressed. CONSTITUTION:The value X(X0) of input data at current time on which the propagation path characteristics are reflected is frequently different from a desired value because of the presence of the noise. For the purpose, a method of least squares which uses an approximation function is applied to the input data X(0) at the current time and M input data series X(0)-X(M-1) of input data X(i) in total which are (i) samples precedent to the current time on the basis of the fact that the time-series conversion of the desired value to determine an arithmetic expression for obtaining the desired value (propagation path estimated value W) at the current time. Then the propagation path estimation unit 10 is constituted on the basis of this expression.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a propagation path estimator and is applicable to, for example, a receiver of a mobile communication system.

[0002]

2. Description of the Related Art In a mobile communication system, its propagation path characteristics change as the mobile moves.
Communication between two fixed points is affected by changes in channel characteristics. Therefore, the receiving apparatus is equipped with a propagation path estimator to obtain an estimated value of the propagation path characteristic and compensate the influence of the propagation path characteristic from the received signal.

By the way, recently, in order to obtain a good estimated value of the channel characteristic, the transmitting side moves a signal for estimating the channel characteristic (pilot signal) separately from the transmission data. Body communication systems have been proposed. For example, a predetermined spreading code sequence (eg PN sequence)
2 is PSK-modulated and is transmitted as a pilot signal in a channel of a CDMA system different from the transmission data. Therefore, if a correlation between such a pilot signal and a predetermined spreading code sequence created on the receiving side in synchronization with this is obtained, a signal in which noise is added to the signal component (desired signal) reflecting the propagation path characteristics Only you get.

From the input signal in which noise is added to the desired signal,
As a transmission path estimator that suppresses noise and improves the S / N ratio of a desired signal, conventionally, it is conceivable to apply a low-pass filter having a non-recursive digital filter configuration described in the following document.

Reference: Oppenheim, Schafer, translated by Gen Gen,
"Digital Signal Processing (above)", Corona Publishing, First Edition 6th
Press, 1984, pp. 237-249. Use of this low-pass filter suppresses noise components located in a frequency band higher than that of the desired signal and improves the S / N ratio of the input signal, that is, the desired signal. You can That is, it is considered possible to obtain an estimated signal of the propagation path characteristic that well reflects the propagation path characteristic.

[0006]

However, in the above-mentioned low-pass filter, when the number of taps is N (N is an odd number), the output signal is delayed from the input signal by (N-1) / 2 samples.

Therefore, as a propagation path estimator of a receiver in a mobile communication system that requires real-time processing, it is difficult to apply due to this delay time. That is, in consideration of the delay time, other processes also have to be delayed according to the delay time, and the whole process is delayed by that amount, so that the real-time property is somewhat deteriorated.

Incidentally, in the case of the receiver of the mobile communication system for obtaining the estimated value signal of the transmission path characteristic from the modulated signal of the transmission data, it may be a little different in the transmission path estimator in relation to the processing of the demodulation processing system main body. Although some delays do not pose a problem (see, for example, Japanese Patent Application No. 6-20965 and the drawings), a receiver of a mobile communication system that obtains an estimated value signal of a transmission path characteristic from a pilot signal estimates the transmission path characteristic. There is preferably no delay in the value signal.

[0009]

In order to solve such a problem, a first aspect of the present invention provides a channel estimator which receives a data sequence reflecting channel characteristics and outputs an estimated value of the channel characteristics. Input data X i samples before the current time
(I) (where i is 0 to M−1 and i is 0, the current time data), and a total of M input data series X (0) to X (M
-1), as the propagation path estimation value W at the current time, the following (1)
It is characterized in that the value obtained by the formula is output.

W = {− 6ΣiX (i) + (4M−2) ΣX (i)} ÷ {M (M + 1)} (1) The second aspect of the present invention is a data sequence reflecting the propagation path characteristics. In a channel estimator that receives an input and outputs an estimated value of channel characteristics, input data X i samples before the current time.
(I) (where i is 0 to M−1 and i is 0, the current time data), and a total of M input data series X (0) to X (M
-1), as the propagation path estimation value W at the current time, the following (2)
It is characterized in that the value obtained by the formula is output.

W = [Σ {30 · i ² + (− 36M + 18) · i + 9M ² −9M + 6} · X (i)] ÷ {M (M + 1) (M + 2)} (2) (However, the sum Σ is i Is about 0-M-1)

[0012]

The value X (0) of the input data at the current time, which reflects the channel characteristics, is often different from the desired value due to the presence of noise. Therefore, according to the first and second aspects of the present invention, the time series change (propagation path characteristic change) of the desired value can be approximated to a predetermined function for a short time, based on the input data X (0) at the current time, Input data X i samples before the current time
(I) A total of M input data series X (0) to X (M-
The least squares method using a predetermined approximation function is applied to 1) to obtain the desired value at the current time, that is, the propagation path estimated value W.

In the first aspect of the present invention, the least squares method is applied on the basis of the assumption that the propagation path characteristic change is "a linear function of time". Under this assumption, the propagation path estimated value W is (1). It is based on what is given in the formula. On the other hand, the second invention is
The change in the propagation path characteristics is obtained by applying the least squares method based on the assumption of "quadratic function of time". Under this assumption, the propagation path estimated value W is based on the equation (2).

Thus, the value X of the input data at the current time
Since (0) is used to estimate the propagation channel estimated value W at the current time, the estimation accuracy can be improved, and the obtained propagation channel estimated value W is at the current time and is not used in the estimation process. There is no delay. Therefore, it is possible to realize a channel estimator suitable for real-time processing and capable of obtaining an accurate estimated value even if the channel characteristic changes rapidly.

[0015]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A first embodiment of a propagation path estimator according to the present invention will be described in detail below with reference to the drawings. The propagation path estimator of this embodiment will be described as being installed in a receiving device of a mobile communication system.

(A) Principle of estimation of propagation path characteristics of the first embodiment In a mobile communication system including a receiving device equipped with the transmission path estimator of the first embodiment, other than transmission data on the transmission side, In addition, a predetermined spreading code sequence (for example, a PN sequence) is two-phase PSK-modulated and transmitted as a pilot signal for estimating a channel characteristic.

Therefore, on the receiving side, the propagation path characteristics are estimated using this pilot signal. When the in-phase component signal of the received baseband signal of the pilot signal is the I signal and the quadrature component signal is the Q signal, the spread code sequence (the same code sequence as the transmitted code) and the I signal are synchronized with the received signal. Considering a complex correlation signal in which the correlation of is the real part and the correlation with the Q signal is the imaginary part, this signal r is a signal that reflects the propagation path characteristics as in equation (3).

R = β · exp (jφ) + d (3) where β, φ, and d are the amplitude characteristic of the propagation path, the phase characteristic of the propagation path, and noise, respectively. Furthermore, exp (jφ)
As is well known, is an exponential expression of cos φ + j sin φ. As shown in equation (3), the real part and the imaginary part of the complex correlation signal r are signals in which noise is added to the desired signal reflecting the propagation path characteristics.

In the first embodiment, it is assumed that the time change of the desired signal β · exp (jφ) reflecting the propagation path characteristic is a linear function of time in a short time section. In addition, the desired signal β · exp (j
The noise d superimposed on φ) is 0, which changes randomly.
Suppose that.

Under these assumptions, the desired signal is a digital signal, the value of the desired signal at the current time (hereinafter referred to as the desired value) is S (0), and the desired value i samples before the current time is S.
Assuming (i), the desired value S (i) can be expressed by the equation (4) using the linear parameters V and W.

S (i) = V · i + W (4) Therefore, if the linear function parameters V and W (variable in the long term, but fixed values in the assumed short term) are optimally obtained. , The desired value S (0) of the propagation path characteristic at the current time can be estimated by the value of the parameter W. To determine the linear function parameters V and W, the formula (4) is used as the principle formula, and the input data X (0) at the current time (the value of the signal r reflecting the transmission line characteristics and the desired value if there is no noise) Equal to S (0)),
Input data X (i) i samples before the current time (however,
i is 1 to M-1) and a total of M input data series X (0) to
It can be determined by the method of least squares using X (M-1). That is, the parameters V and W can be obtained by solving the simultaneous equations shown in equation (5),
The obtained parameter W (therefore, the desired value S (0) reflecting the propagation path characteristic at the current time) can be expressed by the equation (6).

[0022]

[Equation 1] W = {− 6ΣiX (i) + (4M−2) ΣX (i)} ÷ {M (M + 1)} (6) (However, the sum Σ in the equations (5) and (6) is i = 0 to M
−1; the sum Σ obtained below is i
The same applies to the above). Considering the value A (i) shown in the equation (7), the obtained parameter W (hence, the desired value S (0) at the current time) is
It can be expressed by equation (8). Note that the sum total of the equation (8) is for i from 0 to M-1.

A (i) = {(− 6i + 4M−2) / M (M + 1)} (7) W = ΣA (i) · X (i) (8) Then, the propagation path of the first embodiment. In the estimator, the value S (0) of the propagation path characteristic at the current time is estimated by the value W obtained according to the equation (8).

FIG. 2 is an explanatory diagram of such an estimation principle. The estimation method according to the above equation (8) is performed by, for example, input data series X
From (0) to X (9), the linear function S at that time interval
(I) is approximated by the least-squares method, the value W = S (0) at the current time 0 in the approximation function S (i) is obtained, and this value W is reflected in the propagation path characteristics and noise is removed. It is the estimation of the desired value and the equalization.

(B) Device Configuration Including the Propagation Path Estimator of the First Embodiment FIG. 3 shows an example of the configuration of a device including the propagation path estimator of the first embodiment. The position and function of the propagation path estimator in the device will be described below by describing such a device.

The in-phase component signal (I signal) of the received baseband signal of the pilot signal, which is input from the received signal input terminal 11, is given to the synchronization acquisition device 2 and the correlator 3. In addition, the quadrature component signal (Q
Signal) is also provided to the synchronization acquisition device 2 and the correlator 3.

From the I signal and the Q signal, the synchronization acquisition device 2
The synchronization timing of the received signal is captured and the timing signal is output to the correlator 3.

The correlator 3 internally generates a spread code sequence synchronized with the received signal on the basis of the timing signal from the synchronism capturer 2 and combines the I signal input from the received signal input terminal 11 and the spread code sequence. The correlation (I correlation) is calculated, and the correlation (Q correlation) between the Q signal input from the reception signal input terminal 12 and the spread code sequence is calculated and output. The correlation calculation for each signal is the same, and is a product-sum calculation for a predetermined correlation section T (for example, 1-bit section). The correlator 3 outputs the correlation result for each correlation section T.
The I and Q correlation signals from the correlator 3 are
It is given to the corresponding channel estimator 41 and channel estimator 42.

The propagation channel estimator 41 and the propagation channel estimator 42 are the propagation channel estimators of the first embodiment, and have the detailed configuration shown in FIG. 1 described later.

The propagation path estimator 41 estimates the desired value at the current time of the in-phase component of the propagation path characteristic (I propagation path characteristic) from the I correlation signal (data sequence) from the correlator 3 and outputs the propagation path characteristic output terminal. It is output to a data demodulation processing unit (not shown) via 51. Similarly, the channel estimator 42 estimates the desired value at the current time of the orthogonal component (Q channel characteristic) of the channel characteristic from the Q correlation signal (data series) from the correlator 3 and outputs the channel characteristic output terminal. It is output to a data demodulation processing unit (not shown) via 52.

A data demodulation processing unit (not shown) performs data demodulation processing by compensating for the influence of the propagation path characteristic on the received data, using the given estimated value of the propagation path characteristic.

As will be described later, the propagation path estimator of this embodiment can obtain an estimated value of the propagation path characteristic at the current time without delay, and moreover, the propagation path is estimated based on the pilot signal irrelevant to the transmission data. Since the characteristics are estimated, it is easy to provide the data demodulation processing unit with the estimated value of the channel characteristics that is equal to the time of the reception data related to the demodulation processing. Incidentally, in a communication device that obtains an estimated value of a channel characteristic by appropriately processing a modulated signal of transmission data, the estimated value of the channel characteristic is obtained after removing the data component from the received signal,
Since the demodulation output in the past is also used to remove the data component, the input data at the estimated time cannot be input to the channel estimator, and the channel characteristic estimation is delayed. for that reason,
Time adjustment with the data demodulation processing unit is required.

(C) Channel Estimator of First Embodiment FIG. 1 shows a channel estimator 10 (41, 42) of the first embodiment.
It shows the configuration of.

The propagation path estimator 10 of the first embodiment is based on the above-mentioned principle. That is, the channel characteristics are estimated according to the equation (8). There are various methods such as software as a method of estimating the propagation path characteristic according to the equation (8), but the propagation path estimator 10 of the first embodiment has a non-recursive filter (transversal filter) configuration The estimated value of the propagation path characteristic is obtained according to Eq. (8).

That is, as shown in FIG. 1, the propagation path estimator 10 has M-1 memory elements 61 to 1 for one sampling period delay connected in cascade in which the input data series X (i) is input. 6M-1 (these are all shift registers), input data X (0) at the current time fetched in parallel by these memory elements 61 to 6M-1, and input data X i samples before the current time. (I) (where i is 1 to M-1), a total of M pieces of input data X (0) to X (M-
It is composed of M coefficient multipliers 70 to 7M-1 for weighting 1) and a summing device 8 for calculating the sum of the weighted values.

Here, the weighting coefficient (tap coefficient) A (i) of each coefficient multiplier 7i is a value {(-6i) shown in the equation (7).
+ 4M-2) / M (M + 1)} is selected, and thus the summation unit 8 outputs the estimated value W of the desired value at the current time shown in the equation (8).

The number of taps M is determined in consideration of the time when the desired value change of the propagation path characteristic can be approximated by a linear function,
Any number may be used as long as this point is satisfied.

Therefore, according to the propagation path estimator 10 of the first embodiment, the estimated value W of the desired value S (0) that reflects the propagation path characteristic at the current time using the input data X (0) at the current time is used. Can be obtained without causing a delay in the estimation process.

Further, according to the propagation path estimator 10 of the first embodiment, since the estimation is also made by using the input data X (0) at the current time, the conventional estimation which is made only from the past input data. The estimation accuracy can be improved as compared with the instrument.

Further, according to the channel estimator 10 of the first embodiment, since the weighting coefficient A (i) of the coefficient multiplier 7i is fixed, the configuration is simpler than that of the adaptive channel estimator. it can.

(D) Propagation Path Estimator of Second Embodiment Next, a second embodiment of the propagation path estimator according to the present invention will be described. The propagation path estimator of the second embodiment is also realized by a non-recursive filter (transversal filter) configuration. Although the number of taps is generally larger than that in the first embodiment, it will be described here with reference to FIG. 1 according to the first embodiment with M taps as in the first embodiment. Further, in the second embodiment, the weighting coefficient (tap coefficient) of the coefficient multiplier 7i is represented by B (i).

In the second embodiment, the weighting coefficient (tap coefficient) B (i) of each coefficient multiplier 7i is selected to the value shown in the following equation (9), and the summation unit 8 is changed to the equation (10). The difference from the first embodiment is that the estimated value W of the desired value at the present time is output.

B (i) = {30i ² + (− 36M + 18) i + 9M ² −9M + 6} ÷ {M (M + 1) (M + 2)} (9) W = ΣB (i) · X (i) (10) This difference is because the second embodiment follows the following estimation principle, and the estimation principle adopted by the second embodiment will be described below.

In the first embodiment, assuming that "the time change of the desired signal is represented by a linear function of time in consideration of a short time section", the case where the approximate function is represented by a linear function has been described. . On the other hand, in the second embodiment, it is assumed that "the time change of the desired signal is represented by a quadratic function of time in consideration of a short time section", and the approximation function is represented by a quadratic function. It is the one. In addition, it is 1 for too short time
Although it can be approximated to a quadratic function, it cannot be approximated to a quadratic function, and the time interval in which such an assumption holds is generally longer than that of the first embodiment.

In this case, the quadratic function which starts from the present time and approximates the desired signal S (i) past the present time is expressed by the equation (11) using the quadratic function parameters U, V and W. It

S (i) = U · i ² + V · i + W (11) Here, similarly to the first embodiment, if the parameters U, V, and W are optimally obtained, the current time (i = 0) Desired signal S (0)
Can be estimated by the value of the parameter W. Further, in order to obtain the optimum parameters U, V, W, as in the first embodiment,
M input data X including the input data X (0) at the current time
The least squares method may be applied to (0) to X (M-1). In this case, the parameter U,
To obtain V and W, the simultaneous equations shown in equation (12) should be solved.

[0047]

[Equation 2] When this simultaneous equation is solved, the parameter W is expressed by equation (13), and by transforming this equation, equation (14) is obtained.

W = {30Σi ² · X (i) + (− 36M + 18) Σi · X (i) + (9M ² −9M + 6) ΣX (i)} ÷ {M (M + 1) (M + 2)} (13) W = [Σ {30 · i ² + (− 36M + 18) · i + 9M ² −9M + 6} · X (i)] ÷ {M (M + 1) (M + 2)} (14) This equation (14) is the above (9). ) By introducing the value B (i) according to the equation, it can be expressed by the product-sum operation shown in the equation (10),
As a result, the calculation shown in Eq. (14) can be executed by the non-recursive filter, and "the time change of the desired signal of the channel characteristics is expressed by a quadratic function of time when considering a short time section". Under the assumption, the estimated value W of the desired signal S (0) at the current time can be obtained.

Therefore, the channel estimator of the second embodiment also uses the input data X (0) at the current time to obtain the estimated value W of the desired value S (0) reflecting the channel characteristics at the current time. Therefore, there is no delay in the estimation process.
Further, since the input data X (0) at the current time is also used for the estimation, the estimation accuracy can be improved as compared with the conventional estimator which estimates only from the past input data. Furthermore, since the weighting coefficient B (i) of the coefficient multiplier 7i is fixed, the configuration can be simplified as compared with the adaptive channel estimator.

When the second embodiment is compared with the first embodiment, the time interval at which the assumption is established becomes long and the input data to be processed becomes large, but the approximation accuracy becomes high and the estimation accuracy becomes high. There is a difference that it will increase.

(E) Other Embodiments In the above embodiments, the propagation path estimator has the non-recursive filter configuration shown in FIG. 1. However, the present invention is not limited to this.
Any value can be used as long as the value of the expression (6) or the expression (14) can be obtained, and the structure is not limited to that shown in FIG.

The configuration of the propagation path estimator of the above embodiment can be applied not only to the mobile communication receiving device but also to other wireless receiving devices.

[0053]

As described above, according to the present invention, the current time is used as a starting point, and an approximation function approximating a desired signal reflecting the propagation path characteristics past the current time is assumed, and the propagation path is estimated at the current time. Since the value at the current time of this approximation function is output as a value, there is no delay with respect to the input data in which the desired signal that reflects the propagation path characteristics contains noise, and the noise component is suppressed with high accuracy. It is possible to output the estimated value of the channel.

[Brief description of drawings]

FIG. 1 is a block diagram showing a detailed configuration of a propagation path estimator of a first embodiment.

FIG. 2 is an explanatory diagram of an estimation principle of the first embodiment.

FIG. 3 is a block diagram showing a configuration of an on-board device of the propagation path estimator of the first embodiment. is there.

[Explanation of symbols]

61 to 6M-1 ... Memory element for delay (the whole constitutes a shift register), 70 to 7M-1 ... Coefficient multiplier, 8 ... Summer, 10 ... Propagation path estimator.

Claims (4)

[Claims] 1. A channel estimator which receives a data sequence reflecting channel characteristics and outputs an estimated value of channel characteristics, wherein input data X (i) (where
i is 0 to M-1, and when i is 0, the total of M input data series X (0) to X (M-1) (current time data) is used as the propagation path estimation value W at the current time, A propagation path estimator characterized by outputting the value obtained by the following equation (A). W = {− 6ΣiX (i) + (4M−2) ΣX (i)} ÷ {M (M + 1)} (A) (however, for the sum Σ, i is 0 to M−1) 2. A shift register having M taps and a corresponding tap output X (i) of this shift register are input, and this tap output X (i) is expressed by the following equation (B). M coefficient multipliers that multiply the coefficient A (i), and A (i) = (− 6i + 4M−2) / {M (M + 1)} (B) Obtain the sum of outputs from all coefficient multipliers The propagation path estimator according to claim 1, comprising a summer. 3. A channel estimator which receives a data sequence reflecting channel characteristics and outputs an estimated value of channel characteristics, wherein input data X (i) (where
i is 0 to M-1, and when i is 0, the total of M input data series X (0) to X (M-1) (current time data) is used as the propagation path estimation value W at the current time, A propagation path estimator characterized by outputting the value obtained by the following equation (C). W = [Σ {30 · i ² + (− 36M + 18) · i + 9M ² −9M + 6} · X (i)] ÷ {M (M + 1) (M + 2)} (C) (However, the sum Σ is 0 to i About M-1) 4. A shift register having M taps and a corresponding tap output X (i) of this shift register are input, and this tap output X (i) is expressed by the following equation (D). M coefficient multipliers that multiply the coefficient B (i), and B (i) = {30i ² + (-36M + 18) i + 9M ² -9M + 6} ÷ {M (M + 1) (M + 2)} (D) 4. The propagation path estimator according to claim 3, further comprising a summing device that calculates a sum of outputs from the coefficient multipliers.