JPH04208707A - Adaptive equalizer - Google Patents

Abstract

PURPOSE:To improve the equalization characteristic by estimating a ratio Eb/No of the signal energy to the noise power density per bit of a reception signal and the fluctuating speed of a transmission line to set the oblivion coefficient of an adaptive equalizer. CONSTITUTION:A power ratio estimating circuit 21 estimates a ratio of the signal energy to the noise power density per bit of the reception signal in accordance with an error signal and sends this ratio to an oblivion coefficient setting circuit 23. A transmission line fluctuating speed designating circuit 22 obtains an average value and the variance of the tap coefficient value of an FF part 1 in the tracking mode to estimate the fluctuating speed of the transmission line and sends it to the oblivion coefficient setting circuit 23. This circuit 23 determines the oblivion coefficient, with which the equalization characteristic of the adaptive equalizer is best in the current state of the transmission line, in accordance with an oblivion coefficient set table and sends its value to a tap coefficient updating circuit 20.

Description

[Detailed Description of the Invention] [Field of Industrial Application] The present invention relates to an adaptive equalizer used in high-speed digital mobile communication for the purpose of suppressing deterioration of transmission characteristics caused by waveform distortion due to frequency selective fading. It is.

(Prior art) One of the types of baseband waveform adaptive equalizers that can follow fast fluctuations in transmission paths such as mobile communications and can initialize the equalizer in a short training process is the "decision feedback" method. "Frequency selective fading compensation characteristics of 6-wave terrestrial communications using a form-adaptive equalizer" (IEICE Journal B-1)
1, Vol.

J72-B-11, No, 10 pp, 513-52
A decision-feedback adaptive equalizer is known, as described in No. 3 (October 1989).

Figure 'tS7 is a block diagram showing the decision feedback type adaptive equalizer as described above. In the figure, (1) is a transversal filter (hereinafter referred to as FF section) of a feedforward section whose tap interval is a constant delay time of p seconds and has L taps;
In (2), the tap interval is a constant delay time T seconds, and the number of taps is (
M−L) transversal filters of the feedback section (hereinafter referred to as FB section), (3) is the FF section (1)
An adder that adds the output data of the FB section (2) to the output data of the FB section (2), (4) a judger that identifies the output signal sequence of the adder (3) every second, and (5) an FF section (1) and FB section (2)
A tap coefficient update circuit that determines the tap coefficient of (
6) is a switch circuit that switches the input signal series of the FB section (2) to the output signal series of the determiner (4) and the known signal series; (7) is the input signal series input terminal of the FF section (1);
) is a known signal sequence input terminal, and (9) is an output signal terminal of the decision feedback adaptive equalizer.

FIG. 8 shows the configuration of a receiving system using the decision feedback adaptive equalizer configured as described above. In the figure, (10) is the receiving antenna, (11) is the detection section, and (12) is the A /D converter, (13) is a frame synchronization circuit, (14) is a clock recovery circuit, (15) is a buffer memory circuit, (16)
is an equalization section including a decision feedback adaptive equalizer shown in FIG.
17) is an output terminal.

Furthermore, FIG. 9 shows an example of a signal burst format used in mobile communications, etc., where (18) is a known signal sequence used for training and frame synchronization of a decision feedback type equalizer, and (19) is the random data part.

Next, the operation will be explained. In FIG. 8, the received wave is taken in by the receiving antenna (lO) and the detection unit (11
) is converted to a baseband signal by an A/D converter (
12) into a digital signal. and,
This signal is taken into the buffer memory circuit (15),
After frame synchronization is performed in the frame synchronization circuit (13), the signal is sent to the equalization section (16).

At this time, the clock recovered by the clock recovery circuit (14) is supplied to the buffer memory circuit (15) and the equalization section (16).

The decision feedback adaptive equalizer shown in FIG. 7 in the equalization section (16) uses a known signal sequence (18) at the beginning of the burst to estimate the characteristics of the transmission path (training mode), and then uses a random data section (training mode) to estimate the characteristics of the transmission path. (19) is equalized (tracking mode). At this time, the input signal sequence of the FB section (2) is inputted by the switch circuit (6) through the known signal sequence input terminal in the training mode, and the output signal sequence of the determiner (4) in the tracking mode. It can be switched to

The tap coefficient updating circuit (5) uses the input signal sequence of the decision feedback adaptive equalizer, the known signal sequence, the output signal sequence of the decider (4), and the output signal sequence of the adder (3), and uses the Kalman filter algorithm ( RLS algorithm), the tap coefficients of the FF section (1) and the FB section (2) are updated for each symbol.

Next, this tap coefficient updating algorithm will be explained. The input signal vector to the equalization amount at time t=nT is X, (n), the tap coefficient is CM (n), the equalizer output is I (n), the desired output is d (n), and the error signal is e
(n). Here, XM (n), CM (n
), I (n), and d (n) are complex numbers indicating in-phase and orthogonal channels. Also, if the number of taps in the FF section (1) of the decision feedback adaptive equalizer is M, and the total number of taps is M, then xM(n) = [y+” (n), y2” (n)
, −,yL” (n).

d, responsibility n), d2'' (n), -, dM-shi responsibility n
)]” (1)cii(n) = [C+”(n),
C2'(n), ・, CM''(r+)] (2)1
(n) =CM″(n-1) XM (n)
(3) e(n/n-1) = d(n) −
1(n) = d(n) -Cii'(n-1) Xm
(n). Here, ('') represents a complex conjugate transposed matrix (or vector). Also, d(n) is a known signal sequence in the training mode, and is an output signal sequence of the determiner (4) in the tracking mode. , tap coefficient CM that minimizes the evaluation interval number ε expressed by the following formula
(n) becomes the value to be sought.

ε=Σ λ0se″(i/n) e (i/n)
(5)+111 Here, person represents the forgetting coefficient (0<λ≦1).

CM (n) that minimizes equation (5) is CM (n)
= R-' (n) D (n)
(6) p-' (n) = R(n) = Σ λ'-
'XM (i) XM'(i)+δλ'I (δ
: positive constant) (7) D(n)−Σ λ'−' XM (i) d' (i
) (8) 1=1, and furthermore, the commercial at time t = (n-1) T
(n-1), P C at time t=nT from (n-1)
The algorithm for recursively finding M (n) is as follows.

K (n) = P (n-1) XM (n) /
[λ+XM”(n) P(n-1) XM(n)]p
(n) = [P (n-1) −K (n) XM''
(n) P (n-1)] (101CM(n)
= Cii (n-1) 10K(n) e-responsibility n/n-1
) (11)P(0)−δ−′I, (:M(
0)=O(12) Here, K (n) is Kalman Keyne, and P (n) is the estimation error covariance matrix of the tap coefficients.

The forgetting coefficient in the above algorithm is a parameter that determines the followability of the algorithm to transmission path fluctuations. By setting λ to 1, the influence of past data will decrease exponentially, and λ By making it small, it becomes possible to follow fast fluctuations in the transmission path. However, as λ decreases, the amount of data used to estimate CM (n) decreases, and the equalization error increases.

As a result, depending on the speed of fluctuation in the transmission path and the ratio Eb/No between the signal energy per bit of the received signal and the noise power density, there is a forgetting coefficient λ that optimizes the equalization characteristics. Figure 1θ shows the ratio of the signal energy per bit of the received signal to the noise power density Eb/No, and Figure 11 shows the error in the forgetting coefficient λ when the Doppler frequency f, which represents the fluctuation speed of the transmission path, and i are used as parameters. An example of changes in rate characteristics is shown. As shown in Figure 1O, the ratio of the signal energy per bit of the received signal to the noise power density Eb
It can be seen that the lower the /No, the larger the optimal forgetting coefficient λ becomes, and as shown in FIG. 11, the larger the Doppler frequency f, the smaller the optimal forgetting coefficient λ becomes.

(Problem to be Solved by the Invention) However, in the conventional adaptive isolator, the value of the above-mentioned forgetting coefficient λ is set to a constant value. When the ratio Eb/No of noise power density and the Doppler frequency fd vary greatly depending on usage conditions, there is a problem that optimum equalization characteristics cannot be obtained.

This invention was made to solve the above-mentioned problems, and even when the state of the transmission path changes greatly depending on the usage situation, the signal energy per bit of the received signal representing the state of the transmission path can be reduced. and noise power density ratio Eb/No
, the Doppler frequency f, and select the forgetting coefficient λ in the tap coefficient updating algorithm according to its state, with the aim of obtaining an adaptive equalizer with the best equalization characteristics for the transmission path. There is.

[Means to solve the problem]

The adaptive equalizer according to the present invention has a tapped delay circuit, and is configured to multiply each tap output of the tapped delay circuit by a tap coefficient and add the result, and to equalize a received wave based on the addition result. The configured adaptive equalizer includes a transmission path fluctuation speed estimation circuit that estimates the fluctuation speed of the transmission path from the cup coefficient value of the filter, and a transmission path fluctuation speed estimation circuit that estimates the fluctuation speed of the transmission path from the cup coefficient value of the filter, and the sum of products of the predetermined information value of the received signal, the tap coefficient value, and the input signal. A power ratio estimating circuit that estimates the power ratio of the received signal and noise from the error that is the difference, and a transmission path fluctuation speed estimation value by the transmission path fluctuation speed estimation circuit and a power ratio estimation value by the power ratio estimation circuit. A forgetting coefficient setting circuit that calculates a weighting coefficient with respect to time, and a tap coefficient update that updates the tap coefficient of a filter based on the forgetting coefficient and the difference between the value of predetermined information of the received signal, the tap coefficient value, and the sum of products of the input signal. It is equipped with a circuit.

Further, the adaptive equalizer according to the present invention includes a field strength measuring circuit that measures the electric field strength of a received signal, a transmission path fluctuation speed estimation circuit that estimates the fluctuation speed of the transmission path based on the measured value, and a field strength measuring circuit that measures the electric field strength of the received signal. A forgetting coefficient setting circuit that calculates a weighting coefficient for data over time using the measurement value of the intensity measurement circuit and the estimated transmission path fluctuation speed, and equalization of the received signal using the forgetting coefficient set by the forgetting coefficient setting circuit. It is equipped with an equalization section that performs the following.

[Effect]

The adaptive equalizer in this invention uses the tap coefficient CM and the error signal e, which are the calculation results of the adaptive equalizer, to estimate the speed of fluctuation of the transmission path from the fluctuation of the tap coefficient Ci, and also By estimating the ratio of signal energy per bit to noise power density Eb/No from the value of the error signal e, the state of the transmission path can be grasped, and the equalization characteristics of the adaptive equalizer can be adjusted according to the state. The equalization properties are improved by selecting the forgetting coefficient that seems to be the best.

Further, the adaptive equalizer according to the present invention estimates the speed of transmission path fluctuation by estimating the transmission path fluctuation speed based on the measurement value by the electric field strength measurement circuit, so that the equalization characteristics of the adaptive equalizer are maximized. The equalization properties are improved by selecting a forgetting factor that is likely to improve.

〔Example〕

An embodiment of the present invention will be described below with reference to the drawings. 1st
The figure shows a configuration diagram of the adaptive equalizer according to the present embodiment, and the same parts as in FIG. 7 are given the same reference numerals and the explanation thereof will be omitted. In FIG. 1, as a new configuration, (20) is a tap coefficient updating circuit, and (21) is this tap coefficient updating circuit (20).
) is a power ratio estimation circuit whose input signal is an error signal which is one of the calculation results, and (22) is a tap coefficient update circuit (20).
(23) is the power ratio estimation circuit (23) which uses the tap coefficient value which is the calculation result of
21) and a forgetting coefficient setting circuit that manually calculates the calculation results of the transmission path fluctuation speed estimation circuit (22), and the transmission path fluctuation speed estimation circuit (22) has a variance value 61 as shown in FIG.
The correspondence table of the Doppler frequency fd to ~δ6 is also stored in the forgetting coefficient setting circuit (23), and the ratio Eb of the signal energy per bit of the received signal to the noise power density is stored in the forgetting coefficient setting circuit (23).
/No and the forgetting coefficient E1 corresponding to the Doppler frequency fd
-λ13 correspondence tables are provided, respectively.

Next, the operation will be explained with reference to the flowchart shown in FIG. As explained in the conventional technology section, the tap coefficient updating circuit (2o) uses the error signal e (n/n-1) in equation (4) and the tap coefficient C in equation (11).
M (n) is calculated (Sl). The tap coefficient update circuit (20) uses the error signal e (n/n-1) obtained for each symbol time of the received signal and the tap coefficient CM (n) to the power ratio estimation circuit (21) to estimate the transmission path fluctuation speed. Circuit (2
2) respectively. The power ratio estimating circuit (21) estimates the ratio Eb/No of the signal energy per bit of the received signal to the noise power density from the error signal.

Kalman filter algorithm (RLS algorithm)
When estimating tap coefficient values using Since the ratio Eb/No value is obtained, this power ratio estimating circuit (21) calculates the average value of the errors of the last five symbols in the training mode and uses that value as the Eb/No estimated value.
3) (52-54).

In the transmission path fluctuation speed estimating circuit (22), the average value of the tap coefficient values of the FF section (1) in the tracking mode is calculated, and then the absolute value of the average value is calculated with respect to the closest tap coefficient value. The dispersion is determined, and the amount of change in the transmission path during the burst is determined quantitatively. Then, the fluctuation speed of the transmission path is estimated based on the correspondence table between the dispersion value and the Doppler frequency fd determined by computer simulation as shown in Fig. 2, and is sent to the forgetting coefficient setting circuit (23) (5
5-S9).

The forgetting coefficient setting circuit (23) determines the forgetting coefficient that gives the best equalization characteristic of the adaptive equalizer under the transmission path condition, according to the forgetting coefficient setting table determined by computer simulation as shown in Fig. 3. is determined and the value is sent to the tap coefficient update circuit (2o) (S 10 ).

511).

The tap coefficient update circuit (20) uses this forgetting coefficient value to calculate the tap coefficient update algorithm for the next burst received signal sequence (Sll-5ll).
. In addition, in a normal high-speed digital automobile ML system, this burst interval is about 20 m5ec, so optimal forgetting is achieved due to fluctuations in the ratio Eb/No of the signal energy per bit of the received signal and noise power density and changes in the transmission path fluctuation speed during that time. Since the change in the coefficient is considered to be sufficiently small compared to the setting accuracy of the forgetting coefficient, there is no problem in using the forgetting coefficient determined from the characteristics of the transmission path at the time of reception of the previous burst for the next burst.

In the above embodiment, the ratio Eb/No of the signal energy per bit of the received signal to the noise power density and the transmission path fluctuation speed are estimated from the calculation results of the tap coefficient updating circuit (20). may be performed using the output results of other circuits. Such an embodiment is shown below with reference to FIG. FIG. 5 shows the configuration of a receiver similar to FIG. 8,
Identical parts are given the same reference numerals and their explanations will be omitted. Fifth
In the figure, (24) is an electric field strength measurement circuit, (25) is a detection unit including the electric field strength measurement circuit (24), and (26) is a transmission path fluctuation speed using the output signal of the electric field strength measurement circuit (24) as human power. The estimation circuit (27) is the electric field strength measurement circuit (24
) and the output signal of the transmission line fluctuation speed estimation circuit (26) as input signals, and a forgetting coefficient setting circuit that selects a forgetting coefficient based on the state of the transmission line, (28) is a forgetting coefficient setting circuit (27). This is an equalization unit that equalizes the received input signal using the forgetting coefficient set in , according to the Kalman filter algorithm shown in the conventional example.

Next, the operation in the above embodiment will be explained with reference to the flowchart shown in FIG. Electric field strength measurement circuit (24)
Received signal level detectors such as Assuming that Eb/No is proportional to the received electric field strength, the ratio of the signal energy per bit of the received signal to the noise power density Eb/No
Estimate.

The transmission path fluctuation speed estimation circuit (2B) estimates the speed of transmission path fluctuation by counting the number of times the output signal of the electric field strength measurement circuit (24) crosses a certain constant bell per unit time, and converts it into a Doppler signal. After expressing the frequency fd, the value is transmitted to the forgetting coefficient setting circuit (27) (Sll~
515).

The forgetting coefficient setting circuit (27) sets the forgetting coefficient from the ratio Eb/No of the signal energy per bit of the received signal to the noise power density and the estimated value of the Doppler frequency fd in the same manner as in the above embodiment. , the value is equalized by the equalizer (28)
Send to (516-519).

In the above embodiment, the case of a decision feedback adaptive equalizer is shown, but the FB section (2), adder (3), decider (4), switch circuit (6), known signal sequence input terminal (
A case of adaptive equalization without 8) may also be used.

(Effects of the Invention) As described above, according to the present invention, the ratio Eb/No of the signal energy per hit of the received signal to the noise power density and the fluctuation speed of the transmission path are estimated, and the fluctuation speed of the transmission path is estimated. Since the configuration is such that the forgetting coefficient of the adaptive equalizer is set so that the equalization characteristics of the adaptive equalizer are the best, an adaptive equalizer with good equalization characteristics can be obtained.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of an adaptive equalizer according to an embodiment of the present invention, FIG. 2 is an explanatory diagram of a Doppler frequency estimation table provided in the transmission path fluctuation speed estimation circuit (22) of FIG. 1, and FIG. 1 is an explanatory diagram of a forgetting coefficient setting table provided in the forgetting coefficient setting circuit (23) of FIG. 1, and FIG. 4 is an illustration of the overall processing flow of an embodiment of FIG. 1 and the internal operation of each circuit. 5 is a block diagram showing another embodiment of the present invention. FIG. 6 is a flowchart showing the overall processing flow and internal operation of each circuit in the embodiment of FIG. 5.
The figure is a block diagram of a conventional decision feedback adaptive equalizer, and FIG. 8 is a block diagram of a receiver using the decision feedback adaptive equalizer of FIG.
FIG. 9 is a diagram of the burst format of the received signal, and FIG. 10 is the ratio of signal energy per bit of the received signal to noise power density Eb/N. Fig. 11 shows the error rate characteristic curve for the forgetting coefficient of the decision feedback adaptive equalizer when the parameter is taken as a parameter. It is a target line diagram. (20) is a tap coefficient update circuit, (21) is a power ratio estimation circuit, (22) is a transmission path fluctuation speed estimation circuit, (23) is a forgetting coefficient setting circuit, (24) is an electric field strength measurement circuit, (2
6) is the transmission line fluctuation speed estimation circuit, (27) is the forgetting coefficient setting circuit, and (28) is the northern part. In addition, in the figures, the same reference numerals indicate the same or equivalent parts. Agent So Aki Yamazaki --1"1~-2" @1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 10 Figure 11 Figure 1N6 Figure 7-8 rl! J 13 Figure 9

Claims (2)

[Claims] (1) An adaptive equalizer having a tapped delay circuit, configured to multiply each tap output of the tapped delay circuit by a tap coefficient and add the result, and equalize the received wave based on the addition result. A transmission path fluctuation speed estimating circuit estimates the fluctuation speed of the transmission path from the cup coefficient value of the filter, and a transmission path fluctuation speed estimation circuit that estimates the fluctuation speed of the transmission path from the cup coefficient value of the filter, and the reception signal is estimated from the error that is the difference between the value of predetermined information of the reception signal, the tap coefficient value, and the sum of products of the input signal. and a power ratio estimating circuit for estimating the power ratio of the noise to The present invention includes a coefficient setting circuit, and a tap coefficient updating circuit that updates the tap coefficients of the filter based on the difference between the value of predetermined information of the received signal, the tap coefficient value, and the sum of products of the input signal, and the forgetting coefficient. Features an adaptive equalizer. (2) A field strength measuring circuit that measures the electric field strength of the received signal, a transmission path fluctuation speed estimation circuit that estimates the fluctuation speed of the transmission path based on the measured value, and the measured value of the field strength measuring circuit and the transmission a forgetting coefficient setting circuit that calculates a weighting coefficient for time of data using the estimated road fluctuation speed;
An adaptive equalizer comprising: an equalizer that equalizes a received signal using the forgetting coefficient set by the forgetting coefficient setting circuit.