JPH01291530A - Decision feedback type equalizer - Google Patents

Abstract

PURPOSE:To obtain a satisfactory reception characteristic even when the center tap of a decision feedback type equalizer is set at the arbitrary position of an impulse response on a communication path by including a reception signal before a k-th signal in a pre-transversal filter at a time when the k-th transmission symbol of a decision apparatus is decided. CONSTITUTION:The reception signal sampled at one symbol interval at an input terminal 10 is filtered at the pre-transversal filter, and is inputted to a differential circuit 140. The differential circuit 140 takes the difference of the output of the pre-transversal filter and that of a post-transversal filter, and outputs an obtained signal to the decision apparatus 100, and the decision apparatus 100 performs decision based on the the positive or negative condition of the differential circuit 140, and outputs a decision result to an output terminal 11 and a delay element 105, and delays an input signal by one symbol time. Since a k-th reception signal r(k) is included in a pre-filter at the time when a symbol a(k+1) transmitted at a (k+1)th order is decided by the decision apparatus 100, it is possible to comprise the pre-transversal filter capable of obtaining a true least mean square error.

Description

[Detailed Description of the Invention] (Industrial Application Field) The present invention relates to a decision feedback equalizer that reproduces a transmitted signal without distortion from a received signal distorted by code amount interference in a communication channel. .

(Prior Art) Conventionally, at the time when the th transmission symbol is judged in the judgment device, in the pre-transversal filter,
Decision feedback equalizers that do not include received signals before the k-th are known (for example, Broakis, Digital Communications, McGraw-Hill, 1983).
.

(Problems to be Solved by the Invention) However, in the decision feedback equalizer described in the section (Prior Art), in order to obtain good reception characteristics, it is necessary to set the center tap of the decision feedback equalizer. , it must be set at the optimum position for the communication channel's pulse response.

(For example, Ueda, Suzuki, ``Structure and characteristics of decision feedback equalizer for high-speed digital mobile communications'', 1986 Institute of Electronics, Information and Communication Engineers Spring National Conference Communication Group Lecture Proceedings (Volume 1), B- 724).

(Means for solving the problem) In the present invention, a pre-transversal filter consisting of N taps receives a received signal as input, and an output signal from a determiner is input, and the input signal is input for one symbol time T. A delay element to be delayed, a post-transversal filter consisting of M taps whose input is the output of the delay element, and a difference between the output of the pre-transversal filter and the output of the post-transversal filter. A differential circuit,
In a decision feedback equalizer having a decider that receives the difference circuit output as input, at a time when the decider decides the k-th transmitted symbol, the pre-transversal filter contains the k-th and earlier received Contains signals.

(Function) When the impulse response of the communication channel, the signal-to-noise ratio of the reception signal (K), and the setting position of the center tap of the decision feedback equalizer for the impulse response of the communication channel are given, the tap coefficient of the decision feedback equalizer is determined. One method of setting is known to be such that the mean square error between the input of the judge and the output of the judge is minimized (for example, Broakis, Digital Communications, published by McGraw-Hill, 1983). In general, the minimum value of the mean square error between the human power of the judge and the output of the judge, which is the evaluation standard of the method, is determined by: 1) Impulse response of the communication channel 2) Signal-to-noise ratio of the received signal 3) Center tap of the decision feedback equalizer It is a function whose variable is the position relative to the impulse response of the communication channel. Of these variables, 3) can be changed in a decision feedback equalizer, and 1) and 2) cannot be changed in a decision feedback equalizer. . Therefore, in order to find the tap coefficient of the decision feedback equalizer that minimizes the mean square error between the decider input and the decider output, it is necessary to change the center tap position of the decision feedback equalizer with respect to the channel impulse response. Then, the minimum mean square error at each position must be determined, and the tap coefficients obtained when the center tap of the decision feedback equalizer is set at the position that gives the minimum value of the obtained minimum mean square error must be determined. Here, the minimum mean square error obtained when the center tap position of the decision feedback equalizer for the channel impulse response is fixed is the local minimum mean square error, and the minimum value of the local minimum mean square error is the true value. is defined as the minimum mean squared error of

In this way, in general, the decision feedback equalizer uses the optimal channel impulse that can obtain the true minimum mean square error, as described in the section (Problems to be Solved by the Present Invention). Good reception characteristics cannot be obtained unless the center tap of the decision feedback equalizer is set at the response position.

By the way, in the method of setting the tap coefficients so as to minimize the mean square error between the judger input and the judger output, the tap coefficients of the post-transversal filter are determined based on the impulse response of the communication path and the pre-transversal filter. is given by the convolution integral with the impulse response of (
For example, Proakis, Digital Communications, published by McGraw-Hill, 1983). In other words, once the impulse response of the communication path and the impulse response of the pre-transversal filter are determined, the post-transversal filter can be uniquely determined. Therefore, a decision feedback equalizer that provides a true minimum mean square error constitutes a pretransversal filter that provides a true minimum mean square error for a given channel impulse response and S/N ratio. If possible, it can be configured uniquely. Considering this, the above-mentioned discussion regarding the position of the center tap of the decision feedback equalizer with respect to the channel impulse response is based on the condition that the center tap of the decision feedback equalizer is set at a certain position of the channel impulse response. It can be seen below that this is an equivalent argument to whether or not it is possible to construct a transversal pre-filter that gives the true minimum mean squared error.

Now, in conventional decision feedback equalizers, at the time when the decider judges the th transmitted symbol, the pre-transversal filter does not include the received signal before the th transmission symbol. The center tap of the filter must be the last tap of the pretransversal filter. Therefore, the degree of freedom of the pre-transversal filter is limited to a very narrow range, and under the condition that the center tap of the decision feedback shaper is set at a specific position of the channel impulse response, it is difficult to determine which There are characteristics that cannot be realized even if the tap coefficients of the pretransversal filter are set. Therefore, depending on the position of the center tap of the decision feedback equalizer with respect to the channel impulse response, it may not be possible to construct a pretransversal filter that provides a true minimum mean square error. Therefore, in order to obtain good reception characteristics using such a decision feedback shaper, it is necessary to use a pre-transversal filter with a small degree of freedom to achieve characteristics that provide the true minimum mean square error. In this case, the position of the center tap of the decision feedback shaper relative to the channel impulse response must be changed.

However, in the decision feedback shaper of the present invention, at the time when the determiner determines the th transmission symbol, the pre-transversal filter contains the received signal before the th. Therefore, there is no need to set the center tap of the decision feedback shaper to the last tap of the pretransversal filter, and the degree of freedom of the pretransversal filter increases. Furthermore, at the time when the decision unit judges the th transmission symbol, by including a sufficient number of received signals before the th transmission symbol in the pre-transversal filter, the center tap of the decision feedback shaper is set to the channel impulse. It becomes possible to configure a pretransversal type filter that can obtain a true minimum mean square error even if it is set at any position in the response.

(Embodiment) FIG. 1 is a block diagram showing an embodiment of the decision feedback shaper of the present invention. In Figure 1, 1o is an input terminal, 1
1 is an output terminal, 100 is a determiner, 101 to 107 are delay elements for one symbol time, 110 to 117 are memories for storing tap coefficients, 120 and 121 are adders, 130 to
137 is a multiplier, and 140 is a differential circuit. The pre-transversal filter includes delay elements 101-104, memories 110-114, multipliers 130-134, and adder 120C'. Further, the post-transversal type filter includes delay elements 106 and 107, memory 115
117, multipliers 135 to 137, and an adder 121.

Next, the basic operation of the decision feedback shaper of the present invention shown in FIG. 1 will be described. A received signal sampled at one symbol intervals is input from an input terminal 1o.

The signal input from the input terminal 10 is filtered by a pre-transversal type filter and input to the difference circuit 140. The difference circuit 140 calculates the difference between the output of the front transversal filter and the output of the rear transversal filter, and outputs the obtained signal to the determiner 100. The determiner 100 makes a determination based on whether the output signal of the differential circuit 140 is positive or negative, and outputs +1 or -1 to the output terminal 11 as the determination result. In addition, the determiner 10
0 also outputs the determination result to the delay element 105. Delay element 105 delays the input signal by one symbol time,
Output to post-transversal filter.

For example, at the time when the determiner 100 outputs the determination value a(k) regarding the k-th transmitted symbol a(k), the received signal r(k
) is output from the delay element 102. Therefore, determiner 1
At the time of determining the symbol a(k+1) transmitted 00+1st, r(k) will be included in the prefilter.

The tap coefficients of the pre- and post-transversal filters and the constants stored in the memories 110 to 116 are the constants stored in the memories 110 to 114, respectively, by F(-2).
~F(+2), and the constants stored in memories 115 to 117 are respectively B(+1) to B(+3),
For example, the channel characteristics shown in FIG. 2 can be determined by solving the following simultaneous linear equations.

However, Φ is a 5×5-order Hermitian matrix shown below,
The matrix is as follows.

Φ21. n is given by Here, σn2 is the average power of the noise,
δ is the delta function of . Further, Δ is a matrix given by the following, and P is a 3×3 unit matrix. Here, t indicates the transposition of the matrix. Further, r, e, and G are column vectors shown below.

rt=(FC2), F(1), F(0)
, F(-1), F(-2))θ'=(B(1)
), B(2), B(3))Gt=(g(-
2), g(-1), g(0), g(
1), g(2)) (Effects of the Invention) According to the present invention, good reception characteristics can be obtained even if the center tap of the decision feedback equalizer is set at any position in the impulse response of the communication channel. It becomes possible.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a system diagram showing an embodiment of a decision feedback equalizer according to the present invention. In the figure, 10... Input terminal 11... Output terminal 100... Determiners 101-107... Delay elements 110-117... Memories 120-121... Adders 130-137. . . . Multiplier 140 . . . is a differential circuit. FIG. 2 is a diagram showing an example of an impulse response of a communication channel. In the figure, g(k) is the channel gain.

Claims (1)

[Claims] A pre-transversal filter consisting of N taps that receives the received signal as input, a delay element that receives the output signal from the determiner as input and delays the input signal by one symbol time T, and receives the output of the delay element as input. A post-transversal filter consisting of M taps, a difference circuit that takes the difference between the output of the pre-transversal filter and the output of the post-transversal filter, and a determiner that receives the output of the difference circuit as an input. A decision feedback type equalizer comprising: a decision feedback type equalizer, characterized in that the pretransversal type filter includes a received signal before the kth at a time when the kth transmission symbol is determined in the decision unit; Equalizer.