JPH04344731A - Method for extracting correlation value of spread signal - Google Patents

Abstract

PURPOSE:To obtain an original data accurately even when an output signal of a correlation device is largely distorted by correcting distortion in an output signal of each correlation device due to a transmission line when a spread spectrum signal is demodulated. CONSTITUTION:Cells starting from a cell number 1 are inputted to one memory register 17, while cells starting back from a cell number 93 are inputted to the other memory register 18 and a correlation device output is sequentially written to the 93 cells based on a clock signal fed from a clock oscillator. As a result, a correlation device output waveform R (tau) is stored in the memory register 17 and a correlation device output waveform R (-tau) resulting from inverting a time axis of the waveform R (tau) is stored in the other memory register 18 respectively. The contents of both the memories are added at an adder 19 for synthesis based on a timing pulse from a timing extract device 42, the result is transferred to a 3rd memory register 20, and then given to a square circuit 21, in which the amplitude is emphasized, a negative value is inverted into a positive value to obtain a prescribed waveform.

Description

[Detailed description of the invention]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for demodulating spread spectrum modulated signals and a demodulating device for implementing the method, and more particularly to a method and device suitable for demodulating direct spread modulated signals.

[Prior Art] Conventionally, in local area network systems, home control systems, etc., in which transmission lines are laid inside buildings to transmit information signals or control signals, costs associated with construction of the transmission lines have been reduced. In order to do this, it is being considered to use electric light lines and the like that have already been installed inside buildings as transmission lines. On the other hand, although a number of methods have been proposed for spread spectrum modulation, the direct spread method is known as the basic modulation method. FIG. 6 is a block configuration diagram for explaining the direct sequence modulation method,
On the transmitting side M, a multiplier 62 multiplies the transmitted data and the PN sequence 61.
The signal is multiplied by product modulation, and after being amplified to a required level by an amplifier 63, it is sent to a transmission line 65 such as a power line via a transformer 64. On the other hand, on the receiving side D, the spread signal is taken in by a coupling transformer 66, passed through an amplifier 67, and then inputted in parallel to a correlator 68 and a synchronous PN sequence generator 69, and is correlated with the received PN sequence while maintaining synchronization with the received PN sequence. A comparator 70 detects whether the output level of the correlator despread by the comparator 68 is larger or smaller than the reference voltage REF, and the data is demodulated based on the result.

However, if the transmission characteristic of the communication line is ideal, the correlator output value should normally exhibit a maximum value or minimum value for both positive and negative values at the synchronization point, as shown in FIG. In the case of power distribution lines such as electric light lines that do not take impedance characteristics into account and have a lot of noise, the transmission characteristics deteriorate due to the occurrence of dip points within the signal frequency transmission band. ,
As shown in (b) of the same figure, various distortions occur, such as the correlation value at the synchronization point having a negative value, or the positive peak point deviating greatly from the synchronization position, and furthermore, the peak value level becomes small. As a result, normal demodulation was impossible. Conventionally, S. G. Glisic
proposed it as the MDS/SSKSS method in 1988 (
American academic journal IEEE, Transactions o
n Communications, vol. 36,
No. 4, March 1988 issue, pages 519 to 527)
A spread spectrum signal modulation/demodulation method called the CSK method, which is an improved version of this method (NEC Technical Report vol. 1.42 NO.
9/1989: "High Performance SS Power Line Modem") is proposed.

[0003] This method uses "1" and "0" of the data to be transmitted.
”, one of the two PN sequences with the same period length and low correlation is sent out, and the receiving side uses the level of the correlation value determined by the same PN sequence as the sending side to transmit the received data. The principle of this demodulation is explained with reference to FIG. When the state value of the information data to be transmitted is "0", the PN sequence PN0 is set to a timing generator, and when the state value of the information data to be transmitted is "1", the PN sequence PN1 is set to the timing generator. 83, under the control of a synchronization circuit 84, it is sent to a power line 89 via switches 85, 86, an amplifier 87, and a transformer 88. On the other hand, on the demodulation side, a transformer 90
, an amplifier 91, and a bandpass filter 92, a correlator 93 and a PN sequence generator 94 take in the correlation value of the transmitting side PN1 sequence, and a correlator 95 and a P
The correlation values of the PN0 sequence are determined by the N0 sequence generator 96 and each is input to the peak value comparator 97. This peak value comparator 97 includes, for example, an RS flip-flop circuit, and compares each of the correlation values of the PN1 and PN0 series with a predetermined threshold level, and generates a pulse when the threshold level is exceeded. The set terminal S and reset terminal R of the flip-flop circuit
, and depending on which of the correlation values the peak value occurs, the output of the flip-flop circuit will be “1” or “1”.
0" is output.

[0004] In this system, if a matched filter is used in the correlator on the reception side, synchronization control of the PN sequences between the transmitter and the receiver becomes unnecessary, and the synchronization acquisition time can be shortened. In addition to being extremely advantageous for communications, etc., even if the signal is slightly distorted by the transmission path, data demodulation is possible as long as a peak can be obtained in the correlation signal. However, when the transmission characteristics of the electric line are poor, a clear peak cannot often be obtained in the output waveform of the receiving side correlator, and the demodulation error rate is quite high, making it unsuitable for high-speed data communication. As a countermeasure to this, the NEC technical report vol. 1.42
No. 9/1989: The "high-performance SS power line modem" has the absolute value of the signal extracted by the time window F centered at the peak position and the time window E provided on both sides of each correlation value, as shown in Figure 9. A method has been proposed in which the E part of one correlation value is multiplied by the absolute value of the F part of the other correlation value, and the data is demodulated from the polarity of the difference between the two. It is necessary that the value has a relatively clear peak, and if the transmission characteristics deteriorate further and the peak of the correlation value becomes unclear, or if the position of the correlation peak deviates from the synchronization point, As in the past, the error rate increases and the data is no longer available.
It was impossible to demodulate the data. To help understand the present invention, the reason why the correlator output waveform is distorted due to the influence of transmission characteristics will be explained in detail using mathematical expressions.

Now, let the impulse response and frequency characteristics of the transmission path be h(t) and H(f), respectively, the transmit signal output from the modulator be X(t), and the receive signal received by the demodulator side be R( t), and further, when the spectra of the transmitted and received signals are respectively represented by X(f) and R(f), the receiving side correlator output R (τ) is expressed by the following equation (1).

##EQU00001## In this equation, .tau. is the time difference between the received signal in the receiving correlator and the receiving PN sequence. If we transform this equation by substituting each of the above functions, we get equation (2).

[Equation 2] When the time constant τ is zero (τ=0), that is, when the PN sequence code input to the receiving correlator and the received signal are kept synchronized, the correlator output R(0) is expressed by the formula ( 3).

[Formula 3] As is clear from this equation, the correlation value R(0) is affected by the amplitude and phase characteristics of the transmission characteristic H(f) of the transmission line, and can take both positive and negative values. become. Therefore,
Depending on the characteristics, the data undergoes various deformations such as a change in the polarity of the correlation peak, a change in the position of the correlation peak, or a change in the peak value, so it is difficult to perform accurate demodulation using the conventional data demodulation method described above. .

[0006]

OBJECTS OF THE INVENTION The present invention has been made to solve the above-mentioned problem of the spread spectrum signal modulation/demodulation method, and is aimed at solving the problem of the above-mentioned spread spectrum signal modulation/demodulation method. A spread spectrum signal demodulation method that allows the demodulator to obtain the original information data even if the existence of a correlation peak is not clear, and a spread signal correlation value that can be used as a synchronization means in various spread spectrum communications. The purpose is to provide an extraction method.

SUMMARY OF THE INVENTION In order to achieve the above object, in the present invention, when extracting a correlation value from a spread spectrum modulated signal, a despread signal waveform of one cycle or more and a signal waveform of the despread signal waveform are used. The present invention is characterized in that a symmetrical waveform is created with respect to a predetermined position on the time axis, and by combining these two signals, a significant correlation peak is formed at the point defined as the axis of symmetry.

Specific means for this purpose include, for example, a first memory register of a predetermined length that stores the despread signal in the forward direction, and a second memory register of a predetermined length that similarly stores the despread signal in the reverse direction. a memory register, and adds the contents of the two memory registers at a predetermined timing,
It is sufficient to obtain a correlation value from the added signal, and this means is particularly effective for the CSK method described above, and can also be used as one of means for demodulating and establishing synchronization of various spread spectrum signals.

DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be explained in detail below based on embodiments shown in the drawings.Prior to that, the principle of the present invention will be briefly explained. FIG. 1 is a waveform diagram for explaining the principle of the present invention, in which the output waveform R(τ) of the receiving side correlator
is when the characteristics of the transmission path are ideal, that is, |X(f) |
When is 1, it peaks at the synchronization point 0, as shown in the simulation in Figure (a), but as mentioned above, it is affected by the amplitude and phase characteristics of the frequency characteristic H(f) of the transmission path. As a result, the waveform is often distorted, for example, as shown in FIG. In this case, the correlation peak position deviates from the synchronization point, and accurate data demodulation is impossible, as described above.However, in the present invention, as shown in FIG.
generates a symmetrical waveform with respect to synchronization point 0, and adds these two to create the signal shown in FIG. This signal is symmetrical about the axis set at the synchronization point, so that it can be understood visually.

Moreover, since the peak value is exhibited at that point, even if there is waveform distortion due to the transmission path, it is possible to create a pseudo correlation peak that coincides with the synchronization point. Therefore, by using this signal, it is easy to detect the synchronization point from the received signal, and furthermore, it is also possible to demodulate data from the signal. In addition,
In the case of the direct spread modulation explained with reference to FIGS. 6 and 7, the peak value becomes negative depending on the data bit value. If the present invention is used for demodulating CSK modulation, after determining the waveform of (d) above for each correlator output, the square or absolute value is determined, and all are positive values as shown in (e) of the same figure. After that,
Compare the level and determine if the received data is “1” or “0”.
Physically explaining the above operation, the above diagram (c) is the correlator output waveform R(τ) with the time axis reversed, that is, R(-τ). , also figure (
d) becomes φ( τ) expressed by the following equation (4),
The waveform in (e) is φ2 (τ).

[Equation 4] Note that each of the above operations is performed using a memory register that stores at least one cycle of the correlator output in the forward direction and the backward direction with respect to the time axis, as will be explained in detail in the examples below. This can be easily achieved by preparing the following. That is, FIG. 2 is a block configuration diagram showing an example of a spread spectrum communication device for implementing the present invention, and here, in order to facilitate understanding of the operation, the modulation side will be included in the explanation.

In FIG. 2, T is a modulation device, which generates two PN sequences PN0 and PN1 with mutually small correlation values and a period length of 31 bits, respectively, and each of the PN sequence generators 1 and 2, and one of the PN sequences. Switch 3 to switch and output
, 4, an amplifier 5 for amplifying the spread signal to a predetermined level, a circuit coupling transformer 6, and the supplied transmission data S.
In this example, a synchronization circuit 7 which controls the switch 3 in accordance with the bit value of D and a transmission request signal RS are input.
A 300Hz signal is also applied to the two Ps to the synchronization circuit.
The N sequence generator is equipped with a timing generator 8 that supplies a 9.3 kHz signal, and sends out either the PN sequence PN1 or PN0 to the transmission line 9 in accordance with the bit value of the data to be transmitted. It is configured to do so. On the other hand, if the receiving device R is first roughly divided into blocks, there is an input section IN that takes in the spread signal from the electric path, a PN1 demodulation section and a PN0 demodulation section that generate the correlation signals of the PN1 sequence and PN0, and the correlation between the demodulation section. Timing extraction unit TD that extracts the data period signal from the signal obtained by adding the device outputs.
The configuration includes a data demodulation section DD that reproduces data from the output signals of the two demodulation sections, and a control section CONT that controls the entire receiving apparatus.

[0010] Further explaining the configuration of each of the above blocks,
The input section IN includes a transformer 11 that takes in the spread signal from the transmission line 9, an amplifier 12 that amplifies the output thereof, and a band pass filter or cut whose passband is from 1 kHz to 10 kHz to remove noise outside the band. It includes a low-pass filter 13 with an off frequency of about 10 kHz, and a digital-to-analog converter A/D 14 that converts the filter output into a digital signal. Since each of the PN demodulators has the same configuration, the PN1 demodulator will be explained as follows.
A correlator 15 that despreads the output of the D converter 14, a PN1 generator 16 that generates or stores a reference signal PN sequence for the despreading, and a PN1 generator 16 that stores the output of the correlator 15 from opposite directions. Two memory registers 17 and 18 connected for input, an adder 19 for combining the outputs of the two memory registers, a third memory register 20 for storing the output of the adder, and the memory register output. The PN0 demodulator is also comprised of blocks numbered 21 to 32.

The timing extractor TD also includes an adder 41 that adds the second square circuit outputs of the respective PN demodulators and extracts a periodic signal of the transmitted data from the added signal; container 42. Furthermore, the data demodulation section DD includes an adder 43 that adds the outputs of the PN demodulation section, that is, the outputs of the first squaring circuits 21 and 31, and extracts the timing of the pseudo correlation peak point from the output of the adder. a synchronization point extractor 44, a comparator 46 inputting the outputs of the two PN demodulators, the synchronization point extractor 44 and the comparator 4;
5 and a processing circuit 46 for generating demodulated data and a carrier detection signal from the signal from the control unit CONT. In addition, the control unit CONT has 27.9kHz and 300H.
This configuration includes a clock generator 47 that generates a clock signal of z, and a control circuit 48 that controls the entire receiving device.

The operation and data demodulation procedure of the above configuration will be explained in detail. First, on the modulation side T, before supplying the information data SD to be transmitted to the synchronization circuit 7, the transmission request signal RS is supplied to the timing generator 8, and the signal output from the timing generator 8 causes the second The switch 4 is closed and the first switch 3 is controlled to transmit the PN1 and PN0 codes at predetermined cycles, for example, every 30 cycles. If the PN sequence code is 31 bits and the data transmission rate is 300 b/s, 100 ms is required to transmit 30 cycles of the PN sequence using a clock with a frequency of 9.3 kHz. The reason for transmitting a PN sequence of several tens of cycles before transmitting data in this way is to determine the arrival of a received signal, that is, carrier detection, on the receiving side, and to determine the data transmission timing signal component, in this example. In the case of 300H
This is to detect z and prepare for data demodulation. After completing this operation, the data SD is supplied to the synchronization circuit 7, and the synchronization circuit synchronizes with the clock signal of the timing generator and outputs 1 bit of information data every 1/300 (sec). Output. That is, the switch 3 is switched according to the bit value of the data, and when the data is "1", the PN1 is on the generator side, and when the data is "0", the P is on the generator side.
N0 Outputs the PN sequence on the generator side one period at a time.

Next, the operation of the demodulator on the receiving side will be explained. At its input section IN, the spread signal is taken in through the transformer 11, and after being amplified as required by the amplifier 12, the signal is input to the filter 13. removes various noises outside the band, passes only the main lobe of the spread signal, and then
14 into a digital signal. The sampling frequency at this time is an integral multiple of 9.3 kHz, and in this example, it is tripled to 27.9 kHz, and the level (amplitude value) values of three samples per 1 bit of the PN sequence signal are converted into, for example, a 16-bit digital Generate as a signal. The output of the analog/digital converter becomes a digital signal, and all subsequent processing is originally a digital signal.However, in order to make it easier to understand the processing contents, the configuration of the device is divided by function as shown in Figure 2. However, in reality, it can be replaced with a digital processing circuit such as a DSP. Further, for the same reason, the following explanation and waveform diagrams will be explained based on analog signal waveforms. The operation of each PN demodulator is
To explain the demodulation section, first, the correlator 15 converts the digitally converted spread signal and the PN sequence PN stored in the memory.
A so-called matched filter type correlator is common, which calculates the correlation value between the two by multiplying by 1.
This correlator output is input to two 93-bit (93 stages) memory registers (shift registers), but as described above, the input directions are opposite to each other.

That is, as shown in FIG. 3, one memory register 17 receives input from cell number 1, while the other memory register 18 receives input from cell number 93.
27.9 cells supplied from clock oscillator 47
Correlator outputs are written sequentially based on a kHz clock signal. As a result, the memory register 17 has the R(
τ), and the other memory register 18 has R(-
τ) are respectively stored in memory, and the contents of both memories are combined and added by an adder based on a timing pulse from a timing extractor 42, which will be described later, and then transferred to the third memory register 20. The content of this third memory register corresponds to φ(τ) shown in equation (4) above, and its waveform is as shown in FIG. 1(d) above, and the square circuit 2
1 to emphasize the amplitude and invert the negative value to a positive value to obtain the waveform φ2(τ) shown in FIG. 1(e). In addition, the above 2
The multiplication circuits 21 and 31 may be absolute value circuits, and in this case, the waveform in (e) becomes |φ(τ)|. Further, regarding the PN0 demodulator, the reference signal is simply replaced with PN0, and the other operations and configurations are exactly the same as described above. The above processing steps will be explained in more detail.

FIG. 4 shows the values stored in the memory register shown in FIG. 3. The horizontal axis indicates the memory cell number of the memory register, and the vertical axis indicates the correlation output level at each sampling point. This is what is shown. In the same figure (
a) is stored in the memory register 17 in the positive direction of the time axis, and the correlation value output x1 sampled sequentially from the left;
x2 , x3 . . . as the first memory register 17
The information is sequentially input to the cell, the second cell, the third cell, and so on. In addition, (b) is what is stored in the memory register 18 apparently in the negative direction with respect to the time axis, and the first cell of one memory and the 93rd cell of the other cell, the second cell and the 92nd cell, and When the third cell and the 91st cell are added together and stored in the third memory register 20, the result is shown in FIG.
A signal corresponding to the waveform is obtained. This is shown in Figure 1(e) above.
), that is, φ2(τ), and as mentioned above, a waveform that is symmetrical about the axis set at the peak position is obtained. The timing pulse shown in FIG. 4(c) is adjusted by a phase shift adjustment circuit (not shown) so that it is located at the center cell of each memory register, the 47th cell in this example.

Next, the timing pulse extraction will be explained with reference to FIG. That is, FIG. 5(a),(b)
are the correlator outputs of the PN1 demodulator and PN0 demodulator, respectively, and the square circuit outputs thereof are (c) and (d). This example shows the case where "1", "1", and "0" are transmitted as data, and the sum of the above two squared outputs is (e), and from this waveform signal Via a bandpass filter with a passband center frequency of 300Hz (
f) A signal close to a substantially sinusoidal wave having a data transmission period is extracted. Therefore, based on this signal, for example, the timing pulse shown in (g) is generated at the zero cross point that changes from negative to positive, and after adjusting it to near the center cell of the memory register as described above, the timing pulse as described above is generated at this timing. The memory cells of the two memory registers 17 and 18 and 27 and 28 are added together. The manner of addition between the two memory registers at this time is as explained in FIGS. 1, 3, and 4, and the squaring circuit 21 of each PN demodulator,
A symmetrical pseudo-correlation waveform is obtained from the output of 31, and the data
The comparator 45 of the data demodulator DD compares the levels of both, and when the level of PN1 is higher, a positive signal waveform is output, and when the level of PN0 is higher, a negative level signal waveform is output. In addition, the processing circuit

At step 46, the data bit value is determined based on the timing pulse supplied from the timing extractor 44, which is the same as shown in FIG.
As shown in i), the data is demodulated into "1", "1", and "0". In this example, the timing extractor is 42,4
4. First, the timing extractor TD obtains a rough synchronization point, calculates φ( τ) from this, and then calculates the signal φ( τ) with the distortion of the transmission path corrected.
The purpose is to extract a more accurate synchronization point and compare the two correlation values, but the output of the first timing extractor 42 may also be used as the timing signal for the data demodulator. . In this case, the adder 43 of the data demodulation section
, the timing extractor 44 becomes unnecessary, but the processing circuit 46
Some phase shift correction may be necessary so that the timing pulses applied to the If configured and operated as described above, as explained in the summary of the invention, R
(τ) and R(-τ) are found, they are added, and a signal with a pseudo correlation peak at the synchronization point can be created from the signal, which can be used to demodulate data or adjust the synchronization timing. Extraction, etc. becomes possible. The above-mentioned specific example is only one embodiment of the present invention, and it is natural that various modifications can be made without being limited to this example. For example, although synchronization is not particularly taken into consideration between the output pulses of the clock oscillator 47 and the timing extractor 42, the case is not limited to this, and the clock signal of the clock oscillator may be synchronized with the 300 Hz pulse signal output from the timing extractor. By doing so, operations such as writing to a memory register and addition become easier.

[0018]

Effects of the Invention As explained above, the present invention corrects distortion caused by the output signal transmission path of each correlator when demodulating a spread spectrum signal, and generates a correlation peak at a predetermined position. Even if a large delay deviation and amplitude deviation occur in the transmission characteristics of the channel and the output signal of the correlator is greatly distorted, the original information data can be accurately obtained.

[Brief explanation of the drawing]

FIGS. 1A to 1E are signal waveform diagrams for explaining the principle of the present invention.

FIG. 2 is a block diagram showing an embodiment of a spread spectrum signal modulation/demodulation device for implementing the present invention.

FIG. 3 is a diagram for explaining a partial operation of the same embodiment.

FIGS. 4(a) to 4(d) are signal waveform diagrams for explaining the operation of the above embodiment.

5(a) to 5(i) are signal waveform diagrams for explaining the operation of the above embodiment, similar to FIG. 4; FIG.

FIG. 6 is a block diagram of a device configuration for explaining a conventional direct spread communication method.

FIGS. 7(a) and 7(b) are signal waveform diagrams for explaining a conventional direct spread communication method.

FIG. 8 is a block diagram of a conventional improved spreading communication device.

FIG. 9 is a signal waveform diagram of a conventional improved spreading communication device.

[Explanation of symbols]

1, 2...PN sequence generator 3, 4...Switch 7...Synchronization circuit 8...Timing generator 9...Transmission line 15, 25...Correlator 16, 26...PN sequence generator 17, 18, 20, 27, 28 , 30...Memory registers 21, 22, 31, 32...Squaring circuits 19, 29, 41, 43...Adder 45...Comparators 42, 44...Timing extractor 46...Processing circuit 47...Clock oscillator 48...Control circuit T ...Modulation device R...Demodulation device

Claims (5)

[Claims] Claim 1: When extracting a correlation value from a spread spectrum modulated signal, a signal waveform of one period or more despread using the same PN sequence as the modulating side PN sequence, and the time of the signal waveform. 1. A method for extracting a correlation value of a spread signal, characterized in that a pseudo correlation peak is formed at a point defined as the axis of symmetry by combining waveforms that are symmetrical with respect to a predetermined position on the axis. 2. A first memory of a predetermined length for storing in the forward direction a despread signal using the same PN sequence as the modulating side PN sequence when extracting a correlation value from a spread spectrum modulated signal; A register, and a second memory register of a predetermined length for storing the despread signal in the opposite direction, the contents of the two memory registers are added at a predetermined timing, and a correlation value is obtained from the added signal. A method for extracting a correlation value of a spread signal, characterized in that the following is obtained. 3. A first memory of a predetermined length for storing in the forward direction a despread signal using the same PN sequence as the modulating side PN sequence when extracting a correlation value from a spread spectrum modulated signal; A register and a second memory register of a predetermined length for storing the despread signal in the reverse direction, and the contents of the two memory registers are added at a predetermined timing, and the transmitted/received PN is determined from the added signal. A method for extracting correlation values of spread signals characterized by determining synchronization timing of sequences. 4. When the spread spectrum signal is modulated to output one of a plurality of PN sequences having equal period lengths and a small correlation value in accordance with the state value of information data to be transmitted; In this process, the signals despread using the same PN sequence as the modulating side PN sequence are squared or their absolute values are calculated, and then the data transmission period is extracted from the summed signals, and the respective inverse A forward direction memory register of a predetermined length for storing the spread signal in the forward direction, and a backward direction memory register of a predetermined length for similarly storing the despread signal in the reverse direction, respectively at the extracted predetermined timing. 2. A method for extracting a correlation value of a spread signal according to claim 1, wherein the contents of two memory registers, the forward and reverse sides of the despread signal, are added, and the correlation value is obtained from the added signal. 5. When the spread spectrum signal is modulated to output one of a plurality of PN sequences having equal period lengths and a small correlation value in accordance with the state value of information data to be transmitted; In this process, the signals despread using the same PN sequence as the modulating side PN sequence are squared or their absolute values are calculated, and then the data transmission period is extracted from the summed signals, and the respective inverse A forward direction memory register of a predetermined length for storing the spread signal in the forward direction, and a backward direction memory register of a predetermined length for similarly storing the despread signal in the reverse direction, respectively at the extracted predetermined timing. 2. The method for extracting a correlation value of a spread signal according to claim 1, wherein the contents of two memory registers, the forward and reverse sides of the despread signal, are added, and the comparison timing of both correlation value levels is determined from the added signal. 0001