JP3179554B2 - Spread spectrum communication system - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

The present invention relates to relates to a spread spectrum communication <br/> system.

[0002]

2. Description of the Related Art Spread spectrum communication is expected to be applied to wireless LANs in that it is resistant to CDMA and interference waves. There are various types of spread spectrum communication. Among them, the direct spread type is conventionally best known and most often used.

FIG. 12 is a configuration diagram of a spread spectrum communication system based on the direct spread system. Referring to FIG. 12, the spread spectrum communication system includes a transmitter 1
00 and a receiver 101. Spreading coder 102 that outputs spreading code PN (t) to transmitter 100
A multiplication unit 103 for multiplying a binary (± 1) data symbol d (t) for each period of the spreading code;
And a modulator 104 that modulates the carrier with the output from the controller and outputs the modulated signal as a transmission signal. The receiver 101 receives a transmission signal from the transmitter 100 and demodulates the signal, and demodulates a matched filtering process (passive correlation) corresponding to the spreading code PN (t) of the transmitter 100. PN applied to the demodulated signal from section 105
Matched filter 106 and PN matched filter 10
And a sampling circuit 107 that samples the output from the sampling circuit 6 at predetermined time intervals.

In the spread spectrum communication system having such a configuration, first, the transmitter 100 generates a binary (±
After embedding the data symbol d (t) of 1), that is, the information to be transmitted, in the spreading code PN (t) from the spreading encoder 102, it modulates the carrier and outputs this as a transmission signal. The receiver 101 receives and demodulates the transmission signal from the transmitter 100 and inputs the demodulated signal to the PN matched filter 106, where the spread code PN (t) of the transmitter 100 is received.
, And by sampling the output y (t) from the PN matched filter 106 at predetermined time intervals, the data symbol d (t) can be reproduced.

More specifically, the number of chips (code length) of the spread encoder 102 of the transmitter 100 is L, and the spread encoder 102 outputs a spread code PN (t) having the number of chips L. Supposing that one chip time interval (period between codes) is T _c (seconds), the spreading code P
One cycle T of N (t) is LT _c (seconds). Transmitter 10
In the multiplier 103 of 0, ± 1 for each cycle T (= LT _c )
The binary data symbol d (t) of 1 is converted to a spreading code PN (t).
, And outputs the result, for example, as shown in FIG.

On the other hand, on the receiver 101 side, the transmitter 100
Is demodulated into a signal having a form as shown in FIG. 13A, which corresponds to the spreading code PN (t) (one cycle T (= LT _c )) in the transmitter 100. PN matched filter 106 that performs the matched filtering process. Here, a solitary wave signal for one cycle T of the spreading code PN (t) is defined as p (t) as in the following equation.

[0007]

(Equation 1)

When the transmission path noise is n (t),
The transfer function H (ω) of the PN matched filter 106 for p (t) is calculated using the Fourier transform P (ω) of p (t) and the power spectrum density N (ω) of noise n (t).
It is expressed by the following equation. Here, * represents a complex conjugate.

[0009]

H (ω) = P ^* (ω) exp (−jωT) / N (ω)

If the noise n (t) is white noise, N
Since (ω) is a constant, Equation 2 is as follows.

[0011]

H (ω) = P ^* (ω) exp (−jωT)

Since the unit impulse response h (t) at this time is p (Tt), the PN matched filter 106
Output y (t) is as follows.

[0013]

(Equation 4)

For the sake of simplicity, n (t) = 0, d
If (t) = 1, the output y (t) is expressed by the following equation.

[0015]

(Equation 5)

As can be seen from Equation 5, in the matched filter, the pulse waveform p for one cycle T of the spreading code PN (t)
The autocorrelation function of (t) will be output. By the way, many spread signals are pseudo-noise-like, and their autocorrelation functions are as shown in FIG. If the waveform of the spread signal is a binary waveform of ± A, the output y (t) from the matched filter 106 is as shown in FIG. As can be seen from FIG. 14B, since the output y (t) from the matched filter 106 has a peak value when t is T, the sampling circuit 107 outputs the output y (t) from the matched filter 106. What is necessary is just to sample at the time interval T. FIG. 13B shows a waveform of an actual output y (t) when a signal as shown in FIG. 13A is input to the matched filter 106. By sampling at the interval T, the data symbol d (t) from the transmitter 100 can be accurately reproduced based on the above-described principle.

[0017]

As described above, in the conventional spread spectrum communication system based on the direct spreading method, when one cycle of the spreading code PN (t) is T, PN
By sampling the output y (t) from the matched filter 106 at time intervals of T, as can be seen from FIGS. 14 (b) and 13 (b), the output y (t) is generated without causing intersymbol interference. ) Can be taken out only, thereby improving the error rate characteristic and the interference wave removing ability.

However, in the above-described conventional system, the transmission rate is limited by the period T (= LT _c ) corresponding to the number L of chips of the spreading code PN (t). However, there has been a problem that the transmission speed cannot be further increased.

An object of the present invention is to provide a spread spectrum communication system, a transmitter, and a receiver capable of further increasing the transmission rate without deteriorating the error rate characteristics, the interference wave removing ability, and the like. .

[0020]

The inventor of the present application has found that the output y (t) of the PN matched filter 106 is shown in FIG.
4 (a) reflecting the autocorrelation function of FIG. 4 (a), a spreading code (such that the autocorrelation function becomes “0” at regular intervals except for τ = 0) In other words, the side lobe of the autocorrelation function has a constant period.
If there is a spreading code that becomes "0" only at the sampling point ), use this spreading code and shorten the sampling time interval in accordance with the cycle in which the side lobe of the autocorrelation function becomes "0". I found that it can be. That is, since the matched filter is a linear time-invariant system, the period T (= LT) determined by the number of chips (code length) L of the spread coder is sufficient if the timing is accurately obtained.
_It has been found that not only _c ) but also the cycle T _c between codes (that is, the cycle between chips) can be used for information transmission. Therefore, the present invention is characterized by using a spreading code in which the side lobe of the autocorrelation function becomes “0” at regular intervals, and uses this spreading code to spread the data symbols based on the above principle. It is intended to transmit and receive one after another at shorter intervals without leaving T intervals.

FIGS. 1A and 1B are diagrams for explaining the basic concept of a spread spectrum communication system according to the present invention. FIG. 1A shows a processing form in a transmitter. 1 (b) shows a processing mode in the receiver. The transmitter of the present invention basically converts a serial data symbol as information to be transmitted into parallel data shifted by a predetermined period (for example, 2T _c ) as shown in FIG. A spread signal is applied to the parallel data, these are added by an adding means ADD, and a carrier is modulated by the added output and transmitted. Further, the receiver of the present invention demodulates the transmission signal from the transmitter and inputs the demodulated signal to a matched filter MF corresponding to the spreading code used in the transmitter as shown in FIG. Then, the output of the matched filter is, for example, 2
Sampling is performed at time intervals of _Tc to reproduce data symbols. As can be seen by comparing FIG. 1B with FIG. 13B, in the related art, the time interval T (= LT
_c ) Transmission and reception were performed at the transmission rate determined by
In the present invention transmits at the transmission rate determined by the time interval of, for example, 2T _c, can perform reception, it is possible to increase the transmission rate.

When transmission is performed in the manner shown in FIGS. 1A and 1B, the spread signals of two adjacent information symbols mostly overlap on the time axis. That is, they share most of the time with each other. In the ordinary direct spreading method, one data symbol, that is, one information symbol is transmitted at each time interval T. Therefore, when only one-channel communication is performed in baseband, it is transmitted as a binary signal. In the invention, the parallel signal is generally transmitted as a multi-level signal by being added by the adding means ADD. Receivers are required to have a system linearity that can withstand this multi-level signal, but this can be easily achieved with recent device technology by using digital signal processing type matched filters. it can. As long as this linearity is ensured, the error rate characteristics and the interference wave removal capability can be maintained at good levels without any difference from the conventional direct spreading system. In addition, the complexity of the system hardly changes as described later.

[0023]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 2 is a configuration diagram of one embodiment of the spread spectrum communication system according to the present invention. Referring to FIG. 2, the spread spectrum communication system according to the present embodiment includes a transmitter 1,
And a receiver 2. The transmitter 1 applies a predetermined spreading code to an information data symbol d (t) to be transmitted, modulates a carrier with this output, and outputs the modulated signal as a transmission signal.

The receiver 2 receives the transmission signal from the transmitter 1 and demodulates it, and performs a matched filtering process (passive correlation) corresponding to the spreading code in the transmitter 1 on the demodulated signal. The result of the matched filtering process is sampled at predetermined time intervals to reproduce data symbols. Here, the spreading code is assumed to have a chip number (code length) of L (chip) and an inter-chip cycle of _Tc , and assume that one cycle of the spreading code is T (= L
T _c ).

Here, in order to further improve the transmission rate as compared with the prior art, information symbols, that is, data symbols, are spaced at intervals shorter than the period T (= LT _c ) determined by the number L of spreading code chips. Must be sent,
In that case, in order to reproduce the data symbol accurately, it is important to prevent the occurrence of intersymbol interference in the matched filtering processing. Therefore, in the case where data symbols are transmitted at intervals shorter than the period T, a spreading code suitable for the present invention that does not cause intersymbol interference in the matched filtering process will be examined first.

As mentioned above, the output of the matched filter is the autocorrelation function of the signal of interest. In this case, the target signal is the signal p (t) of Expression 1. When the signal p (t) is subjected to the matched filtering processing, its output y (t) is given by Expression 5, and for example, when a normal M sequence is used as the spreading code, the side lobe of the autocorrelation function is “ It does not become 0 ". That is, when data symbols are transmitted at intervals shorter than T using a normal M sequence as a spreading code, intersymbol interference occurs in the output y (t) from the matched filter, and the performance is reduced. Will deteriorate. Therefore, the interference between adjacent symbols of matched filtering is large in an ordinary M sequence or the like, and these cannot be used in the present invention. On the other hand, FIG. 3A shows a pulse satisfying the well-known first Nyquist criterion. This pulse is given by τ =
Except for 0, the value is “0” every fixed period T _s (ie, the side lobe is “0” every fixed period T _s ), and the output of the matched filter, that is, the autocorrelation function of Expression 5 if There characterized in FIG. 3 (a), it is not necessary to cause intersymbol interference in the matched filter output y (t) also when transmitting data symbols at a short time interval T _s than T . In order to further increase the transmission speed, in FIG. 3A, it is better that the period T _{s at} which the autocorrelation function is “0” is as short as possible. As can be understood from the above description, even when data symbols are transmitted at a time interval shorter than T, in order to prevent the occurrence of intersymbol interference in the matched filter output and to prevent the performance from deteriorating, as shown in FIG. In this case, a spreading code may be used such that the side lobe of the autocorrelation function of Equation 5 periodically becomes “0”.

The black spot in FIG. 3B is obtained when the autocorrelation function of the spreading code PN (t) is an integral multiple of the chip (τ = n
Discrete autocorrelation function R from which only the value of T _c ) is extracted
_dpp (n) = R _pp (nT _c ). In order to further increase the transmission rate, the discrete autocorrelation function R
_dpp (n) is, like Kronecker's delta function, n =
It is desirable to be 0 at every point other than 0 (by one interval _Tc ), which means that the original spread code is also a Kronecker delta function, and cannot be a spread code. That is, the solid line in FIG. Thus, the minimum spacing established as the spreading code becomes (2T _c at time intervals) 2 interval (see dashed line in Figure 3 (b)). That is, an even chip excluding τ = 0 (τ = 2n × T _c (n
= 1, 2, 3,..., L / 2-1). Using such spreading codes, each two-chip, i.e., put the data every 2T _c, feed admitted overlap of pulses of adjacent spreading codes, if sampling the matched filter output for each 2 chips on the receiving side, Communication can be performed without causing intersymbol interference after matched filtering.

More specifically, in a system in which L is an even number and an information symbol is transmitted every two chips, R _dpp (n) satisfies the following equation in order to perform transmission without causing intersymbol interference. (Fig. 3 (b) shows this R _dpp (n)
Is indicated by a broken line).

[0029]

(Equation 6)

Considering that | n | ≧ L and R _dpp (n) = 0 and symmetry, Expression 6 is expressed in the form of a matrix as follows.

[0031]

(Equation 7)

In the above equation, L is an even number, and a, b,
c is an arbitrary value. Also, p ₁ , p ₂ , p ₃ ,..., P
_L represents the value of each chip of the spread signal PN (t). The values p ₁ , p ₂ , p ₃ ,..., P _L of each chip are analog values and have a high degree of freedom, but in the following, in order to simplify the apparatus, the value of each chip is set to ± 1 = 2. Consider only the values.

A spread code sequence consisting of ± 1 that satisfies Equation 7 was obtained in the range of L = 4 to 64, and the results are as shown in the following table.

[0034]

[Table 1]

In this way, the spreading code usable in the present invention was actually determined based on the Nyquist's first standard analogy.

The above is a spreading code suitable for the present invention based on the Nyquist first standard analogy. Similarly, a spreading code suitable for the present invention based on the Nyquist second standard analogy can be considered. . FIG. 3 (c) shows a discrete autocorrelation function having pulse characteristics based on the Nyquist second criterion analogy. The spreading code based on the Nyquist second criterion satisfies the following equation, corresponding to Equation 7 in the case of the first criterion.

[0037]

(Equation 8)

In the above equation, L is an odd number, L 'takes a value other than "0", and a, b and c are arbitrary values. Also in the case of being based on the Nyquist second standard analogy, it is possible to obtain a spread code sequence of ± 1 that satisfies Equation 8 in the same manner as that obtained based on the Nyquist first standard analogy. In this manner, the spreading code that can be used in the present invention can also be determined based on the Nyquist second standard analogy.

FIGS. 4 and 5 show examples of the configurations of the transmitter 1 and the receiver 2 shown in FIG. 1, respectively, in the case where spreading codes determined based on the Nyquist's first standard analogy are used. FIG. First, referring to FIG. 4, the transmitter 1 includes a spreading code generator 21 having L chips.
A serial / pipeline / parallel converter 22 for sequentially converting serial data symbols d (t) as information to be transmitted into parallel signals at predetermined time intervals, and a serial / pipeline / parallel converter 2
2, a multiplication unit 23 for multiplying each data symbol converted in parallel by 2 with the spreading code from the spreading code generator 21, an addition unit 24 for adding each multiplication result from the multiplication unit 23, A modulating unit 25 for modulating a carrier with an output signal and outputting the modulated signal as a transmission signal.

Here, since the spreading code determined based on the Nyquist's first standard analogy is used, the number L of chips of the spreading code generator 21 is an even number. The spreading code generator 21 is constituted by, for example, a cyclic shift register of L bits (L chips). In this case, at the time of power-on reset, satisfy formula 1 spread code p (t), and more specifically, the spread code sequence p ₁ satisfying number _{7, p 2, p 3,} ..., p L is of L The spreading code series p ₁ , p ₂ , p ₃ ,..., P _L is set for each chip, cyclically shifted every clock for each chip, and taps every other chip (ie, every other bit). It is output.

As the spreading code generator 21, instead of a cyclic shift register, a conventionally used PN
A configuration can also be made by inputting the output of the code generator to an L-bit shift register. The spreading code generator 21 is generally configured by a logic circuit that outputs a logical value (0 or 1). However, since the spreading code is treated as having a logically arithmetic value, it is specifically specified. , The logic "1" may correspond to the arithmetic value "+1", and the logic "0" may correspond to the arithmetic value "-1".

The serial data symbol d input to the serial / pipeline / parallel converter 22
(T) is generally multi-valued, but when this is binary, the serial / pipeline / parallel conversion unit 22 is composed of an L / 2-bit serial-parallel conversion register and a delay circuit. be able to. More specifically, as shown in FIG. 6, a serial input parallel output shift register 61 of L / 2 bits configuration of the binary serial data symbols d (t) is entered for each time interval 2T _c, of FIG. 4 At the timing of the pulse _TL when the spread code generator 21 enters a specific state (for example, an initial state), L holding the L / 2-bit parallel output output from the serial input / parallel output shift register 61 is used. / 2 bit register (for example, L /
And two D flip-flops) 62, provides a delay time that is offset by one 2T _c (L / 2-1) by the number of delay circuit 63, a serial / pipeline / parallel converter 22
Can be configured. In FIG. 6, to generate the pulse T _L , for example, a pattern detector that detects that the bit pattern of the spreading code generator 21 has entered a specific state (for example, an initial state) may be used. Further, in order to generate a pulse interval of 2T _c may be 1/2 frequency clock of period T _c that is operated spreading code generator 21. Pulse time interval 2T _c are also clock to determine the symbols of the data symbol d (t) Rate (Phase). In this case, between the L / 2-bit configuration of the delay circuit 63 and the register 62 from 2T _c each pipeline parallel output of the time increments shifted L / 2 bits L chip time corresponding to T (= LT _c), It is possible to output stably. Also, the data symbol d (t)
Is multi-valued, the bits can be extended to symbols, so that the data symbol d (t) can be configured in the same manner as in the case of binary. Therefore, hereinafter, the serial data symbol d (t) will be described as binary unless otherwise specified.

Further, the multiplying section 23 is provided with L / 2 multipliers. Each of the L / 2 multipliers has a spreading code sequence shifted by two chip times (2T _c ), and The data from the corresponding serial / pipeline / parallel converter 22 is multiplied every chip time (T _c ). At this time, since the spreading code is output from, for example, a logical cyclic shift register, the logic “1” is changed to the arithmetic “+1”.
And treat the logic "0" as arithmetic "-1" and multiply. When the data symbol is binary, the output of the serial-to-parallel converter is generally treated as a logical value. "0" is treated as arithmetic "-1" and multiplied.

The adder 24 includes an adder 30 and D
/ A converter 31 is provided. The adder 30 has L
The outputs from the / 2 multipliers are added for each chip time (T _c ), and the adder 30 adds and outputs them.
The D / A converter 31 D / A converts the digital output from the adder 24.

Note that the L / 2 multipliers and one adder 24 constitute a product-sum circuit. On this occasion,
When the data symbol is binary, any two inputs to each multiplier can be handled by logical values, and the multiplication output has only +1 and -1 as arithmetic values. it can. For example, (+1) × (+1) = (− 1) ×
(-1) = (+ 1), (+1) × (−1) = (− 1) ×
The operation (+1) = (-1) is performed by adding arithmetic +1 to logic 1
It is obvious that the arithmetic operation can be realized by an exclusive NOR gate if arithmetic -1 corresponds to logic 0. Or exclusive O
The output may be handled by negative logic using an R gate. Furthermore, in these cases, since the input to the adder 30 is a binary logical value, it can be much simpler than a normal arithmetic adder. For example, an addition result can be obtained by counting the numbers of logic 1 and logic 0 of the logic value input to the adder 30 and taking the difference. More specifically, at the beginning of each chip time, the counter is initialized to “0”, and during the chip time, L / 2 multiplication results (logical values) are scanned. In this case, the counter is incremented by +1 (counting up). The output value of the adder is (L /
It can be either even or odd (depending on whether 2 is even or odd), and the sum of the numbers of logic 1 and logic 0 is constant with L, so it can be calculated only with the number of logic 1 It is possible to simplify the circuit by using such properties as: For example, if the number of logic 1 is n,
Is L / 2−n, the output of the adder 30 may be as follows.

[0046]

## EQU9 ## n- (L / 2-n) = 2n-L / 2

At this time, the least significant bit (LSB) of the output of the adder 30 does not change depending on the value of n, and only the upper bits excluding the LSB are involved in the operation. Therefore, it can be directly obtained by a combinational circuit that satisfies Equation 9. Alternatively, the initial value of the counter is [-L / 4] (where [x]
Is an integer value obtained by truncating (rounding down) fractions of x.
When the number n of logic 1 is counted (added), n + [− L / 4] is obtained, and this is doubled, and the least significant bit (LSB) is obtained.
Equation 9 can be obtained by fixedly setting 0 or 1 depending on whether L / 2 is even or odd. In particular, when L / 2 is an even number, even if LSB is not inserted, it is simply １ ／ of Equation 9, and the adder 3
Obviously, it is only necessary to adjust the sensitivity setting of the D / A converter 31 following 0.

In general, the product-sum circuit is complicated, but as long as binary data symbols as in the above example are used, the product-sum circuit can be constituted by an extremely simple (simple) circuit. Further, even if the data symbol is multi-valued, one of the inputs (spreading code sequence) of each multiplier is binary, so that the multiplier can be substantially constituted by a simple (arithmetic) sign inverting circuit, It is not as complicated.

The modulating section 25 is generally constituted by an analog circuit, and multiplies the analog output from the D / A converter 31 of the adding section 24 by, for example, cosωt (ω = 2πf, f is a carrier frequency). , Which modulates the carrier and outputs it as a transmission signal. The D / A converter 3 of the adder 24 described above
1 is when the output of the adder 30 is a digital signal,
It is provided to convert this into an analog signal and input it to the modulator 25 having an analog circuit configuration. In this case, the D / A converter 31 is not the adder 24 but the modulator 2
It may be provided on the fifth side.

[0050] Referring now to FIG. 5, the receiver 2 includes a demodulator 41 which receives a transmission signal transmitted from the transmitter 1 demodulates it, and a matched filter 42, a sampler S ₂ . Here, the demodulation unit 41 includes a multiplier 51 that multiplies a transmission signal transmitted from the transmitter 1 by cosωt (ω = 2πf, f is a carrier frequency), and a low-pass filter (or integrator) 52, The coherent detection (demodulation) is performed by multiplying the transmission signal (that is, the reception signal) by cosωt and passing this through a low-pass filter (or integrator) 52. Note that a carrier must be reproduced for synchronous detection, and there are various well-known methods. As an example, a 2f (f is a carrier frequency) component is obtained by passing the amplified received signal through a squaring circuit, and a narrow band filter having a center frequency of 2f is used to obtain 2f.
It is also possible to extract only the component, divide it by a frequency divider, and obtain the carrier f.

The matched filter 42 is, for example,
Sampler S that samples the output from low-pass filter (or integrator) 52 for each chip time (T _c )
₁ and a multi-level tapped delay line (Multi
level tapped delay line) 54 and a summing circuit 55. The sampler S ₁ is formed of, for example, a 10-bit A / D converter including a sample and hold circuit.
In the sampler S _1, as a method of determining the sample timing, common is can be used in which DLL (Delay Locked Loop) for use in direct spread system, in particular 1 △ form of DLL are suitable. The multi-level tapped delay line 54, the number of chips of the spreading code in the transmitter 1 is constituted by a delay circuit L stages tailored to (code length) L, the sampler S ₁ is for example 10-bit A / When the delay circuit is constituted by a D converter, each stage of the delay circuit is constituted by a 10-bit word. The summation circuit 55 also calculates the values of the spreading code sequences p ₁ , p ₂ , p ₃ ,..., P _L (the arithmetic value “1” or “−1”) during each sampling interval (T _c ) of the sampler S _1. "), The corresponding tap output of the multi-level tapped delay line 54 is added or subtracted. In the above configuration example, the matched filter 42 is of a digital signal processing type (accurately, discrete time), and the output of the low-pass filter (or integrator) 52 is sampled to be a multi-value input. However, instead of this, the matched filter 42 is replaced with a SAW (surface acoustic wave) filter or SA.
It can also be constituted by an analog-like (precisely, continuous-time) device such as a W convolver. In this case, the sampler S ₁ is not required.

The sampler S ₂ samples the output from the matched filter 42, that is, the output from the summing circuit 55, every two chip times (2T _c ).

Next, the operation of the spread spectrum communication system having such a configuration and the operation of the transmitter 1 and the receiver 2 will be described. In the following, the data symbol d (t)
Is a binary code, and a code determined based on Nyquist's first standard analogy (code length L is an even number) is used.

First, the transmitter 1 transmits a binary serial data symbol d (t) as information to be transmitted to the serial / pipeline / parallel converter 22 at a time interval of 2T _c (two chip times). input. As a result, the serial / pipeline / parallel conversion unit 22 outputs
As shown in (1), data is output in parallel in a pipeline manner and input to the L / 2 multipliers of the multiplier 23, respectively. On the other hand, in the spreading code generator 21 having L chips, L spreading codes p ₁ , p ₂ , p ₃ ,..., P _L are initially set in advance, and these L spreading codes p ₁ , p _{2, p 3, ..., p} L
Is shifted by one chip per chip during operation, and is output as a tap output every other chip. L /
The two multipliers, two chip time 2T _c increments shifted data and a tap output by multiplying each data symbol from the spreading code generator 21 corresponding thereto from the serial / pipeline / parallel converter 22 Perform spreading (ie, modulation). That is, in FIG.
Modulate the spreading code output from the th tap,
In step 2, the spreading code output from the (L-1) th tap is modulated. Assuming that the chip phase is represented by an integer of “1” to “L”, each of L / 2 multipliers starts data modulation when the chip phase becomes “1”, and the chip phase “ L "terminates data modulation. At the next moment, the chip phase becomes "1" again, and the next data is similarly modulated. In this way, the data timing is shifted by two chips as shown in FIG. 7, and the (L / 2 + 1) th data (data L / 2)
+1) returns to the first tap and performs multiplication.
Data having a time difference of such 2T _c serial /
By performing pipeline / parallel conversion and multiplying by the corresponding spreading code, the output from the matched filter 42 on the receiver 2 side can be reproduced as it is as a serial data symbol, as described later. There is no need for inverse transformation on the two sides.

The outputs from the L / 2 multipliers of the multiplier 23 are input to the adder 24, and the adder 30 of the adder 24 adds the outputs from the multipliers for each chip time. .
FIG. 8 is a diagram showing an example of the output from the adder 30. The output from the adder 30 generally has a multi-value. Further, in this example, the output from the adder 30 is a digital value as shown in FIG. The output from the adder 24 is
In addition, since the modulating unit 25 is generally constituted by an analog circuit, if the output from the adder 30 is a digital value as shown in FIG. , And then input to the modulation unit 25.

The modulator 25 modulates the carrier having the carrier frequency ω with the output from the adder 30.
Is multiplied by the output from the adder 30, and this is transmitted as a transmission signal, for example, in the form of a radio wave.

Next, the receiver 2 receives the transmission signal transmitted from the transmitter 1 in the form of a radio wave, for example, in the form of a radio wave as a reception signal using, for example, an antenna, amplifies the signal, and inputs the signal to the demodulation unit 41.
In the demodulation unit 41, the multiplier 51 adds cosω to the received signal.
multiply by t and apply this to a low-pass filter (or integrator)
By passing through 52, synchronous detection, that is, demodulation is performed. That is, when the result of addition as shown in FIG. 8 is output from the adder 30 of the transmitter 1 and the result is modulated and transmitted, the demodulator 41 of the receiver 2 outputs the adder 30 of the transmitter 1 The received signal is demodulated into a signal as shown in FIG.

The signal demodulated in this way is applied to the matched filter 42. In the matched filter 42,
First, the demodulated signal is sampled by the sampler S ₁ every chip time (T _c ). In this case, the sample timing of the sampler S ₁ can be determined by using, as described above, for example, a DLL. In the case of using a DLL in the sample timing determination sampler S _1, in the present embodiment, since it is intended to spread the signal is shifted two chips at a time, the synchronization point, present in 2 chip intervals. This is a clear advantage over the conventional direct spreading scheme.
That is, in the conventional direct diffusion system, in the worst case, a phase error close to the L chip is corrected.
Although a long time was required for synchronization, in the present embodiment, even in the worst case, it is sufficient to correct the phase error of two chips, and synchronization can be completed much earlier than in the past. This means that resynchronization can be performed very easily in a wireless (radio wave) environment in which loss of synchronization easily occurs, which is extremely advantageous in many applications.

One chip time (T _c ) by the sampler S ₁
The signal obtained as a result of each sampling is applied to an L-stage multi-level tapped delay line 54, and the summation circuit 55 outputs a multi-level tapped delay line at each sample interval, that is, during one chip time (T _c ). Each tap output of the delay line 54 is converted to a value of the corresponding spread code sequence p ₁ , p ₂ , p ₃ ,..., P _L (arithmetic value “1” or “−”
Addition or subtraction according to 1 ″).

Thereafter, the sampler S ₂ samples the output from the matched filter 42, that is, the output y (t) from the summation circuit 55, every two chip times (2T _c ). At this time, the output y (t) from the matched filter 42 becomes a discrete autocorrelation function of the target data p (t) as shown by a broken line in FIG. As described above, since the spreading code is determined based on Nyquist's first standard, τ
Becomes “0” every 2n × T _c (n = 1, 2, 3,..., L / 2-1). Accordingly, even when the output y (t) from the matched filter 42 is sampled every two chip times (2T _c ), which is much shorter than the time interval T (= LT _c ) determined by the code length of the spreading code, Interference does not occur. That is, as will be described later, it is possible to maintain the same performance as that of the conventional direct spreading method without deteriorating the error rate characteristics, the interference wave removing ability, and the like. Thus while the need without deteriorating the error rate characteristic and the interference wave removing ability and the like, transmits the data at a time interval of 2T _c, significantly enhances be compared with the conventional direct spread system the transmission speed is capable of receiving Can be.

Note that, by sampling every two chip times (2T _c ) in the sampler S ₁ , actually, only one of the two sample (even or odd) columns has no intersymbol interference. Therefore, the one with less intersymbol interference must be selected. Although this is not described in detail, it can be easily distinguished by the following principle. In other words, the decoded symbol sequence with less intersymbol interference takes a binary value in principle (ideally) (actually, noise,
2 due to code distortion, sample timing error, etc.
On the other hand, if the intersymbol interference is large, the deviation from the same binary value becomes abnormally large. Therefore, the deviation value or variance value of the two sequences is obtained, and the smaller one may be selected. . More practically, the deviation or variance of the absolute value of the signal itself may be calculated without calculating the deviation from the binary value. This is because the decoded symbol sequence ideally has a positive-negative symmetric value, and its absolute value becomes a constant value.

[0062] Note that method if the matched filter 42 is of digital configuration, the input to the sampler S ₂ becomes a discrete time signal, selecting one of two sample sequences, as described above Can be used,
If matched filter 42 is of analog configuration, the input to the sampler S _2, since a continuous time signal, it is impossible to use the technique of choosing between the two sample sequences. However, this is generally the case with eye patterns (Eye Pa
ttern) is the same as determining the timing of the sample with the most open eye, and, as is well known, sampling at the most open eye, ie, where there is minimal intersymbol interference. Can be solved by

The inventor of the present application has shown that the performance of the spread spectrum communication system of the present embodiment when the code determined based on the Nyquist's first standard analogy (the number of chips L is an even number) is used for the spreading code. Was evaluated. Less than,
The performance evaluation results are shown.

First, the bit error rate BER under white Gaussian noise was evaluated. In this case, the transmitter 1 uses p (t) of Expression 1 to obtain p (t) when the information is “1”,
"0" was transmitted -p (t) when, in the receiver 2 side, there is no intersymbol interference, both sides of the noise power spectral density and N _o / 2, the energy of the p (t) and E _b Then, the bit error rate BER is theoretically expressed by the following equation. In the following equation, erfc is an error function.

[0065]

(Equation 10)

The simulation in this case was performed with the spreading codes as shown in Table 1 for the three cases where the code length L of the spreading codes was "8", "32", and "64".

FIG. 9 shows the theoretical value (shown by a solid line) of the bit error rate BER obtained by the equation (10) and the above simulation results. As is clear from FIG. 9, the simulation results match the theoretical values.

In this case, L is "8", "32", "6".
4 ", the information transmission rate _Rb is 1 / (2
T _c ), and the L of the system of the present invention using the simple direct spreading method or the simple polar method is "6".
When compared with the case of 4 ", it was found that in the system of the present invention, the transmission rate is 32 times that of the conventional system, but the bit error rate is exactly the same as the conventional system.

Further, the bit error rate in the presence of a single sine wave interference and white Gaussian noise was evaluated. That is, the performance of the system of the present invention was evaluated when white Gaussian noise was further added with single-frequency interference wave interference. Here, the above-described digital processing filter is used as the matched filter 42 on the receiver 2 side of the system. FIGS. 10A, 10B and 10C show that the number of chips L of the spreading code is "8", "32" and "64", respectively.
The average error rate BER when the period of the interference wave, that is, the single sine wave is changed in the range of 5 to 90 times the chip period _Tc
showed that. 10 (a), (b) and (c) show that the magnitude of the simultaneously added white Gaussian noise is different, and FIG. 10 (a) shows that the E _b / N _o ratio (SIR) is 11 dB.
In the case of, FIG. 10B shows that E _b / N _o (SIR) is 15 dB.
Cases, FIG. 10 (c) E _b / N _o ratio (SIR) is 30d
The case of B is shown. FIG. 11 shows all the results of FIGS. 10 (a), (b), and (c) at different scales in order to clarify the characteristics when the interference wave is dominant. From FIG. 11, L is “32” in the region where the interference wave is dominant.
In the case of, L is advantageous by 4 dB with respect to the case where L is “8”,
Further, when L is "64", it is further 4 dB more advantageous than when L is "32". From these results, it was found that the interference wave removal characteristics of the system of the present invention were exactly the same as those of the conventional system using the direct spreading method. However, this is the case where the amount of delay is less than the information symbol interval when there is a multipath.

As is evident from the above, in the system of the present invention, it is desirable to increase the number L of spreading code chips as much as possible in order to improve the interference wave removing capability as in the conventional case. However, in the present invention, even if the number L of spreading code chips is increased as much as possible to remove the interference wave and the interference wave removal capability is increased, neither deterioration in the error rate characteristic nor reduction in the transmission speed is caused. That is, transmission is performed at a high speed of 1 / (2T _c ) based on the inter-chip period without being affected by the number L of chips, information can be accurately reproduced, and the information transmission speed and bit error rate characteristics are degraded at all. Without this, the anti-interference wave characteristics can be improved.

Also, as can be seen by comparing FIGS. 4 and 5 with FIG. 12, the system of the present invention does not increase the complexity of the apparatus as compared with the conventional system of the direct spreading system.
In particular, no structural changes are required on the receiver 2 side. When the matched filter 42 of the receiver 2 is of a digital type, the linearity of the system that can withstand multi-level signals is ensured, and as described above, the error rate characteristics and the interference wave removal capability are improved. Things can be maintained.

FIG. 3 based on Nyquist's first standard
Although the spread spectrum communication system of the present invention using the spreading code having the autocorrelation function as shown in FIG. 3B has been described, the discrete autocorrelation function as shown in FIG. Similarly, the spread spectrum communication system of the present invention can be configured even when a spreading code having That is, the transmitter 1 of FIG.
In the configuration of (1), a spreading code of the number L of chips determined based on the Nyquist second criterion is set in the spreading code generator 21, and L spreading codes are generated at a timing shifted by one chip time. The serial / pipeline / parallel conversion unit 22 converts a serial data symbol as information to be transmitted into L parallel data shifted by one chip time and outputs it.
In addition, the multiplying unit 23 and the adding unit 24 output L data symbols which are output in parallel from the serial / pipeline / parallel converting unit 22 and shifted by one chip time from the spreading code generator 21 corresponding thereto. What is necessary is just to take the product with the spreading code and add L products.
Further, in the configuration of the receiver 2 of FIG. 5, the sampler S _2, it is sufficient to perform the sampling at one chip intervals. In addition, when based on (the analogy of) the Nyquist's second criterion, as shown in Expression 8, the adjacent symbols receive interference. This is due to the L 'is not 0, the amount of the interference is known in advance, it is possible to remove the intersymbol interference included in the output of the sampler S _2.

In this case, when the first reference spreading code is used, as described above, the information transmission rate is ２ ， of the chip rate, that is, 1 / (2T _c ), but the second reference spreading code is used. Is used, transmission can be performed every one chip time _Tc , and the information transmission speed can be made to match the chip speed 1 / _Tc . Therefore, when the second reference spreading code is used, higher-speed transmission can be performed.

[0074]

As described above, claims 1 to 1 are provided.
According to the invention described in No. 0, the transmission rate can be further increased without deteriorating the error rate characteristics, the interference wave removing ability, and the like. In particular, in the invention according to claims 2, 4, and 7, the spreading code used is one determined based on Nyquist's first criterion, the number of chips of the spreading code is an even number, and the spreading code is Since the side lobe of the autocorrelation function becomes “0” every two chip time intervals,
Transmission at intervals of two chips becomes possible, and in the invention according to claims 3, 5, and 8, the spreading code used is one determined based on the Nyquist second standard, and the number of chips of the spreading code is Since an odd number is used as the spreading code such that the side lobe of the autocorrelation function becomes “0” at one-chip time intervals, transmission can be performed every one-chip time.

When the matched filter of the receiver is of a digital type as in the ninth aspect of the present invention, the linearity of the system that can withstand a multilevel signal is ensured.
As described above, it is possible to maintain good error rate characteristics and good interference wave removal capability. On the other hand, when the matched filter is of a continuous-time analog type as in the invention of claim 10, the matched filter can have a simple configuration.

[Brief description of the drawings]

FIG. 1 is a diagram for explaining a basic concept of a spread spectrum communication system according to the present invention.

FIG. 2 is a configuration diagram of an embodiment of a spread spectrum communication system according to the present invention.

FIG. 3 is a diagram showing an autocorrelation function.

FIG. 4 is a diagram illustrating a configuration example of a transmitter in FIG. 1;

FIG. 5 is a diagram illustrating a configuration example of a receiver in FIG. 1;

FIG. 6 is a diagram illustrating a configuration example of a serial / pipeline / parallel conversion unit of the transmitter in FIG. 4;

7 is a diagram showing data output in parallel in a pipeline from a serial / pipeline / parallel converter of the transmitter in FIG. 4;

FIG. 8 is a diagram illustrating an example of an output from an adder of the transmitter in FIG. 4;

FIG. 9 is a diagram showing performance evaluation results of the spread spectrum communication system configured from FIGS. 4 and 5;

FIG. 10 is a diagram showing performance evaluation results of the spread spectrum communication system configured from FIGS. 4 and 5;

FIG. 11 is a diagram showing performance evaluation results of the spread spectrum communication system configured from FIGS. 4 and 5;

FIG. 12 is a configuration diagram of a conventional spread spectrum communication system using a direct spreading method.

FIG. 13 is a diagram showing an outline of processing of the spread spectrum communication system of FIG. 12;

FIG. 14 is a diagram illustrating an example of an autocorrelation function and an output from a matched filter.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 Transmitter 2 Receiver 21 Spreading code generator 22 Serial / pipeline / parallel conversion part 23 Multiplication part 24 Addition part 25 Modulation part 30 Adder 31 D / A converter 34 Multiplier 41 Demodulation part 42 Matched filter 51 Multiplier 52 Low-pass filter (or integrator) 54 Multi-level tapped delay line 55 Summing circuit 61 Serial input parallel output shift register 62 Register 63 Delay circuit S ₁ , S ₂ sampler L Number of chips of spreading code T _c Period between chips

Claims (10)

(57) [Claims] 1. A transmitter comprising a transmitter and a receiver, wherein the transmitter transmits a signal, which is obtained by multiplying information and a spread code having a predetermined number of chips, and is subjected to spread spectrum modulation to a receiver as a transmission signal. The receiver reproduces the transmission signal transmitted from the transmitter using matched filtering, and the spreading code has a side lobe of an autocorrelation function at a sampling point having a constant period. A spread-spectrum communication system, wherein only ones that are "0" are used. 2. The spread-spectrum communication system according to claim 1, wherein a code determined based on Nyquist's first criterion is used as the spread code, and the number of chips of the spread code is an even number. A spread-spectrum communication system, wherein a spreading code whose side lobe of an autocorrelation function becomes "0" at intervals of two chips is used. 3. The spread-spectrum communication system according to claim 1, wherein the spreading code is a code determined based on Nyquist's second standard, and the number of chips of the spreading code is odd. A spread-spectrum communication system wherein a spreading code whose side lobe of an autocorrelation function becomes "0" at one-chip time intervals is used. 4. The spread-spectrum communication system according to claim 1, wherein the transmitter is configured based on Nyquist's first criterion, and the side lobe of the autocorrelation function is “0” every two-chip time intervals. A number L of spreading codes are set, and L / L
Spreading code generating means for generating two spreading codes, and serial pipeline parallel for converting serial data symbols as information to be transmitted into L / 2 parallel data shifted by two chip times for output Conversion means;
The product of the L / 2 data symbols output in parallel from the serial pipeline / parallel conversion means and shifted by two chips each time and the spread codes from the spread code generation means corresponding to these, respectively, is obtained. A spread-spectrum communication system, comprising: a sum-of-products means for adding a number of products; and a modulating means for modulating and transmitting a carrier wave by an output of the sum-of-products means. 5. The spread spectrum communication system according to claim 1, wherein the transmitter is configured based on Nyquist's second criterion and has a number of chips whose side lobe of an autocorrelation function is “0” in one chip time interval. L spreading codes are set, and spreading code generating means for generating L spreading codes at a timing shifted by one chip time, and L spreading codes which shift serial data symbols as information to be transmitted by one chip time. Serial pipeline / parallel conversion means for converting the data into parallel data and outputting the data; and L data symbols which are output in parallel from the serial pipeline / parallel conversion means and shifted by one chip each time, and the spreading codes respectively corresponding thereto. A product-sum means for taking the product with the spread code from the code generation means and adding the L products, and carrying by the output of the product-sum means And a modulating means for modulating a transmission wave and transmitting the modulated signal. 6. The spread spectrum communication system according to claim 1, wherein the receiver is a demodulating means for demodulating a received signal, and matched filtering corresponding to a predetermined number of chips of a spreading code used on the transmitter side. a matched filter for performing processing on the demodulated signal, and a sampling means for reproducing the sampled data symbols output of the matched filter at a predetermined time interval, sampling means, predetermined chip based on prior symbol a constant cycle A spread spectrum communication system, wherein sampling is performed at time intervals. 7. The spread spectrum communication system according to claim 6, wherein said spread code is determined based on Nyquist's first criterion, and said sampling means performs sampling at a two-chip time interval. A spread spectrum communication system, characterized in that: 8. The spread spectrum communication system according to claim 6, wherein said spread code is determined based on Nyquist's second criterion, and said sampling means performs sampling at one-chip time intervals. A spread spectrum communication system, characterized in that: 9. The spread spectrum communication system according to claim 6, wherein said matched filter is of a digital type, and has a sampling circuit for sampling a demodulated signal at every one-chip time interval; A multi-level tapped delay line having a combined number of stages and sequentially delaying the sampling signal from the sample circuit, and a summing circuit for summing tap outputs from the multi-level tapped delay line A spread spectrum communication system, comprising: 10. The spread spectrum communication system according to claim 6, wherein said matched filter is constituted by a continuous-time analog type device.