JPH07221810A - Linear modulation mode communication system having symbol timing recovery function using dispersion operation - Google Patents

Abstract

PURPOSE:To realize symbol timing recovery in an excellent way even under fading by improving the recovery accuracy of the symbol timing recovery based on a criterion close to an absolute reference. CONSTITUTION:The communication system is provided with a transmission system 1 in which a synchronization symbol signal and an information symbol signal of a constant energy symbol series are generated and a carrier is subjected to linear modulation with a slot signal formed by arranging the symbol signals and with a reception system 2 sampling a demodulated symbol signal in a sampling timing based on the synchronization symbol signal included in a demodulation symbol signal obtained by demodulating a transmission signal. The transmission system 1 samples the demodulated symbol signal at a rate being a multiple of M of the symbol rate, calculates the dispersion in sample values in each of M-sets of sample value sets based on the sample value by a prescribed symbol as to each of M-sets of sampling points per symbol and outputs one of sampling points corresponding to the sampling set representing a dispersion smaller than a prescribed threshold value as an optimum symbol timing.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION This invention is applicable to cellular telephones and JSs.
The present invention relates to a communication device that employs linear modulation such as MR. Further, the present invention relates to a transmission / reception device that takes measures against fading, and more particularly to a mobile radio device and system. Further, the present invention is
The present invention relates to a system, such as a TDMA (Time Division Multiple Access) system, in which symbols of a predetermined pattern such as a preamble are embedded in a slot signal for transmission.

[0002]

2. Description of the Related Art As a symbol timing reproduction method in linear modulation, the maximum amplitude method: MAM (Maximum Amplitud)
e Method) and difference method: WDM (Wave Difference Meth)
od) etc. are used. The contents of these methods are described in the literature as “Technical Research Report of the Institute of Electronics, Information and Communication Engineers (Wireless Communication System), vol.92 No. 411, January 20, 1993, pp.43 to pp.48, Seiichi Sampei, Kamil.
O Feher, especially page 45, left column, line 29 to right column, line 32 ”.

In such MAM and WDM, a logarithmic likelihood function (logaristic likelihood) is used as a function for probabilistically displaying a predetermined standard.
It is based on the relative comparison between the calculation results of sample values using (function). Therefore, even when the received symbol is sampled at the true symbol timing to obtain the symbol timing, the symbol timing is relative and is difficult to serve as an absolute reference.

Further, under fading in mobile communication, the symbol itself may be distorted and it may be difficult to reproduce the symbol timing.

[0005]

SUMMARY OF THE INVENTION The present invention has been made in view of the above-mentioned points, and an object of the present invention is to perform symbol timing reproduction based on a judgment criterion close to an absolute reference and reproduce its reproduction accuracy. And improve
An object of the present invention is to provide a communication system capable of excellent symbol timing recovery even under fading.

[0006]

A communication system according to the present invention generates a sync symbol signal composed of symbols having substantially constant energy and an information symbol signal carrying a predetermined information signal at a predetermined symbol rate, and these symbol signals are generated. A transmission system that linearly modulates a carrier wave by a slot symbol signal formed by linking each other to generate a transmission signal, and a demodulation symbol signal by demodulating the transmission signal, and the synchronization symbol signal included in the demodulation symbol signal A communication system comprising a reception system for sampling the demodulated symbol signal at a sampling timing based on, wherein the transmission system samples the demodulated symbol signal at M (M> 1) times the predetermined symbol rate. The oversampling means and the system in the oversampling means. For each of the M sampling points per volume, the variance of the sample values in each of the M sets of sample values consisting of sample values for a predetermined symbol is calculated and corresponds to the sample value set showing a variance value smaller than a predetermined threshold value. And a timing detecting means for deriving one of the sampling points as the optimum symbol timing.

[0007]

According to the linear modulation communication system having the symbol timing recovery function using the dispersion calculation according to the present invention, the demodulated symbol signal is sampled at M (M> 1) times the predetermined symbol rate in the receiving system. For each of the M sampling points per symbol in the oversampling, the variance of the sample values in each of the M sets of sample values consisting of the sample values for the predetermined symbol is calculated, and the sample value indicating the variance value smaller than the predetermined threshold value. One of the sampling points corresponding to the set is derived as the optimum symbol timing.

[0008]

The present invention will be described in detail below with reference to the drawings. FIG. 1 shows a QAM (Quadra) according to an embodiment of the present invention.
FIG. 1 is a schematic block diagram showing the configuration of a Tture Communication Amplitude Modulation (TDMA) communication system. In FIG.
The transmission system 1 has a data generation system (not shown) for generating bit data which is a digital binary signal indicating a predetermined information signal such as a voice signal to be transmitted, and the generated data is transmitted to the mapping circuit 11. Forward. Mapping circuit 1
1 maps the input bit data and transfers it to the frame signal forming circuit 12. In this mapping, for example, 4
Converting to symbol data bit by bit, in detail, the predetermined 2 bits in the 4 bits are code-converted into a baseband signal, that is, symbol data corresponding to an in-phase component in 16QAM, and The other 2 bits of 16 QAM
e) Code conversion into a baseband signal corresponding to the component, that is, symbol data. In short, the input binary signal is converted into a two-sequence four-valued signal and output.

The frame signal forming circuit 12, which is also called a burst format circuit, supplies the mapping output data to the digital-analog (D / A) converter 13 in a predetermined burst format according to the frame of the TDMA system. In such formatting,
Preamble signals, pilot signals not described in detail, and signals used for various synchronization controls, communication controls, etc. (sometimes referred to as data or control signals in distinction from the above information signals) are performed for each series of mapping output data. It is added to the mapping output data to form a TDMA frame signal. Formatted data corresponding to the in-phase component is transferred to the D / A converter 13 as I channel symbol data, and formatted data corresponding to the anti-phase component is transferred to the D / A converter 13.

The D / A converter 13 converts the transferred symbol data of the I and Q channels into an analog signal and outputs an LP signal.
Supply to F14. The LPF 14 passes only a predetermined low frequency region of the input signal, that is, removes its harmonic component and supplies it to the QAM modulator 15. The QAM modulator 15 uses the oscillation output signal of the local oscillator 16 for LP
QAM modulation is performed using the output signal of F14 as a signal to be transmitted. The oscillation output signal of the local oscillator 16 has a predetermined carrier frequency that is predetermined in the communication device, and this becomes the center frequency in QAM modulation. The modulation output of the QAM modulator 15 is applied to the transmission antenna 17 and radiated as a transmission wave.

The reception system 2 captures a transmission wave of the transmission system 1 that has arrived via, for example, a mobile propagation path by a reception antenna 21. The antenna 21 uses this as a received signal for Q
Guided to the AM demodulator 22, the QAM demodulator 22 demodulates the received signal based on the oscillation output signal of the local oscillator (or reference carrier recovery circuit) 23 in accordance with QAM. The oscillation output signal of the oscillator 23 has the same frequency as the oscillation output signal of the modulation local oscillator 16 in the transmission system 1. The demodulated output of the demodulator 22 is supplied to the low-pass filter 24 so that only a predetermined low frequency range is analog-digital (A
/ D) is supplied to the converter 25.

The A / D converter 25 serves as one of the components of the oversampling means and the symbol rate sampling means as will be apparent from the following description. The sampling clock of the A / D converter 25 is
It is obtained by the delay / frequency setting circuit 252 which delays, divides, or multiplies the clock signal based on the clock signal having the predetermined frequency and the duty cycle from the clock signal generation source 251. The delay / frequency setting circuit 252 is simply composed of a divider, and appropriately sets the frequency of the sampling clock to M times the symbol data rate fs in transmission. Therefore, the A / D converter 25 operates with a period in which one symbol period of the supplied demodulated signal is divided into approximately M equal parts as a sampling cycle. Conversion output of A / D converter 25, that is, LPF2
The sample signal of the output signal of No. 4 is transferred as I and Q channel demodulation data to the optimum timing detection unit 26 and the fading correction processing circuit 27 which serve as timing detection means.

The optimum timing detection section 26 obtains optimum synchronization timing information of received symbols and frames as will be described in detail later, and controls the delay amount and the set frequency in the delay / frequency setting circuit 252 according to such timing information. The fading correction processing circuit 27 corrects the distortion of the input data due to fading with respect to the amplitude and phase other than the fluctuation component of the time axis, and transfers the data to the demapping circuit 28. The demapping circuit 28 is
The transferred symbol data is inversely converted by the mapping circuit 11, outputs bit data, and transfers it to a decoding system (not shown).

An example of the transmission / reception mode in such a TDMA communication system is shown in FIG. The figure shows a case where the number of multiplexing is 2 in so-called ping-pong transmission. In FIG. 2 (a), each communication time zone (time slot) from a base station which carries channels 1 (CH1) and 2 (CH2) to a mobile station in one frame, and a mobile station which carries CH1 and CH2 Each communication time zone to the base station is shown. Further, it is shown that a preamble signal (shaded portion) as a synchronization symbol signal which is a known signal for both the transmitting and receiving systems is arranged at the beginning of these communication time zones, and an information symbol signal is arranged subsequently thereto.

FIG. 2B shows a form of a transmission signal from the base station of CH1 to the mobile station in the communication time zone (downstream slot), which corresponds to the output of the QAM modulator 15 of FIG. . As can be seen from this figure, the transmission system 1
Is the symbol data rate fs, that is, the symbol period T
A symbol sequence of a preamble signal composed of symbol data having a complex baseband signal whose value can be changed for each s and a symbol sequence of an information signal are sequentially transmitted and output.

The sequence of preamble signal symbols is shown in FIG.
It is generated based on the I and Q channel data as shown in (c) and (d). These data are
Output data I (n) of the frame signal forming circuit 12 of FIG.
And Q (n). Further, these data have an amplitude with a maximum peak level of 3 [unit is arbitrary] and a minimum peak level of -3, and have no other level at least at a true sampling point (symbol point). That is, the absolute values of the amplitudes of the I (n) and Q (n) data signals are constant. Therefore, the preamble signal as shown in FIG. 2B obtained by combining these has a constant energy level (I ² + Q ² ). In the figure, k = 0, 1, 2, ..., 13 indicates an example of a number (symbol number) indicating the time axis position of each symbol in the preamble part in the time slot.

FIG. 2E shows I and Q channel data obtained by QAM demodulation when the reception system 2 in the mobile station receives the transmission signal from the base station as shown in FIG. 2B. There is. The demodulated data is the QAM demodulator 2 of FIG.
2 corresponding to output signals I ′ (t) and Q ′ (t). In the figure, the actual demodulation waveforms of the symbols of k = 7 and 8 among the symbols of the preamble signal in FIG. 2D are simply drawn. This demodulated waveform signal is A /
The D converter 25 is oversampled with M times the symbol rate, that is, Ts / M as a sampling period. In the figure, m = 0, 1, 2, ..., M−1 is a number (sampling point number) indicating the time axis position of the sampling point in one symbol in the oversampling. The demodulated data oversampled in this way is the symbol timing recovery circuit 26.
It is supplied to A.

Next, the configuration and operation of the optimum timing detection section 26 and the timing recovery processing by this detection output will be described. FIG. 3 is a diagram showing a specific configuration of the detection unit 26, specifically, the symbol timing reproduction circuit 26A.
In FIG. 3, the demodulated data signal I ′ (m) from the A / D converter 25 is supplied to one square calculator 261 and the demodulated channel data signal Q ′ (m) is supplied to the other square calculator 262. The operation output of each operation unit is the adder 2 respectively.
At 63 they are added up. As a result, a signal of the sum of squares of the symbol data indicating the energy level is obtained. This sum of squares signal is a sampling cycle of the A / D converter 25 (M of the symbol rate at the time of preamble reception).
Be one which level changes at intervals determined by multiplying), m per symbol is passed to the distributed processing unit 264S as the signal value y _m that counts from 1 to M-1.

In the distributed arithmetic processing unit 264S, the distributed value var _m of each symbol is calculated for each number of _m (var _m ← VAR [y _m ]; m = 0, 1, 2, ...,
M-1; VAR is a variance value operator). In the variance minimum value detection processing unit 265S, which is a selection unit, the smallest number var _min among the obtained variance values is selected, and the number m of m indicating that is selected.
get _min (var _min / m _min ← MIN [var _m ]
/ M; m = 0, 1, 2, ..., M-1; MIN is a minimum value selection operator).

The threshold comparison processing unit 266S compares the selected minimum dispersion value var _min with the dispersion threshold var _th to determine the magnitude relationship between the two. var _min <v
If it is determined as art _th , the optimum timing information update processing unit 267S passes m _min indicating the minimum dispersion value as a pointer m _op indicating the optimum symbol synchronization timing, and passes it to the delay / frequency setting circuit 252 (FIG. 1). . That is, the variable m _op is updated by m _min which stores the variance minimum value. On the other hand, when it is determined that var _min ≧ var _th , the minimum variance value selected this time is not treated as an optimum value, a new sample value set is taken in, and the same calculation is repeated.

The processing units 266S and 267S realize a derivation means for finally deriving the optimum symbol timing. With the optimum symbol timing information signal m _op , the delay / frequency setting circuit 252 changes the delay amount and the frequency shift amount of the output signal with respect to the input clock signal. That is, a sampling for activating the operation of the A / D converter 25 to sample each symbol of the information signal symbol sequence following the preamble signal symbol sequence at the point indicated by the pointer m _op among the M sampling points in the symbol. It emits a pulse. A / D converter 2 when receiving such an information signal
The sampling timing of 5 is fixed until the pointer m _op is updated by the processing of the next incoming preamble signal. As a result, the symbol data after the preamble is satisfactorily sampled in the A / D converter 25 for each time slot at a rate equal to the symbol rate without error on the time axis.

Details of the processing units 264S and 265S are shown in FIG.
This will be described with reference to the flowchart of. In FIG. 4, first, the signal values y _m (k) sequentially transferred in time series are fetched in a form suitable for arithmetic processing (step S1). As described above, the A / D converter 25 performs oversampling at a speed higher than the actual symbol rate and M (M> 1) times the symbol transmission rate. That is, assuming that the sampling cycle of the A / D converter 25 is T _{A / D} ,

[0023]

[Equation 1]

[0024]

[Equation 2]

[0025] At this time, the time t _{k, m} sampled by the A / D converter 25 is

[0026]

[Equation 3]

It is Sum of squares y of signals (data) I ′ (m) and Q ′ (m) sampled at such time
_{Taking m} (k) as u (t _{k, m} ), that is, u (k, m), the following set of M sample values is obtained.

[0028]

[Equation 4]

It is important to determine which m represents the true symbol timing using such M sets of samples. In other words, u (k, 0) to u
By grasping the properties of each of the sets up to (k, M-1), the optimal 1 of the number of m points to these sets.
One is selected as the optimum sampling point.

In the processing after step S2, the variance of each set is calculated, and m indicating the set having the smallest variance value is adopted as the optimum symbol timing information. As initialization, the content of the variable m is cleared to zero, and the content of the minimum variance value var _min is set to a sufficiently large variance value v.
Let ar ₀ (steps S1 and S2). Here, considering the case where k takes a finite number of integer values from 1 to I,

[0031]

[Equation 5]

Is calculated (step S4), and the variance value var is calculated.
_m

[0033]

[Equation 6]

Is calculated every time the number of m changes (step S5). Next, the calculated variance value var _m is compared with the minimum variance value var _min (step S6), and var _m is va
If it is smaller than r _{min, the} variance value var _{m is set} to the minimum variance value v
ar _min , and the number of m at that time is moved to m _min (step S7). On the other hand, if var _m is equal to or larger than var _min , the process of step S7 is not executed. After step S6 or S7, the content of the variable m is incremented by 1 (step S8).

The processes of steps S4 to S8 are repeated until the content of the variable m matches the number of M, and at the time of the match, the flow is exited (step S).
9). After all, in the iterative flow, the variance value in each of the sets of symbol data sampled from m = 0 to M−1 is calculated, and the smallest value among them and the number of m indicating the variance value are each variable var.
so that the determined the final value of the _min and m _min. Then, these final values are passed to the threshold value comparison processing unit 266S and the optimum timing information update processing unit 267S in FIG.

The principle of such processing will be briefly described as follows. The data symbol to be transmitted (here, x (n) as a complex number of I (n) and Q (n)) is a digital signal and is sent at the symbol rate Ts as described above. At this time, so that symbol interference does not occur with each other, and a certain bandwidth Ω
To output an analog signal x (t) that fits within (−1 / Ts ≦ Ω ≦ 1 / Ts),

[0037]

[Equation 7]

It is necessary to satisfy the following relationship. In FIG. 1, the equation (7) is obtained by the processing of the LPF 14 of the transmission system 1.
You are getting an analog signal that meets the requirements. Actually, the LP
F is the raised cosine filter
er) is often adopted. If the analog signal x (t) is sampled at the true symbol timing (t _T : see FIG. 2E) in the demodulated data in the reception system 2 without any disturbance, the digital signal x ( n) should be obtained. When sampled at a false symbol timing, the relevant x is sampled.
A signal that is completely unrelated to (n) may be output.
When disturbance is applied, the influence of fading is exerted particularly in mobile communication, so that there is a possibility that the digital signal x (n) of the transmitted symbol itself cannot be obtained even if the sampling is performed at the true symbol timing. However, if the transmitted symbol sequence is

[0039]

[Equation 8]

If the energy consists of a constant symbol data string, the amplitude of the sample signal at the true symbol timing observes a random signal according to the Rayleigh distribution even when the above disturbance is applied. be able to. Further, this amplitude has a characteristic of having a low frequency component, and its variation is smaller than that of a signal sampled at a false symbol timing. Therefore, it is possible to estimate the timing very close to the true symbol timing by examining the variation of each sample signal in each set of M sample signals indicated by m.

Generally, in the QAM modulation / demodulation system, as shown in FIG. 2, a signal having a constant energy is inserted in the transmission signal as a preamble (synchronization symbol signal). Then, symbol timing reproduction is performed in the receiving system using this preamble. Therefore, the symbol timing reproduction circuit 26A can be used as a pre-process for frame timing estimation (reproduction). That is, by obtaining the optimum symbol timing and performing symbol synchronization, it is possible to shift to the processing of frame timing reproduction with the arrival of the synchronization signal relating to the frame being narrowed down to some extent, which also saves processing time.

However, the frame timing estimation process, which has been performed before, becomes unnecessary. As a simple processing method, for example, in the optimum timing detection processing 26 of FIG. 1, the frame timing estimation processing is performed separately from the symbol timing reproduction processing. The frame timing estimation process in this case is based on preparing a reference preamble signal having the same symbol pattern as the transmission preamble signal in advance, comparing the reference preamble signal with the received symbol, and monitoring the correlation between the two. I was there. Further, the MAM or WDM as described above is applied to the symbol timing reproduction process, and when the symbol timing is obtained by the process, the information is given to the frame timing estimation process. When the frame timing estimation process receives the information, it starts the frame timing detection, and when it detects that the degree of the correlation is sufficiently large, it detects the arrival (reception) of the preamble signal. Fading is a reason for such a two-step process. That is, the symbol timing reproduction process itself is easily affected by fading, and even if the proper reproduction of the symbol timing is destroyed due to the fading after once obtaining the frame timing, the frame timing can be maintained and compensated for.

On the other hand, in the present invention, the symbol timing is reproduced by using the preamble signal arranged at the predetermined time axis position in the frame as the transmission symbol of constant energy as described above. The amplitude fluctuation of the sample signal due to the true symbol timing of the preamble signal is focused on following the Rayleigh distribution even when disturbance due to fading is added, and the variation of the sample signal at each sampling point in the symbol is examined,
Sampling points corresponding to those with sufficiently small variations are used as symbol timings. Therefore, regardless of fading or the like, it is possible to detect the predetermined time axis position within the frame, that is, the frame timing, and to perform the symbol timing reproduction, so that it is possible to obtain the result without performing the frame timing estimation processing as a separate means. I'm out.

As described above, according to the present invention, the accuracy and capability of estimating the symbol timing are improved, and the use of the frame timing estimation as a pre-processing is not limited to the processing used for the frame timing estimation. Since the timing estimation processing itself can be eliminated, the optimum timing detection processing unit is DSP
Especially when it is configured in the (Digital Signal Processor), it is possible to significantly reduce the time and power required for the process up to the frame timing estimation.

FIG. 5 shows the outline of the dispersion characteristic with respect to the maximum Doppler frequency when the preamble signal as described above is sampled. The thick line shows the variance of the sample values in each symbol at the true symbol timing t _T shown in FIG. 2 (e). The thin line shows the variance of the sample value in each symbol at the timing temporally distant from the true symbol timing t _T, with the width of the spacing varied, and the larger the width of the spacing, the normalized variance value It can be seen that has a tendency to grow. Further, according to this, the dispersion characteristic at the true symbol timing is about 0.5 or less in the standardized dispersion value regardless of the moving speed of the mobile station (the maximum Doppler frequency depends on the moving speed). While remaining, the dispersion characteristics at other timings are:
Since it has a variance value of more than 0.5, the variance threshold var _th to be set in the processing unit 266S of FIG.
It can be seen that the normalized value is about 0.6.

On the other hand, when a totally random signal whose amplitude is not uniquely determined, such as information signal symbol data, is sampled, the outline of the dispersion characteristic with respect to the maximum Doppler frequency is similarly shown, as shown in FIG. become. According to this, the dispersion characteristic at the true symbol timing cannot be distinguished from the dispersion characteristic at other timings. That is, at any point of the maximum Doppler frequency, there are some that are below the variance value at the true symbol timing, some are above the variance value, and some are close to each other, so it is difficult to determine the criteria for determining the true symbol timing. .

FIG. 7 is a schematic block diagram showing the structure of a communication system of another embodiment according to the present invention, in which the same parts as those in FIG. 1 are designated by the same reference numerals. The main difference of this embodiment from the previous embodiments is the configuration of the timing recovery processing by the detection output of the optimum timing detection unit 26. Specifically, the sampling frequency of the A / D converter 25 is fixed to f _M , and each of the output sample data is temporarily stored in the burst memory 29 as a storage unit for each symbol, and the pointer m obtained in the optimum timing detection unit 26 is obtained.
One of the sample data indicated by _op is transferred to the fading correction processing circuit 27 as symbol data.

The burst memory 29 is a so-called buffer.
It is also called a memory device and stores, for example, M sample data.
It has a memory area. Read operation of burst memory 29
Are controlled by the read control circuit 291. Read control circuit 29
1 is the M number of subsamples generated by oversampling.
Only the most suitable sample data of the sample data
Symbol timing recovery circuit for transfer as digital data
Pointer m from 26A _opRead address corresponding to
It is specified in the host memory 29. This allows the burst message
From the memory 29, the demodulation sequence from which distortion on the time axis has been removed.
The symbol data is transferred to the fading correction processing circuit 27.
Is done.

Also in this embodiment, the same effect as the previous embodiment can be exhibited. The sampling clock of the A / D converter 25 may be configured not to be directly supplied from the clock signal generation source 251, but to be supplied via a frequency divider or a multiplier. Also, A
Even if the sampling frequency of the / D converter 25 is not always fixed at f _M , only f during the optimum timing detection processing period
_It may be configured to set to _M and switch to the normal symbol rate fs in other cases.

Thus, as is apparent from the above-described two embodiments to which the present invention is directed, a method using a criterion close to an absolute criterion, which is deviated from the relative criterion, is proposed as a criterion for symbol timing reproduction. The present invention is not limited to the ping-pong transmission as shown in FIG. It is not limited to the values and numbers used in the figure. Further, as long as it is a signal composed of symbol data having a constant amplitude, it is basically possible to use this signal instead of the preamble signal to implement the present invention.
In addition to this point, in the embodiment, the preamble signal and the information signal are always stored in the time slot for transmission.
A method is conceivable in which, when the TDMA system is started up, first, all the symbol signals in the preamble signal form are stored in a time slot and transmitted, and after obtaining the optimum symbol timing, the slot signal in the form as in the embodiment is transmitted. . Further, the processes in FIGS. 3 and 4 are not limited to the process of obtaining the variance minimum value after the process of obtaining the variance minimum value, and the order of these processes may be reversed to the pointer m _op. The number of may be derived. In this embodiment, the number of one type of pointer m _op may be derived for each time slot. It is further noted that the present invention only needs to adopt linear modulation and is not limited to application to a communication system using QAM. The present invention can be appropriately modified by those skilled in the art as necessary without departing from the principle based on all the descriptions in the present specification.

[0051]

As described in detail above, according to the linear modulation type communication system having the symbol timing recovery function using the dispersion calculation of the present invention, the demodulated symbol signal in the receiving system is M (M > 1) times, the variance of the sample values in each of the M set of sample value sets consisting of the sample values for the predetermined symbols is calculated for each of the M sampling points per symbol in the oversampling, and the variance is calculated from the predetermined threshold value. Since one of the sampling points corresponding to the sample value set showing a small variance value is derived as the optimum symbol timing, it is possible to perform the symbol timing reproduction according to the criterion close to the absolute reference, and the reproduction accuracy thereof. Not only can you improve the Oite can also be satisfactorily form a symbol timing recovery.

[Brief description of drawings]

FIG. 1 is a schematic block diagram showing the configuration of a QAM type TDMA communication system according to an embodiment of the present invention.

FIG. 2 is a diagram showing an example of a transmission / reception mode in the communication system of FIG.

3 is a diagram showing a specific configuration mode of a symbol timing reproduction circuit 26A in the optimum timing detection unit 26 of FIG.

FIG. 4 is a flowchart showing details of a dispersion calculation processing unit 264S and a dispersion minimum value detection unit 265S in FIG.

FIG. 5 is a diagram showing an outline of dispersion characteristics with respect to the maximum Doppler frequency when a preamble signal is sampled.

FIG. 6 is a diagram showing a summary of dispersion characteristics with respect to the maximum Doppler frequency when a completely random signal in amplitude is sampled.

FIG. 7 is a schematic block diagram showing the configuration of a communication system of another embodiment according to the present invention.

[Explanation of symbols]

 1 Transmission System 11 Mapping Circuit 12 Frame Signal Forming Circuit 12 13 D / A Converter 14 LPF 15 QAM Modulator 16 Local Oscillator 17 Transmission Antenna 2, 2'Reception System 21 Reception Antenna 22 QAM Demodulator 23 Local Oscillator 24 LPF 25 A / D converter 251 Clock signal generation source 252 Delay / frequency setting circuit 26 Optimum timing detection unit 26A Symbol timing reproduction circuit 261,262 Square calculator 263 Adder 264S Distributed calculation processing unit 265S Dispersion minimum value detection unit 266S Threshold comparison processing unit 267S Optimal Timing information update processing unit 27 Fading correction processing circuit 28 Demapping circuit 29 Burst memory 291 Read control circuit

Continuation of the front page (51) Int.Cl. ⁶ Identification number Office reference number FI technical display location 9297-5K H04L 27/00 F

Claims (9)

[Claims] 1. A slot symbol signal formed by generating a synchronization symbol signal composed of symbols having substantially constant energy and an information symbol signal carrying a predetermined information signal at a predetermined symbol rate, and linking these symbol signals with each other. A transmission system that linearly modulates a carrier wave to generate a transmission signal, a demodulation symbol signal by demodulating the transmission signal, and the demodulation symbol signal at a sampling timing based on the synchronization symbol signal included in the demodulation symbol signal. A communication system comprising a receiving system for sampling, the transmitting system sampling the demodulated symbol signal at M (M> 1) times the predetermined symbol rate, and a symbol in the oversampling unit. Each of M sampling points per Is calculated, the variance of the sample values in each of the M sets of sample values consisting of sample values for a predetermined symbol is calculated, and one of the sampling points corresponding to the sample value set showing a variance value smaller than a predetermined threshold is optimized. A communication system comprising: a timing detection unit that derives as a symbol timing. 2. The timing detecting means calculates a variance value computing means for computing a variance value of the sample values in the set for each of the M sets of sample value sets, and a minimum value of the variance values. Selecting means for selecting the resulting set, and selecting one of the M sampling points corresponding to this set when the value of the variance of the set selected by the selecting means is less than the predetermined threshold value as the optimum symbol timing. The communication system according to claim 1, further comprising: deriving means for deriving as. 3. The timing detection means calculates, for each M sets of sample value sets, a variance value calculation means for calculating a variance value of the sample values in the set, and the variance value is less than the predetermined threshold value. Selecting means for selecting a set corresponding to the value, and one of the M sampling points corresponding to the set having the smallest variance value among the sets selected by the selecting means as the optimum symbol timing. The communication system according to claim 1, further comprising deriving means for deriving. 4. The communication system according to claim 1, further comprising a symbol rate sampling means for sampling the information symbol signal included in the demodulated symbol signal at the optimum symbol timing. 5. The oversampling means and the symbol rate sampling means include an A / D converter,
For the synchronization symbol signal included in the demodulated symbol signal, the sampling operation is activated at a rate of M (M> 1) times the predetermined symbol rate, while the optimum for the information symbol signal included in the demodulated symbol signal. 5. The communication system according to claim 4, further comprising sampling clock supply means for supplying a sampling clock signal for activating a sampling operation at the symbol timing to the A / D converter. 6. An accumulating means for accumulating at least M sample values by the oversampling means, and one of the sample values accumulated in the accumulating means, which corresponds to the sampling point derived as the optimum symbol timing, is read out. 4. The communication system according to claim 1, further comprising read control means. 7. The predetermined threshold value is set to a value equal to or higher than a maximum value of a variance of sample values at the true symbol timing of the demodulated symbol signal within a predetermined range of the maximum Doppler frequency. The communication system according to claim 1, 2, 3, 4, 5 or 6. 8. The communication system according to claim 1, wherein the synchronization symbol signal is a preamble signal arranged at the head of the slot symbol signal. . 9. A TDMA system using the communication system according to any one of claims 1, 2, 3, 4, 5, 6, 7 or 8, wherein the slot symbol signal has one predetermined channel in a frame. A TDMA system characterized by being transmitted during a communication time zone.