JPH0563633A - Network control system - Google Patents

Abstract

PURPOSE:To easily match controlled objects by controlling a control object of its own station at a succeeding time so that the relative error between the control object subject to filtering and its own control object is eliminated. CONSTITUTION:A communication means is provided between adjacent nodes, each node sends its own control object, and a node receiving the control object from an adjacent station applies being spatial filtering Fs or time filtering Ft to the control object of the adjacent station and controls the control object of its own station at a succeeding time so that the relative error between the control object subjected to filtering and its own control object is eliminated. Thus, the information relating to the control object is sent and received between its own node and a node in the vicinity of its own node to find out a relative error, then the relative error with respect to the control object at least in the vicinity of its own node is eliminated without provision of a special control station.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a network control system, and more particularly to a network control system in which a mobile station in each of a plurality of cells performs mobile communication with a base station in which it is located.

In a mobile communication network in which a service area is divided into a plurality of wireless cells (zones), and a wireless base station (hereinafter sometimes referred to as a node) is installed at the center of each wireless cell, the mobile stations are wireless. TDMA system with base station,
A network control method for performing communication by the FDMA method or the CDMA method is known. In these network control methods, a mobile station communicates while synchronizing with a signal from a radio base station of a radio cell in which the mobile station is located.

Here, since the radio base stations have separate reference clocks and perform transmission based on the reference clocks, when a mobile station moves between cells, the mobile station receives the reference clock from the destination radio base station. It is necessary to resynchronize the transmission signal such as frequency and timing, and there is a problem that communication is interrupted until the synchronization is reestablished. Therefore, it is necessary to eliminate the relative error of the controlled object value such as the phase difference or the frequency difference of the clocks between the adjacent wireless base stations by some means.

[0004]

2. Description of the Related Art In the conventional network control system, as shown in FIG. 55, a plurality of radio base stations 700 are used as nodes, and a single radio line central control station 800 monitors them to allow the plurality of radio base stations 700 to operate. A centralized control network for matching and correcting the frequency or timing of each transmission signal of each wireless base station 700 based on information from each is configured. This eliminates the need for the mobile station to resynchronize when the mobile station moves from one zone to the adjacent zone.

[0005]

In the above conventional method, the size of each cell is selected in consideration of the frequency utilization efficiency, the transmission power of the mobile station, etc., and the size of each cell is large to some extent. There was (for example, radius Km).

However, in response to the recent demand for smaller and lighter mobile stations, there is a demand for lower transmission power and higher frequency utilization efficiency, and the size of each cell is, for example, radius 5
In a network that is minimized to around 0m-100m,
In the above-mentioned conventional method, the number of wireless base stations dramatically increases,
There is a problem in that the load on the wireless line central control station and its host device becomes extremely large and control becomes difficult.

The present invention has been made in view of the above points,
In a network control method in which a mobile station performs mobile communication with a base station in which it exists in each of a plurality of cells, the transmission signal of each radio base station without configuring a centralized control network by a central control station. The purpose is to match the control target values of.

[0008]

[Means and Actions for Solving the Problems] The present invention shown in FIG .
Basically means and the action of light: FIG. 1 is a diagram for explaining the network control system according to the present invention as a method, now, the control target of the node of interest as A (t) at time t, It is controlled at the next time t + 1 according to A (t + 1) = A (t) + C (t) (1). Here, C (t) is a value when time filtering Ft is applied at time t, and C (t) = Ft {B (t), B (t-1), ..., B (tm)} (2). Here, B (t) is a value when spatial filtering Fs is applied at time t, and B (t) = Fs {E1 (t), E2 (t), ..., En (t )} (3) Here, Ei (t), i = 1,2, ..., n is a relative error between the node of interest at time t and the control target value Ai (t) of the node referred to by the node. , Ei (t) = A (t) −Ai (t), i = 1,2, ..., n (4).

In this way, if the information about the controlled object value can be exchanged with the nodes in the vicinity of the own node and the relative error can be found, at least the vicinity of the own node is not necessary without providing a special control station at the center. In, it is possible to eliminate the relative error related to the controlled object value, and finally,
It is possible to obtain a correctly controlled state regarding the controlled object value in the entire system.

Further, since the error signal is spatially filtered, even if there is a disturbance or anomaly in some of the plurality of spatially spread reference signals, the influence can be removed or mitigated by the spatial filtering. Even if there is a disturbance or anomaly in some of the plurality of temporally consecutive error signal sequences, the effect can be removed or mitigated by temporal filtering.

However, the order of the spatial filtering Fs and the temporal filtering Ft may be reversed, and at least one of them may be applied.

Embodiments of the Invention Shown in FIGS. 2-6: Preferred embodiments of the apparatus used directly to carry out the basic method of FIG. 1 above include first those shown in FIGS. ..

In the present invention shown in FIG. 2A, a network control system in which a plurality of radio base stations each transmitting a transmission signal are used as nodes, and a mobile station communicates with a radio base station in its own zone. In, each of the plurality of nodes includes a first calculation unit 11, a first error detection unit 12, and a control unit 13,
The first arithmetic unit 11 receives reference signals from peripheral nodes (see FIG.
(Corresponding to the control target value of 1) is subjected to a spatial filter calculation. Further, the first error detection unit 12 outputs an error signal between control target values indicating the difference between the output signal of the first calculation unit 11 and its own output signal. Then, the control unit 13 controls the synchronization target input signal so as to minimize the error signal, and outputs it as a reference signal indicating its own information to peripheral nodes.

Further, in the present invention shown in FIG. 2B, a plurality of second error signals, which output the difference between the reference signal and its own output signal, are output instead of the first error detecting section 12 described above. the error detection unit 12 ₁ to 12 _k is provided, the second arithmetic unit 14 for performing a spatial filtering operation on the output signals of the second error detection unit 12 ₁ to 12 _k instead of the first calculation portion 11 Provided,
The output signal of the second calculation unit 14 is given to the control unit 13 as a control signal.

In the case of the present invention shown in FIG. 3A, the first
Is provided for storing reference signals from peripheral nodes at a plurality of time points. Then, the third calculation unit 16 performs a time filter calculation on the reference signals from the first storage unit 15 at a plurality of time points, and outputs the calculation result to the first error detection unit 12. The other points are the same as those in FIG.

In the present invention shown in FIG. 3B, the error signal between the self output signal and the reference signal is generated by the third error detecting section 17, and the error signal is stored by the second storing section 18. The data is stored at each of a plurality of time points, and at least the error signals from the second storage section 18 at a plurality of time points are subjected to time filter calculation by the fourth calculation section 19, and the calculation result is used as a control signal in the control section 13. Giving to. The other points are the same as in FIG.

In the present invention shown in FIG. 4 (A), FIG.
With respect to the configuration of FIG.
In addition, a fifth arithmetic unit 22 for performing a time filter arithmetic operation on the output signals from the storage unit 21 at a plurality of time points and outputting the arithmetic operation result to the first error detection unit 12 is added.

Further, in the present invention is shown in FIG. 4 (B), to the arrangement of FIG. 2 (A), a first storage unit 15 and the third arithmetic unit 16 15 ₁ to 15 _k, 16 ₁ ˜16 _k , a plurality of them are provided, respectively, and the third computing units 16 _{1 to} 16 _k are provided.
It is characterized in that the reference signals which have been respectively subjected to the time filter calculation and are subjected to the spatial filter calculation in the sixth calculation unit 24.

In the present invention shown in FIG. 5, a fourth storage section 26 and a seventh arithmetic section 27 are added to the configuration of FIG. 2 (B). Thereby, the error signal subjected to the spatial filter calculation from the second calculation unit 14 is subjected to the time filter calculation in the fourth storage unit 26 and the seventh calculation unit 27, and is given to the control unit 13 as a control signal.

In the present invention shown in FIG. 6, FIG. 1 (A) and FIG.
The configuration of (B) is combined.

Aspects of the Invention Shown in FIGS. 7-9: The apparatus used directly to practice the basic method of FIG. 1 above further includes the preferred aspects shown in FIGS. 7-9.

That is, in order for each node to synchronize its own signal timing or signal parameter corresponding to the controlled object value shown in FIG. It is common to update at a fixed time interval, but generally, if the time interval for updating its own signal is shortened, the convergence time of the system will be shortened and it will be susceptible to disturbances etc., and the system will become unstable. Cause problems. Further, in the case of applying the time filter, if the time width of the time filter is lengthened, the effect of suppressing the influence of disturbance is high, but there is a problem that the convergence time of the system becomes long. Further, in the case of performing quantization or the like, if the quantization precision is increased, the residual error is reduced, and the synchronization control can be performed with high precision, but the convergence time of the system becomes long.

Therefore, first, in a preferred mode of the network control system shown in FIG. 7, the detection section 2-1 outputs its own output signal Q (corresponding to the control target value shown in FIG. 1) and a signal from another node. The difference signal E from R (also corresponding to the control target value) is detected, and the signal generator 2-2 is based on the difference signal E by the bias (timing signal T) from the controller 2-3. The own signal Q is controlled so that the difference signal E becomes smaller. In this state, the control unit 2-3 changes the time interval for energizing the signal generation unit 2-2 in the case where the system is in the initial state or according to the magnitude of the difference signal E, for example, by a power-on reset signal or the like. When the system is in the initial state or the difference signal E
Is larger than a predetermined value T _H , the energizing time interval is shortened to accelerate the convergence to the signal R. In other cases, the energizing time interval is lengthened to reduce the influence of disturbance or the like. ing.

In the preferred embodiment of FIG. 8, the detecting section 2-1 detects the difference signal E between its own signal Q and the signal R from another node, and the signal generating section 2-2 outputs the difference signal E. Based on this, the self signal Q is controlled so that the difference signal E becomes smaller. At that time, the filter 2-4 filters the signal R from the other node, the difference signal E from the detector 2-1 or the control signal C of the signal generator 2-2 based on the difference signal E ( The control unit 2-5 is arranged so as to perform spatial filtering or temporal filtering processing.
Changes the filter characteristic of the filter 2-4 when the system is in the initial state or in accordance with the magnitude of the difference signal E. For example, when the system is in the initial state or the difference signal E is a predetermined value T
When it is larger than _H , the response of the filter 2-4 is accelerated to speed up the convergence to the signal R, and in other cases, the response is delayed so that it is less affected by disturbance or the like.

In the preferred embodiment of FIG. 9, the detecting section 2-1 detects the difference signal E between its own signal Q and the signal R from another node, and the signal generating section 2-2 outputs the difference signal E. Based on this, the self signal Q is controlled so that the difference signal E becomes smaller. At that time, the input / output level converter 2-6 converts the input / output level with respect to the signal R from the other node, the difference signal E, or the control signal C of the signal generator 2-2 based on the difference signal E. It is arranged so as to perform, and in this state, the control unit 2-7
Changes the level conversion characteristic of the input / output level conversion unit 2-6 when the system is in the initial state or according to the magnitude of the difference signal E. For example, when the system is in the initial state or the difference signal E
Is larger than a predetermined value T _H , the level conversion characteristic is coarsely quantized or linearized to speed up the convergence to the signal R. In other cases, the quantization accuracy near the input level of 0 is increased. The accuracy of convergence to the signal R is improved.

Aspects of the invention shown in FIG . 10: FIG. 1 above.
As a device directly used for carrying out the basic method of (1), there is a preferable mode in which the frequency shown in FIG.

First, in the base station as an adjacent node, the frequencies f _A and f _B (corresponding to the control target value in FIG. 1) of the respective radio channels are different by a nominal frequency interval f _{S in} order to suppress the interference to be small. When the mobile station moves between zones and moves toward the adjacent base station, the mobile station switches its frequency from f _A to f _B for communication.

By the way, since the oscillation frequency of the local oscillator that determines the frequency of the radio channel has a finite stability, the frequency assigned to each base station always includes a finite error. In this case, this frequency error was absorbed by an automatic frequency control (hereinafter abbreviated as AFC) circuit mounted on the mobile station.

As a result, when the mobile station moves from one zone to another (hereinafter, the movement of the mobile unit from one zone to another or from one cell to another) is referred to as a handover, the frequency f _B of the radio channel of the base station in the new zone is newly added. Even if the information is given to the mobile station, the mobile station must reabsorb the error frequency (f _B −f _B ') between the nominal frequency f _B of the base station and the actual frequency f _B ' by the AFC circuit. I have to. In reality, since the signal of the base station before the handover also includes the frequency error, the maximum frequency error that must be acquired is twice the frequency error allowed by the base station.

As described above, the control of the AFC circuit works transiently at every handover, and the call or data transmission / reception is interrupted until the control is completed, and the user feels inconvenience. From the viewpoint of effective use of frequency, it has been pointed out that the zone (in this case, a cell) is used spatially and repeatedly by making the zone small, but in this case, frequent handover is performed. Therefore, the interruption of communication will further deteriorate the serviceability of the system.

Therefore, in the present invention, a relative frequency error between adjacent base stations which can be handed over is eliminated, so that the mobile station sets the frequency by the nominal frequency interval f _S of the radio channel at the time of handing over. The focus is on making enough changes.

For this purpose, each base station detects the frequency corresponding to the error between the control target values of the frequency of the signals of the surrounding base stations and the frequency of its own station, and eliminates the error frequency of the transmission frequency of its own station. Control it as follows.

Therefore, in the network control system of this preferred embodiment, as shown in FIG. 10, first, in the reception automatic frequency control unit 3-3 in the base station 3-1, AFC control matching the transmission frequency of the adjacent base station is performed. Then, the frequency error detector 4 compares the frequency f _{R at} this time with the frequency f _A + f _{S obtained} by adding the specified interval frequency f _S between the reference stations to the transmission frequency f _A of the own station and detects the frequency error Δf. To do.

Such frequency error Δf is detected with a plurality of adjacent base stations, and the averaging unit 3-5 averages this to perform the spatial filtering shown in FIG. ..

That is, suppose that the nominal frequency of the i-th base station is f _i and the actual frequency at time t is f _i '(t).
Neighboring base station j from which handover can occur from this base station
The relative frequency error Δf _i (t) over (j = 1, 2, ..., N) is Δf _i (t) = [ _{j = 1} Σ ^N {(f _i '(t) −f _i ). -(F _j '(t) -f _j }] / N = [ _{j = 1} Σ ^N {(f _i ' (t) + (f _j -f _i ) -f _j '(t)}] / N = [ _{J = 1} Σ ^N {(f _i '(t) + f _sij ) −f _j ' (t)}] / N (5), where f _sij is f _sij = f _j −f _i ( 6) is the difference between the nominal frequencies of base stations i and j, and is an integral multiple of the nominal interval frequency fs, and the addition of _{j = 1} Σ ^N is j
= 1, 2, ..., N is meant to be added.

This error information Δf _i (t) is used in the averaging unit 3-5.
Is given to the transmission frequency control unit 3-6 from the transmission frequency control unit 3-6, and the transmission frequency (f _A ) of the own station at the next time t + 1 is represented by f _i (t + 1) = f _i (t) It is controlled as −α · Δf _i (t) (7). However, α is a numerical coefficient, and 0 <α ≤ 1.

In the averaging of the frequency error relative to each adjacent base station, a weighted average with the C / N of each signal as a weight is taken, or by skipping a large error by averaging. The stability of the control system can be increased.

In the present invention, the averaging unit 3-5 can also add temporal averaging to the frequency error Δf _i (t) (temporal filtering). In this case, in the communication channel. It is possible to eliminate the adverse effects of noise, etc., and to have a smoothing action against a large detection error in an instant, thereby improving the stability of the entire system. This is especially necessary when the C / N of the received signal is low.

When a certain base station is considered in this way,
Relative frequency differences across the surrounding base stations can be eliminated, but this is only local. This is because when the absolute numerical value regarding the frequency needs to be limited, the relative frequency control between the base stations described above is insufficient, and the absolute value of the frequency is not guaranteed at all.

Therefore, in order to guarantee the elimination of the global frequency error over the entire communication system, in the present invention, FIG.
1, one of the base stations 3-1 is provided with a reference station R that transmits an absolute transmission frequency,
The absolute frequency of the entire communication system is guaranteed by the control of absorbing the frequency error Δf by each of the other base stations by being dragged by this absolute transmission frequency, and the reliability is assured. Will be improved.

The present invention is a communication system in which the transmission and reception frequencies are the same, and the transmission and reception frequencies are different, and each base station and mobile station constantly control the frequency difference between the reception frequency and the transmission frequency. It is possible to adapt to each.

Aspects of the invention shown in FIG . 12: FIG. 1 above.
As a device used directly for carrying out the basic method of (1), there is a preferable mode in which the timing shown in FIG.

In the network control system according to this preferred embodiment,
Each of the wireless base stations 101 as a plurality of nodes notifies the surrounding wireless base stations 101 of the transmission timing information of its own station, and the above transmission timing information as an input signal, and at least the transmission signal of its own station. And a correcting means 104 for correcting the phase of the signal so as to reduce the difference from the timing at which the transmission signal of the surrounding wireless base station 101 is subjected to the spatial filtering process.

That is, in the plurality of wireless base stations 101, the notifying means 103 notify each other of the own station transmission timing information with the neighboring wireless base stations, and each correcting means 104 uses the transmission timing information to notify the neighboring wireless base stations. Correction is performed so that the difference in transmission timing with the station 101 becomes small, and the phase difference finally becomes zero. Therefore, as shown schematically in FIG. You are configuring a network.

The correction operation will be described in more detail. The size of a node (radio base station 101) in the distributed control type network (that is, the frame phase and the clock frequency in the case of the TDMA method is indicated, and FDMA is used). In the system and the CDMA system, the clock phase and the clock frequency are shown) as B (i, t). However, i indicates the node number and t indicates the time. Each node compares the size with the surrounding nodes and corrects only the difference.

That is, the node number to be compared is 1 to
Assuming N (however, i> N for convenience), the difference in size between the local node and the surrounding nodes subjected to the spatial filtering process is

ΔB (i, t−1) = B (i, t−1) ^−N Σn ₌ 1B (n, t−1) / N (8) Therefore, the size B of the node with the node number i
(I, t) is corrected as shown in the following equation. B (i, t) = B (i, t−1) − [α · ΔB (i, t−1)] _AVE (9) where α is a coefficient (0 <α ≦ 1), and [] _AVE means averaging.

As described above, in the present invention, all the nodes finally have the same size in any of the TDMA system, the FDMA system, and the CDMA system, and the respective radio base stations are brought into a synchronized state.

In the present invention, each node (radio base station 1
01) is connected to multiple adjacent nodes,
Even if the line with one of the nodes is disconnected, the line can be finally prevented from being affected by the timing information from the remaining adjacent normal nodes. Further, even if the number of nodes (radio base station 101) increases, the number of nodes adjacent to each node does not increase so much, and thus the burden (load) on each node remains almost unchanged.

In addition, according to the present invention, each node has a pull-in range, and by controlling only the control target value existing in this pull-in range, some nodes become uncontrollable in the network. In such a case, the controlled object value can be removed from the mutual synchronization control, and the unstable period of the entire network can be reduced.

Further, according to the present invention, by giving a certain regularity to the selection of the base station that outputs the controlled object value, it is possible to prevent the variation of the controlled object value from spreading randomly to neighboring nodes. The overall instability period can be reduced.

Further, in the present invention, the above-mentioned spatial filtering gives a weighted average to the relative error of the controlled object value based on the received electric field strength selected by the diversity receiving system, so that a special S / N detection is performed. Eliminates the need for circuits.

[0053]

FIG. 14 shows the configuration of an embodiment of the network control system according to the present invention shown in FIG. 2A, and in FIG.
The same components as in (A) are designated by the same reference numerals, and the description thereof will be omitted. In FIG. 14, 11a is a read only
In a memory (ROM), 4-bit reference signals (for example, frequency signals, timing signals, and received power signals as control target values) from four nodes around this node
Is inputted to the address line, and data is preliminarily tabulated and stored so that an average value of 4 bits × 4 address input is outputted from 4 bits of the 8-bit data line.

Reference numeral 12a is a comparator, which is composed of, for example, a 4-bit digital subtractor, and constitutes the error detecting section 12. Comparator 12 which is a 4-bit digital subtractor
a outputs a result obtained by subtracting the other 4-bit input value from the one 4-bit input value in 4 bits.

13a is a coefficient multiplier, 13b is an adder, 1
Reference numeral 3c is a control device, which constitutes the control unit 13. The coefficient multiplier 13a is composed of, for example, a 4-bit digital multiplier, and outputs a 4-bit result obtained by multiplying one 4-bit input value by another 4-bit coefficient α (where 0 <α <1). To do.

The adder 13b is composed of, for example, a 4-bit digital adder (adder), and as one 4-bit input, a signal to be synchronized at its own node (for example, a clock as a control target value, transmission power information). , Etc., and the output value of the coefficient multiplier 13a is input as the other 4-bit input, and these are added to output a 4-bit addition result. This adder 13
Since b is provided to give an initial value or offset, it can be omitted if they are not present.

The control device 13c is composed of, for example, a 4-bit digital accumulator (accumulator: ACC), and the 4-bit result is obtained by adding one 4-bit input value and the stored output value before the temporary point. Output. The output value of the control device 13c is input to the comparator 12a, is output as its own synchronization target signal, and is further transmitted as a reference signal to the peripheral nodes by the output means.

Next, the operation of this embodiment will be described.
The reference signals from the four peripheral nodes are input to the ROM 11a, converted into the average value of the four reference signals here, and then input to the comparator 12a, where they are subtracted from the output value of the control device 13c. Here, if the own node number is i, the node numbers of other nodes are j, and the signal amounts of the node numbers i and j at time n are Si (n) and Sj (n), the output error signal of the comparator 12a Is expressed by the following equation.

ΔSi (n) = { _{j = 1} Σ ^N Sj (n)} / N-Si (n) (10)

However, in the above equation, N is the total number of input reference signals, which is "4" in this embodiment. Further, in the above equation, the first term on the right side represents the average output value of the ROM 11a.

Output error signal ΔSi (n) of the comparator 12a
Is supplied to the coefficient multiplier 13a, where it is multiplied by the coefficient α, then added with the synchronization target input signal A by the adder 13b, and further input to the control device 13c and added with the signal before the temporary point. As a result, the signal Si (n + 1) represented by the following equation is taken out from the control device (ACC) 13c.

Si (n + 1) = Si (n) + α · ΔSi (n) + A (11) In a steady state, the value of its own signal and the average of the values of the reference signals match, so ΔSi (n) = 0, and its own signal does not change. On the other hand, when the value of the self signal changes for some reason or the value of the reference signal changes, control is performed so that ΔSi (n) is set to 0, and the value Si (n + 1) of the self signal becomes Change. As a result, each node operates so as to be synchronized with the surrounding nodes, and each node in the network as a whole is synchronized.

In such a network, when an abnormally large value or an abnormally small value is input to one of the four reference signals, the influence thereof is reduced to 1/4 by the averaging operation by the ROM 11a, that is, the spatial filter operation. Therefore, it is possible to suppress the instability of the signal amount to be synchronized, which is the control target.

Here, the spatial filter calculation by the ROM 11a is not limited to the above-mentioned average value calculation, and an average value excluding a value above (or below) a certain threshold may be used.

15 and 16 show simulation results of the first embodiment of FIG. 14, respectively. FIG. 15 shows characteristics of signal amount Si (n) vs. time (simulation result) when a simple average as in the embodiment is used as the spatial filter calculation,
FIG. 16 shows characteristics of signal amount versus time when an average of values excluding a threshold value or more is used as the spatial filter calculation,
The changes in the signal amounts of all the nodes (36 nodes) are plotted in the same graph.

In FIG. 15, it can be seen that when one reference signal changes to an abnormally large value at a certain time point as shown by a, the signal amount of the synchronization target of each node gradually converges to a constant value.
On the other hand, in FIG. 16, even if one of the reference signals is fixed to an abnormally large value exceeding the threshold value as indicated by a, the reference signal is excluded, so that other nodes are not affected as indicated by b. I understand.

Other examples of the spatial filter calculation include a method of obtaining the most input value (median) and an average value calculation excluding the maximum and minimum values. In these cases, the influence of a value far from the others is removed. be able to.

In the above embodiment, the result of the spatial filter operation on the input reference signal is compared with its own signal.
As shown in FIG. 2B, spatial filter calculation is performed on a plurality of error signals obtained by comparing each input reference signal with its own signal amount, and the calculation result is used as a control input of the control unit 13. Good.

That is, in this case, the spatial filter calculation result ΔSi (n) of the following equation is obtained from the equation (10). ΔSi (n) = { _{j = 1} Σ ^N ΔSij (n)} / N (12) ΔSij (n) = Sj (n) −Si (n) (13)

FIG. 17 is a block diagram of an embodiment of the network control system according to the present invention shown in FIG.
The same components as those in (A) and FIG.
The description is omitted. In FIG. 17, reference numeral 15a is a 3-bit shift register having a 4-bit width, which constitutes the first storage unit 15. The 4-bit input reference signal is shifted rightward every clock, and the past 3 Store the reference signal in the clock period.

Reference numeral 16a denotes a ROM, which is the third arithmetic unit 16
The data is preliminarily tabulated and stored so that the average value of 4 bits × 4 address inputs is output from 4 bits of the 8-bit data line.

In this embodiment, the reference signal from one peripheral node is shifted to the right by every one clock by the shift register 15a and applied to the 4-bit address terminal of the ROM 16a. Other three 4 of ROM16a
Each of the bit address terminals has been output in parallel from the shift register 15a in the past three time points (one clock before,
Reference signals 2 clocks before and 3 clocks before) are supplied. The ROM 16a calculates the average value (moving average value) of the reference signals at the four consecutive time points from the present time to the past, and supplies it to the comparator 12a. That is, the ROM 16a
Performs a time filter calculation on the reference signal and supplies the calculation result to the comparator 12a, where the control device 13c
From the output signal Si (n) and converted into an error signal ΔSi (n) represented by the following equation.

ΔSi (n) = { _{m = 1} Σ ^M Sj (n-m + 1)} / M-Si (n) (14) Here, M represents a period (the number of time points) for which the moving average is taken (this In the example, M = 4).

The error signal ΔSi (n) at this time n
Is input as a control signal to the control unit 13 having the same configuration as that of the control unit 13 shown in FIG.
A +1 signal Si (n + 1) is output. This output signal is input to the comparator 12a, is used as a clock of its own node, and is transmitted as a reference signal to peripheral nodes. Incidentally, A is an input signal to be synchronized. Si (n + 1) = Si (n) + α · ΔSi (n) + A (15)

In the present embodiment, in the steady state, the value of its own signal and the value of the moving average of the reference signal match, so ΔSi
Since (n) = 0, its own signal does not change. On the other hand, when the value of the self signal changes for some reason or the value of the reference signal changes, control is performed so that ΔSi (n) is set to 0, and the value Si (n + 1) of the self signal becomes Change. As a result, each node operates so as to be synchronized with the surrounding nodes, and each node in the network is synchronized as a whole.

When the value of the reference signal changes abnormally large or small, the moving averaging operation suppresses the rate of change due to the effect to 1/4, and the signal amount of its own changes gently. ..

This will be further described with reference to FIGS. 18 and 19. In FIGS. 18 and 19, the number of nodes is 3 respectively.
6 shows the case where the number of reference signals is 3, FIG. 18 shows the signal amount vs. time characteristics when the time filter is not applied (in the embodiment of FIG. 9, the shift register 15a is not provided), and FIG. The signal quantity-time characteristic of a structure is shown. Further, FIGS. 18 and 19 plot the changes in the signal amounts of all the nodes in the same graph, as in FIGS. 15 and 16.

In FIG. 18, when the value of one reference signal changes significantly as shown in d, the signal amount of the other node changes to e.
As shown in (3), the change in the signal amount is not large, but after a sharp change, it gradually converges to a constant value. On the other hand, in the present embodiment, when the value of one reference signal changes abnormally greatly as shown by f in FIG. 19, the signal amount of the other node changes little as shown by g, and is gentle. You can see that it changes. Therefore, stable network synchronization control that is hardly affected by an abnormal change in the reference signal or disturbance can be performed.

In the above-mentioned embodiment, the simple moving average is used as the calculation (time filter). However, in addition to the simple moving average, a section average within a certain time and a value corresponding to the time from the present are weighted. The weighted average, maximum input value (median), average excluding maximum / minimum, etc. may be used. When using a weighted average with the value corresponding to the time from the present as the load, for example, if the reciprocal of the time from the present is used as the load, the past control value will not affect the current control value. , The convergence time from the initial state to the steady (stable) state is faster than when the simple moving average is used. When using the average excluding the median and maximum / minimum, the effect of sudden disturbance can be eliminated.

Further, in the above embodiment, the result of applying the operation (temporal filter) to the stored values of the input reference signal from the present time point to the past several time points is compared with its own signal.
As shown in (B), the result of applying a calculation (time filter) to a plurality of error signals obtained by comparing the stored signal value at each time point with its own signal amount is used as a control input. Good.

That is, in this case, the time filter calculation result of the following equation is obtained from the equation (14). ΔSi (n) = { _{m = 1} Σ ^M ΔSij (n−m + 1)} / M (16) ΔSij (n) = Sj (n) −Si (n) (17)

FIG. 20 shows an embodiment of the present invention shown in FIG. 4A. In the figure, the same components as those in FIGS. The description is omitted.
In FIG. 20, reference numeral 21a is a 3-bit shift register having a 4-bit width, which constitutes the third storage unit 21 and has the same configuration as the shift register 15a. Also, 2
Reference numeral 2a denotes a ROM, which constitutes the fifth arithmetic unit 22 and has the same configuration as the ROM 16a. The average value of 4 bits × 4 inputs input to the address line is 8 bits for the data line. Data is written in advance so that 4 bits are output.

Next, to explain the operation of this embodiment,
The average value of the four reference signals extracted from the ROM 11a is input to the shift register 21a, is shifted to the right every one clock, and is supplied to the 4-bit address terminal of the ROM 22a. To the other three 4-bit address terminals of the ROM 22, the reference signal average values at the past three points output from the shift register 21a in parallel are input. The ROM 22a calculates a moving average value of the reference signal average value at four consecutive time points from the present time to the past and supplies the moving average value to the comparator 12a. Ie ROM
The reference numeral 22a supplies a calculation result obtained by further time-filtering the spatially filtered reference signal to the comparator 12a.

The comparator 12a takes the difference between the calculation result from the ROM 22a and its own output signal to generate an error signal ΔSi (n) represented by the following equation. ΔSi (n) = [ _{m = 1} Σ ^M { _{j = 1} Σ ^N Sj (n-m + 1)} / N] / M-Si (n) (18)

This error signal ΔSi (n) is input as a control signal of the control unit 13, and is output from the control unit 13 as the signal Si (n + 1) shown in the equations (11) and (15).

In this embodiment as well, as in the above embodiments, when the value of the reference signal changes, control is performed so that the error signal ΔSi (n) becomes 0, and the nodes constituting the network are synchronized. Acts like. Furthermore, in the present embodiment, when an abnormally large value or an abnormally small value is input to the reference signal, the total change amount is 1/4 by the spatial filter calculation.
And the rate of change due to the influence is suppressed to 1/4 by the time filter calculation.

It is needless to say that the above-mentioned various operations other than the simple average can be used as the above spatial filter operation, and the above-mentioned various operations other than the simple moving average can be used as the above time filter operation. .. Also,
The order of spatial filter calculation and temporal filter calculation is the third above.
The present invention is not limited to the embodiment, and the spatial filter calculation may be performed after the temporal filter calculation is performed on the reference signal as shown in FIG.

Further, as shown in FIG. 5, the result of applying the spatial filter and the temporal filter to the plurality of error signals obtained by comparing the value of the reference signal with the signal amount of its own is the control input ΔS.
It may be used as i (n). In this case, ΔSi (n) is expressed by the following equation from the equation (18). ΔSi (n) = [ _{m = 1} Σ ^M { _{j = 1} Σ ^N ΔSij (n-m + 1)} / N] / M (19) where ΔSij (n) = Sj (n) -Si (n) (20) ).

Further, as shown in FIG. 6, the spatially filtered error signal may be temporally filtered.

Each of the above embodiments is applied to a radio base station of a mobile communication system, and clocks, frequencies, transmission powers, etc. can be synchronized in the whole mobile communication network. It is effective when applied to a mobile communication network with an extremely large number of base stations.

FIG. 21 shows the construction of the embodiment of the network control system according to the present invention shown in FIG.
10 is a phase detector (corresponding to 2-1 in FIG. 7), 2-20 is a signal generator (also corresponding to 2-2), 2-21 is a multiplier, 2-22 is an adder / subtractor, and 2-23 is 2-23. Register, 2-2
4 is an oscillator, 2-25 is a shift number variable shift register that can change the number of shift stages from input to output according to a control signal, 2-30 is a control unit (also equivalent to 2-3) 2-3
1 is a comparator, 2-32 is an OR gate circuit (O), 2-3
3 is a variable frequency divider.

The phase detection unit 2-10 detects the phase difference between the clock signal Q as its control target value and the clock signal R as the control target value from another node and detects the absolute value signal | E of the phase difference. | And the symbol S are output, and the signal generator 2-20 is self-energized according to the phase difference signal | E |, S at that time by the bias (timing signal T ₃ ) from the controller 2-30. The phase of the generated clock signal Q is controlled so that the phase difference signal | E | becomes small. More specifically, the multiplier 2-21 determines that the phase difference signal | E |
Multiply by (0-1) to keep the loop gain of the system below 1
The adder / subtractor 2-22 adds the new phase difference α | E |, S to the accumulated phase difference (control signal | C |, S) of the register 2-23 to update the control signals | C |, S. On the other hand, oscillator 2-
24 is a clock signal Q ′ having the same frequency as the clock signal R
The variable stage number shift register 2-25 receives the clock signal Q ', and advances or delays the phase of the output clock signal Q according to the control signals | C |, S.

In this state, the control section 2-30 changes the time interval for energizing the signal generating section 2-20 when the system is in the initial state or in accordance with the magnitude of the phase difference signal | E |
For example, when the system is set to the initial state or the comparator 2-31
The phase difference signal | E | if is determined to be greater than a predetermined value T _H is to shorten the time interval of generation of the timing signal T ₃ by reducing the frequency division ratio of the variable frequency divider 2-33, which Thereby speeding up the convergence to the signal R. Also, in other cases, to increase the frequency division ratio of the variable frequency divider 2-33 longer time interval of generation of the timing signal T ₃ and thereby are not easily affected by disturbance.

FIG. 22 is a block diagram of the phase detecting section in the above embodiment. In the figure, 2-10 is a phase detecting section (corresponding to 2-1 in FIGS. 7 to 9) and 2-11 is a D type. Flip-flop (FF), 2-12 is a counter, 2-
13 is an adder / subtractor, 2-14 is a latch, 2-15 is a set / reset type flip-flop (FF), 2-1
6 is an up / down counter (U / D counter), 2-
17 is a set / reset type flip-flop (F
F), D is a delay circuit, I is an inverter circuit, N is NAN
It is a D circuit.

FIG. 23 is an operation timing chart of the above-mentioned phase detecting section. The operation of the phase detecting section will be described below with reference to FIGS. 22 and 23. First, when the clock signal R is input from another node, each rising edge thereof. Then, a series of timing signals T ₁ and T ₂ are generated. Flip-flop 2-11
The output gate signal G is a signal which is reset by the timing signal T ₁ and is set at the last rise of the delay circuit D. While the gate signal G is at the HIGH level, the counter 2-12 has a high frequency. Clock signal CL
Count up K. The state is indicated by the count signal A.

On the other hand, the clock signals R and Q are input to the set and reset terminals of the flip-flop 2-15, respectively, and it is now assumed that three typical phase states of the clock signal Q with respect to the clock signal R are shown. become Q ₁ ~Q _3. Here, it is considered that the signal Q ₁ is in phase delay, the signal Q ₂ is in an intermediate state, and the signal Q ₃ is in advance of phase b. Therefore, the signal Q ₁ is pulled in the direction of arrow a, and the signal Q ₃ is pulled in the direction of arrow b. Such control is performed. It should be noted that the signal Q ₂ may be drawn into either one.

The output of the flip-flop 2-15 corresponding to the signals Q _{1 to} Q ₃ , that is, the up / down counter 2
The -16 up / down control signals U / D _{1 to} U / D ₃ are as shown. The up / down counter 2-16 is energized by the gate signal G, counts up the clock signal CLK while the up / down control signal is HIGH level, and counts up when the up / down control signal is LOW level. Count down the clock signal CLK. Therefore, when the count signal B corresponding to each of the up / down control signals U / D _{1 to} U / D ₃ of the up / down counter 2-16 is shown, the path is taken. In the case of and, there is a state in which the count B is changed from 0 to -1, and at that time the up / down counter 2-16
Outputs a borrow signal BO, which is flip-flop 2
It is stored in -17 and becomes the code S of the phase difference signal | E |.

The adder / subtractor 2-13 is (A
+ B) and the phase difference signal | a | = | A + B |
The HIGH level (delayed phase) is set in the latch 2-14. In the other case, (A-B) is performed and the difference signal | b |
= | AB | and the code S = LOW level (leading phase) are set in the latch 2-14.

FIG. 24 shows the configuration of the embodiment of the network control system according to the present invention shown in FIG.
40 is a time filter (corresponding to 2-4 in FIG. 7), 2-41
2 to 43 are shift registers (Z ^-1 ) for sequentially shifting the phase difference signals | E | and S, and 2-44 are shift registers 2-
42, 2-43 output or selector for selecting data "0", 2-45 ROM for outputting the arithmetic mean value corresponding to address inputs 0 to n, 2-50 control section, 2-51 D type Is a flip-flop (FF). 2
-40 'is an analog time filter.

The phase detector 2-10 detects the phase difference between its own clock signal Q and the clock signal R from another node, and outputs the absolute value signal | E | of the phase difference and the code S. The generator 2-20 controls the phase of the clock signal Q generated by itself based on the phase difference signal | E |, S so that the phase difference signal | E | becomes smaller.

At that time, the time filter 2-40 'or 2
-40 is a signal R from another node, a phase difference signal | E |, S
Alternatively, the control signal | C |, S of the signal generator 2-20 based on the phase difference signal is subjected to time filter (moving average) processing. In this state, the controller 2-50 sets the system to the initial state. Or the filter time width of the time filter 2-40 'or 2-40 is changed according to the magnitude of the phase difference signal | E |.

More specifically, for example, when the system is set to the initial state or when the phase difference signal |
When it is determined that E | is larger than the predetermined value T _H , the flip-flop 2-51 is set and its output signal F
The TW connects the input terminal of the selector 2-44 to the B side, whereby the data width for taking the moving average is reduced, the response of the time filter 2-40 is accelerated, and the convergence to the signal R is accelerated. In other cases, the flip-flop 2-51 is reset and its output signal FTW connects the input terminal of the selector 2-44 to the A side, which increases the data width for taking the moving average. The response of the time filter 2-40 is delayed, and is less likely to be affected by disturbance or the like.

FIG. 25 shows the configuration of an embodiment of the network control system according to the present invention shown in FIG.
60 is an input / output level conversion unit (corresponding to 2-6 in FIG. 9), 2
-61 is an RO that stores a plurality of types of level conversion characteristics
M and 2-70 are control units, and 2-71 is a D-type flip-flop (FF). 2-60 'is an analog type input / output level converter.

The phase detector 2-10 detects the phase difference between its own clock signal Q and the clock signal R from another node, and outputs the absolute value signal | E | of the phase difference and the code S. The generator 2-20 controls the phase of the clock signal Q generated by itself based on the phase difference signal | E |, S so that the phase difference signal | E | becomes smaller.

At this time, the input / output level converting unit 2-60 '
Alternatively, 2-60 is a signal R from another node, a phase difference signal | E
, S or the control signal | C |, S of the signal generator 2-20 based on the phase difference signal, the input / output level is converted, and in this state, the controller 2-70 initializes the system. The level conversion characteristic of the input / output level converter 2-60 'or 2-60 is changed depending on the state or the magnitude of the phase difference signal | E |.

More specifically, for example, when the system is in the initial state or when the phase difference signal |
When it is determined that E | is larger than the predetermined value T _H , the flip-flop 2-71 is set and its output signal L
C selects one of the level conversion characteristics of the ROM 2-61 (for example, a coarse quantization characteristic or a linear quantization characteristic), which widens the input / output dynamic quantization range or makes the quantization range linear, and Convergence to R becomes faster.
In other cases, the flip-flop 2-71 is reset so that its output signal LC has the other level conversion characteristic of the ROM 2-61 (for example, a non-linear quantization characteristic with improved quantization accuracy when the input level is near 0). Etc., and thereby, the accuracy of quantization in the vicinity of the input level of 0 is increased and the accuracy of convergence to the signal R is increased.

The present embodiment can be applied not only to the phase synchronization of the clock signal but also to the synchronization of various signal parameters, control parameters and the like.

FIG. 26 is a block diagram of the error detection unit in the above embodiment. In the figure, 2-80 is an error detection unit (corresponding to 2-1 in FIGS. 7 to 9), 2-81 is a reception unit, and 2 is a reception unit. -8
2 is a demodulation unit, 2-83 is an A / D conversion unit, and 2-84 is a subtractor.

For example, when an amplitude modulated signal having a predetermined signal parameter is sent as the signal R from another node, the receiving section 2-81 receives the modulated signal and the demodulating section 2-82 receives it. The amplitude of the signal parameter is demodulated and A / D converted by the A / D converter 2-83. Then, the subtractor 2-84 obtains the difference signal between the output of the A / D converter 2-83 and the amplitude of the signal parameter P of its own, and the signal parameter synchronization control can be performed in the same manner as above. ..

In this case, the signal generator 2-20 may control the amplitude of its own signal parameter P based on the error signals | E | and S so that the error signal | E | becomes smaller.

In the above embodiment, the number of stages variable shift register 2-25 shifts the phase of the clock signal Q ', but the control signal | C |, S directly controls the oscillation frequency of the oscillator 2-24. You may do it.

[0112] The difference signal in the comparator 2-31 in the above Examples | E | but is determined whether greater than a predetermined value T _H, further the difference signal provided threshold T _Hi plural stages | E | May be determined in multiple stages and the controlled object may be controlled in multiple stages accordingly.

In the above embodiment, the case where the signal R from another node is one is shown, but it is also possible to synchronize with a plurality of nodes at the same time, and in this case, for example, A signal obtained by performing averaging or the like on a plurality of signals R may be input to the detection unit, or a difference signal E _i between the signal R _i from each node and its own signal Q is obtained,
The difference signal E may be obtained by averaging these.

The loop gain α may be changed when the system is in the initial state or in accordance with the magnitude of the difference signal E.

In the above embodiment, the time filter 2-
Although 40 is used, other filter characteristics such as a high pass filter, a notch filter and a band pass filter may be prepared and these filter characteristics may be changed.

Example of averaging in frequency control (FIG. 2
7 to 35): An example of spatial averaging (filtering) in the averaging unit 3-5 shown in FIG. 10 is shown below. In this case, the following g _1i to g _6i are the second right side of the equation (7).
It is replaced with the term Δf _i (t) to control f _i (t + 1).

(1) The simplest averaging example is the arithmetic mean, which is equal to the equation (5), g _1i (t) = Δf _i (t) = [ _{j = 1} Σ ^N {(f _i '(t _{) + f sij) -f j '} (t) ] _{/ N = {j = 1 Σ} N Δf ij (t)} / N becomes (21). Here, Δf _ij (t) is, relative to the frequency f _i of the i th base station as described above '(t) and frequency f _j of the j-th station of the neighboring base stations' (t) The frequency error is shown. _{That, Δf ij (t) = (} f i - represented by _{'(t) f sij) -f} j' (t) (22).

(2) In the arithmetic mean of the above equation (21), Δ
Excluding the maximum and minimum values of f _ij (t) from equation (21), g _2i (t) = [ _{k = 1} Σ ^N-2 Δf _ik (t)] / (N-2) (23) May be averaged. However, in this case, N ≧
There is a need for 3. Further, k represents an N-2 station in which the maximum value and the minimum value are excluded from j = 1, 2, ..., N.

(3) Further, as described above, at the same time as Δf _ij (t) is detected, C / N (carrier power / noise power) or S / N
When an N (signal / noise) detector (not shown) is provided to detect the C / N of each adjacent base station, C / N = γ _j is set and a load is applied to each frequency error Δf _ij (t). If given, g _3i (t) = [ _{j = 1} Σ ^N γ _j (t) Δ · f _ij (t)} / { _{j = 1} Σ ^N γ _j (t)}] / N (24) Can also be asked.

(4) Furthermore, by combining the above average values g _2i and g _3i , g _4i (t) = [ _{k = 1} Σ ^N-2 γ _k (t) · Δf _ik (t)} / [ _An average value of _{k = 1} Σ ^N-2 γ _k (t)}] ÷ (N-2) (25) may be used.

(5) Further, by setting a threshold value γ _th to the above C / N = γ _j and excluding Δf _ij (t) when γ _j <γ _th from averaging, g _5i (t) = ^{_{[l = 1 Σ N 'Δf i1}} (t) ] / N' (26) can also be used as the average value. However, _{l = 1} Σ ^{N 'is} the number N satisfying γ _l> γ _th' is meant for adding Δf _il (t) of the.

[0122] (6) In addition, by combining the average value g _3i and g _5i above, g _6i (t) = ^{_{[l = 1 Σ N 'γ l}} (t) · Δf il (t)} / _{[l ^{= 1 Σ N 'γ l (}} t)} ] ÷ N' (27) the average value may be used comprising.

FIG. 27 shows an embodiment of temporal averaging (filtering) which is performed as necessary after performing spatial averaging in the averaging unit 2-5 as described above. .. In this case, the following G ^a _ki (t) (k = 1 to
3, a = 1 to 6) is α · Δf of the second term on the right side of the equation (7).
_{It replaces i} (t) and controls f _i (t + 1).

FIG. 11A shows the case where a so-called transversal filter is used, and the spatial average value Δ of the frequency error is
G ^a _1i (t) = _{n = 0} Σ ^M α _n · g _ai with M delay elements T that input f _i (t) in series, M tap coefficients α _n, and one adder ADD1. A frequency control signal of (tn) (28) is obtained. However, α _n <1, and a = 1 to 6 represents the type of spatial average.

Further, a complete integration type filter is shown in FIG. 9B, and in this example, by each multiplier of the multiplication coefficients α and β, the adders ADD2 and ADD3 and the delay device T, G ^a _2i (t) = α · g _ai (t) + β · h _ai (t) (29) h _ai (t) = h _ai (t-1) + g _ai (t) (30) Be done.

Further, an incomplete integral type filter is shown in FIG. 7C, and in this example, by each multiplier of the multiplication coefficients α and β, the adder ADD4 and the delay device T, G ^a _3i The filter average value of (t) = α · g _ai (t) + β · G ^a _3i (t-1) (31) is obtained. This filter is
In (a), Is equivalent to.

When the averaging unit 2-5 performs the averaging operation using each of the filters shown in FIG. 27 and the transmission frequency control unit 2-6 shown in FIG. 10 controls the transmission frequency. 28 to 35 show simulations of the frequency error as the error between the control target values. However, in this example, as shown in FIG. 11, six base stations are adjacent to each other in the base station A. In the base stations B, C, D,
Control is performed by detecting a frequency error between adjacent two, three, and four stations at the shortest distance. Further, in the initial state at t = 0, the frequency error of each base station is random, and −1
It is assumed that they are distributed almost uniformly over ˜ + 1.

FIG. 28 shows the control of the transmission frequency when a simple spatial arithmetic average of g _1i (t) in the above equation (21) is performed, and the frequency error between adjacent base stations with the passage of time. Is approaching "0". The horizontal axis represents time and the vertical axis represents normalized frequency error.

According to G ^a _1i (t) when the temporal averaging by the transversal filter of FIG. 27 (a) is added to the simple spatial arithmetic mean of g _1i (t) in the above equation (21). The frequency error in the case of control is shown in FIG. here,
It is assumed that α _n = 0.5 and M = 8. It can be seen that the frequency error converges (draws in) faster than in FIG. 28, and the time constant is small. However, a small time constant has a side that it is weak against noise.

FIG. 30 shows the frequency error when the temporal averaging by the perfect integral filter of FIG. 27 (b) is added to the simple spatial arithmetic mean of g _1i (t) in the above equation (21). It can be seen that the frequency error converges more quickly than in FIG. 29. In this example, α = Equation (29)
0.8 and β = 0.6.

FIG. 31 shows the frequency error when the temporal averaging by the incomplete integral filter of FIG. 27 (c) is added to the simple spatial arithmetic mean of g _1i (t) in the above equation (21). As shown, it can be seen that the convergence of the frequency error is almost similar to the case of FIG. In this example, α and β in equation (32) are set to α = 0.8 and β = 0.6.

In the simple spatial arithmetic averaging operation of g _1i (t) in the above equation (21), the transition of the frequency error when the reference station R in the base station is set as shown in FIG. 11 is shown in FIG. As shown, it can be seen that the transmission frequency of the reference station R is gradually approached. In this example, the frequency error of the reference station is set to 0.75 so that the effect of the existence of the reference station can be easily understood.

When the reference station R is set,
FIG. 33 shows the frequency error when the temporal averaging by the transversal filter of FIG. 27 (a) is added to the simple arithmetic mean of g _1i (t) in the equation (21), and FIG. By comparison, it can be seen that the asymptotic speed to the transmission frequency of the reference station R is high.

When the reference station R is set,
FIG. 34 shows the frequency error when the time averaging by the perfect integration type filter of FIG. 27 (b) is added to the simple arithmetic mean of g _1i (t) of the equation (21), and It can be seen that the convergence to the transmission frequency is performed while oscillating.

When the reference station R is set, the frequency error when the temporal averaging by the incomplete integral type filter of FIG. 27 (c) is added to the simple arithmetic average of (1) is shown in FIG. It can be seen that the asymptote to the transmission frequency of the reference station R exhibits a waveform similar to that of the transversal filter.

Example of Timing (Phase) Control (FIG. 36)
-FIG. 45): FIG. 36 shows a configuration diagram of a first embodiment of the essential part of the network control system according to the present invention as the preferable mode shown in FIG. 12, and in this embodiment, as shown in FIG. , The radio base station BS1 is located around the radio base stations BS2 to BS
5, a main part of the radio base station BS1 has the configuration shown in FIG. 36, but the same applies to the other radio base stations BS2 to BS5. Each of the radio base stations BS1 to BS5 performs mobile communication with the mobile station by, for example, the TDMA method, and transmits a signal at an arbitrary frame clock frequency and frame phase (control target value). In the following embodiments, the TDMA system will be described as an example, but in the case of the FDMA system and the CDMA system, each of the radio base stations BS1 to BS B is independent of the frame.
S5 can be similarly applied by transmitting a signal at an arbitrary clock frequency and clock phase.

In FIG. 36, reference numerals 201 to 204 denote comparators, which are the peripheral radio base stations BS2 to BS2 shown in FIG.
The frame clock frequency f _{2 of} each of the wireless base stations BS2 to BS5 notified from the BS5 directly by wire or wirelessly.
~f ₅ and frame phase theta ₂ information at time t-1 of _{~θ 5 I (f 2, θ} 2, t-1) ~I (f 5, θ 5, t-1) through the terminal 206 to 209 From the voltage controlled oscillator (VCO) 210 to the radio base station BS.
1 Information I (f ₁ , θ ₁ , t-1) at time t-1
Is input and these are compared.

From the comparators 201 to 204, the frame clock frequency and the frame phase of the peripheral wireless base stations BS2 to BS5 with respect to the frame clock frequency and the frame phase of the wireless base station BS1 itself at time t-1 are obtained. The error information is extracted respectively. in this case,
In each of the comparators 201 to 204, the number of clocks that gives a frame defining a TDMA time slot, that is, the clock frequency and the phase of the clock indicating the beginning of the frame are compared, and the respective error information is extracted. Then, the output signals of these comparators 201 to 204 are respectively supplied to an adder 211 to be added, and then a divider 212 calculates the number of neighboring wireless base stations (BS in this case).
2 to BS5 in total), spatial filtering is performed, and the multiplier 213 multiplies the coefficient α from the coefficient generator 214. As a result, the multiplier 213 outputs the
The first control signal indicating the error ΔB (i, t-1) between the frame clock frequency and the frame phase between the wireless base station BS1 of the own station and the wireless base stations BS2 to BS5 in the surroundings shown in the equation (8) is Taken out.

The first control signal is averaged by the averaging circuit 216, further supplied to the signal processing circuit 215, converted into the second control signal suitable for controlling the VCO 210, and then applied to the VCO 210. , The output oscillation frequency and the phase ΔB (i, t−1) are variably controlled to be small. As a result, from the VCO 210, the above (9)
A signal indicating the size of the node B (i, t) shown in the equation is taken out. The output signal of the VCO 210 is also notified to each of the peripheral wireless base stations BS2 to BS5 as transmission timing information by a wired or wireless line.

In this way, according to the present embodiment of the TDMA system, the radio base stations BS2 to BS having the frame clock frequency and the frame phase of the radio base station BS1 are in the peripheral.
Corrected to be the same as those of 5. In addition, this T
In the case of the DMA method, both the clock phase and the frame clock frequency are controlled by the VCO 210.
Since the MA method and the CDMA method have no relation to the frame, the clock frequency and the clock phase are VCO210.
Will be controlled by.

Next, the second main part of the network control system according to the present invention
An example will be described. In this embodiment, the mobile station and the TDM
In a wireless base station that performs mobile communication according to the A system, and FIG.
As shown in FIG. 5, the wireless base station control unit 401 controls the wireless base stations 402 to 404 that transmit signals at the same frame phase and clock frequency as those of other wireless base stations. However, in the case of this embodiment, only the frame phase is controlled.

In FIG. 38, reference numerals 301 to 304 denote comparators, respectively, which are the frame phase information I of the radio base station itself at time t from a phase control circuit 309 described later.
While (θ ₁ , t) is input, the frame phase information I (θ at time t of the peripheral wireless base stations notified by wire or wirelessly from the peripheral wireless base stations (here, there are four stations). ₂ , t) to (θ ₅ , t) are terminals 305 to 308
Respectively, and the difference between them is detected. Here, the phase control circuit 309 receives the reference frame phase information I from the wireless line central control station 401 shown in FIG.
(Θ ₀ , t) has been input.

The output signals from the comparators 301 to 304 are
After being added by the adder 310, a divider 311 divides by a peripheral wireless base station (here, “4”) to obtain an average value, and then a multiplier 312 outputs the coefficient α from the coefficient generator 313. Is multiplied. As a result, the first control signal indicating the frame phase error between the wireless base station of the own station and the wireless base stations in the vicinity is extracted from the multiplier 312.

This first control signal is averaged by the averaging circuit 315 and then supplied to the signal processing circuit 314, where the second signal having a signal form suitable for the control of the phase control circuit 309 is provided.
After being converted into the control signal of ( ₁ ), it is supplied to the phase control circuit 309 and controlled so that the difference between the reference frame phase I (θ ₀ , t) and the frame phase of the peripheral wireless base stations becomes zero.
That is, although the input signal of the phase control circuit 309 is the reference frame phase, the output signal indicates the frame phase signal of the own station by adjusting the phase of the reference frame phase by the second control signal. The output signal from 304 indicates the phase error information between the frame phase of the local station and the frame phase of the surrounding wireless base stations. The phase control circuit 309 controls the clock phase in the case of the FDMA system and the CDMA system.

Next, a simulation result of a network having "80" nodes as shown in FIG. 40 will be described based on the above embodiment. 41 and 42 show the simulation results when the initial value is given by a random number and the coefficient α is set to 0.5. In both figures, the vertical axis represents the node size B (i, t), and the horizontal axis represents time. FIG. 41 (B),
41C shows the pull-in characteristics of two different nodes, and FIG. 41A shows the pull-in characteristics in which these pull-in characteristics are overlapped for all 80 nodes. As can be seen from FIG. 41A, the sizes of the nodes are all equal after a predetermined time.

In addition, in FIG. 40, when the node 81 is newly added to the network, the characteristic of each node of the network is slightly disturbed at the time of the additional addition as shown by I in FIG. 42 (A). Immediately synchronized. Further, in FIG. 40, when the node 73 is disconnected from the network, the characteristics of each node of the network are slightly disturbed at the time of disconnection as shown by II in FIG. 42 (A), but also immediately synchronized. It turns out that it will be in a state.

In FIG. 42B, the initial value is set to 0 and the node 8
The pull-in characteristic a of the node 80 when 0 is added and the characteristic b when the value B (i, t) of the node 73 is forcibly set to 0 and then disconnected from the network are shown. In this way, it can be seen that each node in the network is in a stable synchronized state even when nodes are added or disconnected.

Therefore, according to the present embodiment, T is used for mobile communication.
When the DMA, FDMA, or CDMA system is applied, timing synchronization of the wireless base station can be taken for non-disruptive handover. In this case, if each node corresponds to a radio base station and its size is frequency or phase information, it can be seen from the simulation results that synchronization can be achieved. Further, according to the present embodiment, it is understood that the frame synchronization is stable without being greatly disturbed even when the number of wireless base stations is increased or the wireless base stations are out of order.

The network configuration example according to the present invention is not limited to FIG. 40, and the examples shown in FIGS. 43 to 45 can be considered. In the embodiment shown in FIG. 43 and the embodiment shown in FIG. 44, the adjacent nodes are basically connected, but in the embodiment shown in FIG. 45, one of the five nodes (radio base station ) Can be locally synchronized, and the remaining four radio base stations can be connected around the radio base station.

An embodiment in which the pull-in range is set (FIGS. 46 to 4)
9): For example, if the timing phase is set as a control target value, and the timing phase of a certain node (base station) in the network already in the mutual synchronization state becomes the mutual synchronization control disabled, this system Since there is no absolute standard of, the mutual average value will settle down in the mutual synchronization state. Therefore, when one node in the network falls out of control, the other nodes follow the value of the node out of control, and the time shown in FIG. Although it takes time as shown in the characteristic graph of the simulation of the signal amount Ti (n), it can be pulled into the mutual synchronization state.

However, especially when a number of uncontrollable nodes occur, it takes a long time to pull back to the mutual synchronization state or the pulling cannot be performed. This state is shown in the simulation graph of FIG. 47, and shows that the signal amount Ti (n) does not decrease because the pull-in does not proceed even if time elapses.

Therefore, as shown in FIG. 48, the received signal from the hybrid 501 is demodulated by the demodulator 502, the received signal from the adjacent node is separated by the separator 503 for each node, and the difference detectors 504-1 are respectively provided. -504-n are compared with the reference value and the respective difference values are compared with the pull-in range detector 50.
Give to 5. The received data is received from the separator 503 via the line terminating circuit 509.

When the pull-in range detector 505 detects one of these difference values that exceeds a predetermined threshold value, the display unit 506 displays the number of the node as alarm information and the node. The reset signal generation unit 507 generates a reset signal corresponding to the number and sends it to the synthesis unit 508.

In the combining unit 508, the line termination circuit 509
When the above-mentioned reset signal is added to the transmission data from the transmission data to the delay adjustment circuit 510, the delay adjustment circuit 510 gives a predetermined delay time to the transmission data by the control signal from the pull-in range detector 505, It will be transmitted from the device 504 through the hybrid 501.

At the node receiving such a reset signal, the separator 503 and the reset signal separator 511 separate the reset signal, reset the pull-in range detector 505, and delay-control the delay adjusting circuit 510 to pull-in. Adjust the timing so that it is within the pull-in range.

As a result, it can be seen that the pull-in time is greatly shortened as shown in the simulation graph of FIG.

Example of node selection (FIGS. 50 to 52):
It has been shown that the time until the mutual synchronization control stabilizes is shortened by setting the pull-in range as described above, but in addition to this, when the timing difference between certain nodes suddenly fluctuates. Shows that if each node randomly selects a comparison target node, its fluctuation spreads to neighboring nodes, and it takes too long (t = 140) until the entire network stabilizes (Fig. 50).

Therefore, as shown in FIG. 51, a certain regularity is given to the selection method of the nodes to be compared, and the node in the direction of the arrow, that is, the node in the direction of the arrow, in order to prevent the fluctuation in the network from returning to its own station, that is, The center of the network is decided, and directionality is given such that control is performed so that only the node closer to the center is synchronized. In practice, the node to be selected may be determined in advance by the directional antennas in each node, and the control may be executed with the directionality shown in FIG. 51 as a whole. Then, the control information of each node uses a broadcast channel such as a wireless BCCH to put the information therein, and each node arbitrarily receives the information and corrects the difference from its own station. ..

By doing so, as shown in the simulation result of FIG. 52, the convergence time until the whole network becomes stable can be shortened (t = 27).

Synchronous control using diversity reception method
Embodiment (FIGS. 53 to 54): As described above, for example, in the case of the network control system of FIG. 10, the frequency error detection unit 3
-4, when calculating the spatial average of the frequency error Δf by the averaging unit 3-5, normally, this Δf is multiplied by the received C / N from the S / N detection unit (not shown) to calculate the load. Is supposed to give. This also applies to the timing control shown in FIG.

However, in such a case, since it is necessary to provide the S / N detection section, in this embodiment, the reception electric field strength measuring device normally provided for each system using the reception diversity system is used. The measured value is used as a substitute for the S / N detector.

FIG. 53 is a combination of the network control system by the frequency shown in FIG. 10 and the reception diversity system, and a reception automatic frequency control unit 3-31 for each of the reception signals from the two reception antennas RA1 and RA2. And 3-32, and further, measuring devices (E1, E2) 3-7 and 3-8 for measuring the reception electric field strength of the intermediate frequency signals in the reception automatic frequency control units 3-31 and 3-32. The comparison and selection unit 3-9 compares the measured values, the larger measured value is given to the averaging unit 3-5, and the selection result is given as a selection control signal for the received data in the selector 3-10. ..

FIG. 54 is a combination of the network control method by timing control and the reception diversity method. The difference from FIG. 53 is that the averaging section 3-5 is provided in the transmission control section 3-6.
Timing control section 3-11 which controls the timing of the modulator by receiving the output signal of the above, the reception timing signal from one reception control section 3-31 and the above timing control section 3-
11 and the transmission timing signal from
-5 is provided with a timing error detection unit 3-11 that gives a timing error signal ΔT.

In this way, in both frequency control and timing control, the S / N detector is not required by taking the weighted average by adding the one with the larger received electric field strength to the frequency error or timing error. ing.

[0165]

As described above, in the network control method according to the present invention, the communication means is provided between the adjacent nodes, and each node transmits its own control target value and receives the control target value from the adjacent station. At the node, one of spatial filtering and temporal filtering is applied to the control target value of the adjacent station, and the following is performed so that the relative error between the filtered control target value and its own control target value is eliminated. Since it is configured to control the control target value of its own station at the time, if a relative error can be found by exchanging information about the control target value with the nodes in the vicinity of its own node and finding a relative error, a special control station will be placed in the center. It is possible to eliminate the relative error related to the controlled object value at least in the vicinity of the own node without the need to provide it, and finally, it is possible to eliminate the relative error related to the controlled object value in the entire system. Properly controlled conditions Te can be obtained.

Further, since the error signal is spatially filtered, even if some of the plurality of spatially spread reference signals have a disturbance or abnormality, the influence can be removed or mitigated by the spatial filtering. Even if there is a disturbance or anomaly in some of the plurality of temporally consecutive error signal sequences, the effect can be removed or mitigated by temporal filtering.

Further, according to the present invention, when the own signal is controlled based on the difference signal between the own signal and the signal from another node so that the difference signal becomes small, and the system is in the initial state. Alternatively, since the control time interval, the filter characteristic, or the level conversion characteristic is changed according to the magnitude of the difference signal, the convergence speed of network synchronization and the stability of synchronization can be optimally maintained according to the state. it can.

Further, according to the present invention, in all base stations, the transmission frequency of the base station is compared with the transmission frequencies of neighboring base stations, and the relative frequency error becomes the nominal interval frequency between the reference stations. Since the transmission frequency of the own station is controlled as described above, in the steady state, the transient AFC operation for absorbing the frequency error of the mobile station is performed only immediately after the power is turned on. In addition, the base station transiently sets A to the mobile station that has moved in or newly transmitted.
There is no need to operate the FC. As a result, it is possible to significantly reduce communication interruption during handover.

Further, if the frequency intervals of the radio channels between the adjacent base stations are accurately controlled, it is necessary to make the frequency intervals of the radio channels between the adjacent base stations wide in advance in consideration of the frequency error. As a result, the frequency utilization efficiency can be improved in the entire communication system.

Further according to the present invention, TDMA, FDM
In both the A and CDMA systems, since a distributed control type network having radio base stations as nodes is configured, even if the number of nodes is increased compared to the conventional centralized control type network, the load on each node is almost the same. In addition, each radio base station is connected to multiple adjacent radio base stations, and even if the line with one radio base station is disconnected, the other normal radio base stations are connected to each other. It is possible to perform stable synchronization control based on the connection, and to have an effect such as having a backup function.

[Brief description of drawings]

FIG. 1 is a block diagram showing a basic method principle of a network control method according to the present invention.

FIG. 2 is a block diagram showing a preferred apparatus embodiment (using a spatial filter) of the present invention.

FIG. 3 is a block diagram showing a preferred apparatus embodiment (using a temporal filter) of the present invention.

FIG. 4 is a block diagram illustrating a preferred apparatus aspect of the invention (using spatial and temporal filters).

FIG. 5 is a block diagram illustrating a preferred apparatus aspect of the invention (using spatial and temporal filters).

FIG. 6 is a block diagram illustrating a preferred device aspect of the invention (using spatial and temporal filters).

FIG. 7 is a block diagram showing a preferred device mode (variable control time interval) of the present invention.

FIG. 8 is a block diagram showing a preferred apparatus mode (variable filter characteristics) of the present invention.

FIG. 9 is a block diagram showing a preferred apparatus mode (variable level conversion characteristic) of the present invention.

FIG. 10 is a block diagram showing a preferred device aspect (frequency control) of the present invention.

FIG. 11 is a diagram showing an arrangement example of base stations.

FIG. 12 is a block diagram showing a preferred apparatus mode (timing control) of the present invention.

FIG. 13 is a network configuration diagram applied to the present invention.

FIG. 14 is a block diagram showing the embodiment of FIG. 2 (A).

15 is a simulation diagram showing a conventional signal amount vs. time characteristic with reference to FIG.

16 is a simulation diagram showing the signal amount vs. time characteristic of FIG.

FIG. 17 is a block diagram showing the embodiment of FIG. 3 (A).

FIG. 18 is a simulation diagram when the time filter is not applied.

FIG. 19 is a simulation diagram for explaining the characteristics of FIG.

FIG. 20 is a block diagram showing the embodiment of FIG. 4 (A).

21 is a block diagram showing the embodiment of FIG. 7. FIG.

FIG. 22 is a block diagram showing an embodiment of a phase detector used in the present invention.

FIG. 23 is an operation timing chart of the phase detector of FIG. 23.

FIG. 24 is a block diagram showing the embodiment of FIG.

FIG. 25 is a block diagram showing the embodiment of FIG. 9.

FIG. 26 is a block diagram showing an embodiment of an error detection unit used in the present invention.

FIG. 27 is a diagram showing an example of time averaging in an averaging unit used in the network control method according to the present invention.

FIG. 28 is a diagram of a simulation result showing changes over time in frequency error between adjacent base stations when simple spatial arithmetic averaging is performed in the averaging unit used in the present invention.

FIG. 29 is a diagram of a simulation result showing the transition of the frequency error with time, when the temporal averaging by the transversal filter is added to the simple spatial arithmetic average in the averaging unit used in the present invention.

FIG. 30 is a diagram of a simulation result showing a transition of a frequency error with time when simple temporal arithmetic averaging is added to a simple spatial arithmetic average in an averaging unit used in the present invention. ..

FIG. 31 is a diagram of a simulation result showing changes in frequency error with time when simple temporal arithmetic averaging is added to simple spatial arithmetic average in the averaging unit used in the present invention. is there.

FIG. 32 is a diagram of a simulation result showing the transition of the frequency error with time when the reference station R in the base station is set in a simple spatial arithmetic averaging operation in the averaging unit used in the present invention. ..

FIG. 33 shows a reference station R in the averaging unit used in the present invention.
FIG. 7 is a diagram of a simulation result showing the transition of the frequency error with time when simple temporal arithmetic averaging is added to the simple spatial arithmetic average, with the setting of.

FIG. 34 shows a reference station R in the averaging unit used in the present invention.
FIG. 9 is a diagram of a simulation result showing a transition of a frequency error with time when a simple spatial arithmetic average is added to a temporal spatial averaging by a perfect integration type filter in the case where is set.

FIG. 35 is a diagram showing an example of a standard station R in the averaging unit used in the present invention.
FIG. 7 is a diagram of a simulation result showing the transition of the frequency error with time when a simple spatial arithmetic average is added to the temporal spatial averaging by the incomplete integral type filter when is set.

FIG. 36 is a configuration block diagram of the first embodiment of FIG. 12.

FIG. 37 is an explanatory network configuration diagram of FIG. 36.

38 is a configuration block diagram of the second embodiment of FIG. 12. FIG.

FIG. 39 is a network configuration diagram for explaining FIG. 38.

FIG. 40 is a block diagram of an embodiment of a network according to the present invention.

41 is an explanatory diagram of characteristics of the network in FIG. 40.

42 is an explanatory diagram of characteristics by adding or disconnecting nodes in the network of FIG. 40.

FIG. 43 is a block diagram of another embodiment of the network according to the present invention.

FIG. 44 is a configuration diagram of still another embodiment of the network according to the present invention.

FIG. 45 is a configuration diagram of still another embodiment of the network according to the present invention.

FIG. 46 is a simulation diagram showing a synchronization pull-in characteristic (time vs. signal amount) when there are few uncontrollable nodes.

FIG. 47 is a simulation diagram showing a synchronization pull-in characteristic (time vs. signal amount) when there are many uncontrollable nodes.

FIG. 48 is a block diagram showing an example of a node having a pull-in range function.

49 is a simulation diagram showing the synchronization pull-in characteristic (time vs. signal amount) of the embodiment of FIG. 48.

FIG. 50 is a simulation diagram showing a synchronization pull-in characteristic (time vs. signal amount) when a neighboring node is randomly selected in the present invention.

FIG. 51 is a diagram for explaining the regularity of node selection in the present invention.

FIG. 52 is a simulation diagram showing a synchronization pull-in characteristic (time versus signal amount) when regularly selecting adjacent nodes in the present invention.

FIG. 53 is a diagram showing an embodiment in which the frequency control and the reception diversity system are combined in the present invention.

FIG. 54 is a diagram showing an embodiment in which the timing control and the reception diversity system are combined in the present invention.

FIG. 55 is a block diagram of a conventional network.

[Explanation of symbols]

11 first calculation portion 12 first error detecting unit 12 ₁ to 12 _k second error detecting section 13 control section 14 second calculation portion 15 first memory unit 16 the third arithmetic unit 17 third error Detection unit 18 Second storage unit 19 Fourth calculation unit 21 Third storage unit 22 Fifth calculation unit 24 Sixth calculation unit 26 Fourth storage unit 27 Seventh calculation unit 2-1 Detection unit 2 -2 signal generation part 2-3 control part 2-4 filter 2-5 control part 2-6 input / output level conversion part 2-7 control part 3-1 base station 3-2 mobile station 3-3 own station automatic frequency control Part 3-4 Frequency error detection part 3-5 Averaging part 3-6 Transmission frequency control part 100 Cell (zone) 101 Radio base station 103 Notification means 102 Mobile station 104 Correction means 201-204, 301-304 Comparator 210 Voltage Controlled oscillator (VCO) 309 Phase control In the circuit diagram, the same reference numerals denote the same or corresponding parts.

 ─────────────────────────────────────────────────── ─── Continuation of the front page (31) Priority claim number Japanese Patent Application No. 3-75399 (32) Priority date Hei 3 (1991) April 8 (33) Priority claiming country Japan (JP) (31) Priority Claim number Japanese patent application No. 3-189403 (32) Priority date Hei 3 (1991) July 3 (33) Priority claiming country Japan (JP) (72) Inventor Satoshi Takema 1015 Kamiodanaka, Nakahara-ku, Kawasaki-shi, Kanagawa Address within Fujitsu Limited (72) Inventor Atsushi Yamashita 1015 Kamiodanaka, Nakahara-ku, Kawasaki-shi, Kanagawa Address Fujitsu Limited (72) Inventor Takeshi Inoue 1015, Ueda-anaka, Nakahara-ku, Kawasaki City, Kanagawa Within Fujitsu Limited

Claims (24)

[Claims] 1. A communication means is provided between adjacent nodes, and each node transmits its own control target value and at the node receiving the control target value from the adjacent station, a space is provided for the control target value of the adjacent station. Either the filtering or the temporal filtering is performed, and the control target value of the own station at the next time is controlled so that the relative error between the filtered control target value and the self control target value is eliminated. A characteristic network control method. 2. In a network control system in which a plurality of radio base stations each transmitting a transmission signal are used as nodes, and a mobile station communicates with a radio base station in a zone in which the mobile station is located, each of the plurality of nodes is A first calculation unit (11) for performing a spatial filter calculation on a reference signal as a control target value from the peripheral node; and an output signal of the first calculation unit (11) and its own output signal. A first error detection unit (12) that outputs an error signal between control target values indicating a difference, and controls a synchronization target input signal as its control target value so as to minimize the error signal. A network control method comprising: a control unit (13) that outputs to a peripheral node as a reference signal indicating information. 3. A plurality of second error detectors (12 _{1 to} 12 _k) that output an error signal indicating the difference between the reference signal and its own output signal instead of the first error detector (12). ) Is provided, and the plurality of second error detection units are provided instead of the first calculation unit (11).
A second arithmetic unit (14) for performing spatial filter arithmetic on each output signal (12 _{1 to} 12 _k ) is provided, and the output signal of the second arithmetic unit (14) is sent to the control unit (13). A network control method characterized in that it is given as a control signal of. 4. A first storage unit (15) for storing a reference signal from the peripheral node at a plurality of time points is provided, and at least a plurality of time points are referred from the first storage unit (15). A third arithmetic unit (16) which performs time filter arithmetic on the signal and outputs the arithmetic result to the first error detection unit (12).
3. The network control system according to claim 2, further comprising: a first arithmetic unit (11) in place of the first arithmetic unit (11). 5. A third error detection unit (17) is provided instead of the first error detection unit (12), which outputs an error signal indicating a difference between the reference signal and its own output signal, The third error detection section (17) in place of the first calculation section (11)
Second storage unit for storing the error signal from
(18) and at least a plurality of time point error signals from the second storage unit (18) are subjected to a time filter operation, and the operation result is given to the control unit (13) as a control signal.
3. The network control system according to claim 2, further comprising an arithmetic unit (19). 6. A third storage unit (21) for storing the output signal of the first arithmetic unit (11) at a plurality of time points, and at least a plurality of time points from the third storage unit (21). The second calculation unit (2) that performs a time filter calculation on the output signal of the above and outputs the calculation result to the second error detection unit (12).
The network control system according to claim 2, further comprising: 7. The first storage unit (15 _1-
15 _k ) and a plurality of third calculation units (16 _{1 to} 16 _k ) are provided in total, and the reference signals extracted by the time filter calculation from the third calculation unit (16 _{1 to} 16 _k ) are extracted. A sixth arithmetic unit (24) is provided, which receives each as an input signal, performs spatial filter arithmetic on the input signal, and outputs the arithmetic result as a control signal to the first error detection unit (12). The network control method according to claim 4. 8. A fourth storage unit (26) for storing the spatial filter-calculated error signal extracted from the second calculation unit (14) at each of a plurality of time points, and at least the fourth storage unit. A seventh arithmetic unit for performing a time filter arithmetic operation on output signals from a plurality of points from (26) and giving the arithmetic result to the control unit (13) as a control signal.
4. The network control method according to claim 3, further comprising: (27). 9. A first error detector (12) for outputting an error signal indicating a difference between the output signal of the first arithmetic unit (11) and its own output signal, and the error signal for a plurality of time points. A second storage unit (18) that stores each
And a fourth arithmetic unit for performing time filter arithmetic on error signals at a plurality of time points from at least the second storage unit (18) and giving the arithmetic result to the control unit (13) as a control signal.
3. The network control system according to claim 2, further comprising (19). 10. In a network control method for synchronizing control target values between a plurality of nodes, a difference between a signal (Q) as its own control target value and a signal (R) as a control target value from another node. A detector (2-1) for detecting the signal (E), and a signal generator (2-) for controlling the own signal (Q) based on the difference signal (E) so that the difference signal (E) becomes smaller. 2)
And a control unit (2-3) for energizing the control of the signal generation unit (2-2), the control unit (2-3) being used when the system is in the initial state or when the difference signal (E ) A network control method characterized by changing the time interval of energization according to the size of. 11. A network control method for synchronizing control target values among a plurality of nodes, wherein a difference between a signal (Q) as its control target value and a signal (R) as a control target value from another node. A detector (2-1) for detecting the signal (E), and a signal generator (2-) for controlling the own signal (Q) based on the difference signal (E) so that the difference signal (E) becomes smaller. 2)
And a filter (2-4) for filtering the signal (R) from another node, the difference signal (E) or the control signal (C) of the signal generator (2-2) based on the difference signal. , Control unit (2-5) that controls the filter characteristics of the filter (2-4)
The network control method is characterized in that the control unit (2-5) changes the filter characteristics when the system is in the initial state or according to the magnitude of the difference signal (E). 12. A network control method for synchronizing control target values between a plurality of nodes, wherein a difference between a signal (Q) as its control target value and a signal (R) as a control target value from another node. A detector (2-1) for detecting the signal (E), and a signal generator (2-) for controlling the own signal (Q) based on the difference signal (E) so that the difference signal (E) becomes smaller. 2)
And an input / output level conversion for converting an input / output level with respect to a signal (R) from another node, a difference signal (E) or a control signal (C) of a signal generator (2-2) based on the difference signal. The controller (2-6) and the control unit (2-7) for controlling the level conversion characteristics of the input / output level converter (2-6) are provided. In the network control system, the level conversion characteristic is changed according to the magnitude of the difference signal (E). 13. In a network control system in which a fixed base station (3-1) and a mobile station are configured as nodes and the transmission / reception frequency as a control target value is the same, each base station (3-1) is an adjacent base station. Of the reception automatic frequency control unit (3-3) that matches the transmission frequency of, and the output frequency of the reception automatic frequency control unit (3-3) and the frequency obtained by adding the nominal interval frequency between the reference stations to the transmission frequency of the own station. A frequency error detection unit (3-4) that detects a frequency error, an averaging unit (3-5) that calculates a spatial average value of the frequency errors for the number of adjacent base stations and performs spatial filtering, A network control method comprising: a transmission frequency control unit (3-6) for controlling the transmission frequency of the local station so as to eliminate a frequency error. 14. The network control system according to claim 13, wherein the averaging unit (3-5) also averages the frequency errors over time. 15. The network control system according to claim 13, wherein a reference station for transmitting an absolute transmission frequency is set in the base station (3-1). 16. The transmission / reception frequency is different, and each base station (3-1) and the mobile station constantly control the frequency difference between the reception frequency and the transmission frequency, respectively. 15. The network control method according to any one of 15. 17. A radio base station (101) for transmitting a transmission signal at a timing as an arbitrary control target value exists as a node in each of a plurality of cells (100), and a mobile station (102) exists in its own existence. In the network control method for performing mobile communication with a wireless base station serving the area, each of the plurality of wireless base stations (101) notifies the nearby wireless base station (101) of transmission timing information of the own station. Notifying means (103) for receiving the transmission timing information notified from the peripheral wireless base station (101) as an input signal, and at least the timing of the transmission signal of the local station is the peripheral wireless base station (101) A network control method comprising: a correction unit (104) that corrects so that a difference between the transmission signal and the timing of the spatial filtering process is reduced. 18. The plurality of radio base stations (101) perform time division multiple access communication with the mobile station (102) and transmit signals at arbitrary frame clock frequencies and frame phases. A base station for transmitting, the correction means (104) compares the frame clock frequency and frame phase information from the surrounding wireless base station with the frame clock frequency and frame phase information of the own station, respectively. Bowl (201
~ 204), and a calculating means (211 to 214, 216) for calculating the difference between the clock frequency and the frame phase between the own station and the surrounding wireless base stations from the output signals of the comparators (201 to 204), and the calculating means ( 18. The network control system according to claim 17, further comprising: oscillating means (210) for variably outputting the frame clock frequency and frame phase of the local station based on the output result of (211 to 214, 216). 19. The plurality of radio base stations (101) perform time division multiple access communication with the mobile station (102), and also perform signal transmission at the same frame frequency and arbitrary frame phase. Is a base station for transmitting, the correction means (104), the frame phase of its own station generated from the frame phase information from the surrounding wireless base station and the reference frame phase information from the wireless link central control station (401) Comparators (301 to 304) for respectively comparing information with each other, and calculating means (310 to 31) for calculating the frame phase difference between the own station and the peripheral wireless base stations from the output signals of the comparators (301 to 304).
3,315), and control means for changing the reference frame phase information from the wireless line central control station (401) based on the output result of the calculating means (310 to 313,315) to generate the frame phase information of the own station ( 309), and the network control system according to claim 17. 20. A frequency division multiple access system, a clock frequency, instead of the time division multiple access system, the frame clock frequency, and the frame phase, respectively.
And a clock phase is used.
Alternatively, the network control method described in 19. 21. The network control method according to claim 20, wherein a code division multiple access method is used in place of the frequency division multiple access method. 22. The network control method according to claim 1, wherein each node has a pull-in range, and control is performed only for a control target value existing in this pull-in range. 23. The network control method according to claim 1, wherein a certain regularity is given to the selection of the node that outputs the control target value. 24. The spatial filtering is a method in which a weighted average is given to a relative error of the controlled object value based on a received electric field strength selected by a diversity receiving system. 23. The network control method according to any one of 23.