CA1232023A

CA1232023A - Method and apparatus for supervising digital radio transmission line

Info

Publication number: CA1232023A
Application number: CA000464445A
Authority: CA
Inventors: Masahisa Kawai
Original assignee: Fujitsu Ltd
Current assignee: Fujitsu Ltd
Priority date: 1983-10-05
Filing date: 1984-10-01
Publication date: 1988-01-26
Also published as: US4677619A; EP0138693A1; DE3470643D1; JPS6077546A; JPH048981B2; EP0138693B1

Abstract

ABSTRACT OF THE DISCLOSURE

This invention provides an apparatus for supervising a digital radio transmission path between a transmitter and a receiver, in which the transmitter transmits a message including a data signal and a predetermined bit pattern portion, the portion being used for error rate detection; the receiver receives the portion, produces error pulses from the portion by comparing the portion to a reference and supervises the transmission path quality based on the produced error pulses to issue a path switching command when the detected error rate exceeds a predetermined threshold value. The apparatus comprises a device for producing a first number equal to the number of received error pulses by counting the received errors in the portion and for producing a second number proportional to a time interval between two successive error pulses; and a comparison unit which receives the first number obtained by counting the received error pulses and the second number produced in accordance with the time interval between two successive error pulses, and produces the line switching command when the first number is greater than or equal to the second number. The supervisory system rapidly detects deterioration in the quality of the digital radio transmission line and performs the line switching when such deterioration is detected.

Description

32~

METHOD AND APPARATUS FOR SUPERVISING
DIGITAL RADIO TRANSMISSION LINE

BACKGROUND OF THE INVENTION

1.- Field of the Invention The present invention relates to a supervisory system for a radio transmission line, more particularly, it relates to a supervisory system which rapidly detects deterioration in the quality of the digital radio transmission line and performs a line switching when such deterioration is detected.

2. Description to the Prior Art The recent trend in radio communication systems is for various information sets to be transformed into digital signals and the thus transformed information sets transmitted to a remote office through a digital radio transmission line. In such radio communication systems, the level of the received signal at the receiver side often is subject to extreme variation due to fading and the like occurring in the radio transmission line, and therefore, a required transmission quality cannot be - attained. In a conventional system, this quality dotter-ration is checked by continuous scanning of an error rate in the normal and emergency lines. If the error rate in a normal line increases to a predetermined threshold value, operation through the normal line is changed to an emergency line by, for example, space diversity, frequency diversity, or polarization wave diversity, so that a desired transmission quality can be assured.
A problem occurring in the prior art supervisory system, is that where a line switching command cannot be issued rapidly. That is, too long a period must elapse before the line can be switched. Thus, by the time the line switching command is actually provided, the quality deterioration in the transmission line quality has ~X320~3 become serious; particularly when fading occurs not gradually but suddenly. The reason for the above will be given in detail hereinafter, but briefly, it is because the line switching command in a prior art system is generated in accordance with each average error rate, which is determined every time a predetermined constant observation term is completed.
SUMMARY OF THE INVENTION
It is an object of the present invention to provide an apparatus for supervising a digital radio transmission line, which will enable the line switching command to be rapidly issued.
The above object is attained by measuring each time interval of successive error pulses which define the error rate and determining whether or not the line switching command is to be issued in accordance with variation in the time intervals.
In accordance with one particular aspect of the present invention, there is provided an apparatus for supervising a digital radio transmission path between a transmitter and a receiver, in which the transmitter transmits a message including a data signal and a predetermined bit pattern portion, the portion being used for error rate detection, the receiver receives the portion, produces error pulses from the portion by comparing the portion to a reference and supervises the transmission path quality based on the produced error pulses to issue a path switching command when the detected error rate exceeds a predetermined threshold value. The apparatus comprises:
means for producing a first number equal to the number of received error pulses by counting the received errors in the portion and for producing a second number proportional to a time interval between two successive error pulses; and ~232023 - pa -a comparison unit which receives the first number obtained by counting the received error pulses and the second number produced in accordance with the time interval between two successive error pulses, and produces the line switching command when the first number is greater than or equal to the second number.
In accordance with another particular aspect of the present invention, there is provided an apparatus for producing a transmission line switching command, the apparatus comprising:
error detection means for detecting error in a received digital radio signal; and change detecting means for detecting changes in a number of the errors detected by the error detection means and producing the switching command when the number is increasing, the change detecting means comprising:
projection means for producing an estimated count number in dependence upon the time between successive errors or a minimum count number when the time between the successive errors exceeds a predetermined value as a projected number;
count means for counting an actual number of errors; and comparison means for comparing the actual number with the projected number and producing the switching command when the actual number is greater than or equal to the projected number.

12320~3 - 2b -BRIEF DESCRIPTION OF THE DRAWINGS
The present invention will be more apparent from the ensuing description with reference to the accom-paying drawings, wherein:
Fig. 1 is a general block diagram of a con-ventional digital radio communication system;
Fig. 2 depicts pulse patterns for explaining the relationship in time between the location of the observation time slots and the issuance of the line switching command;
Fig. 3 is a general block diagram of a super-visor system of the digital radio transmission line according to an embodiment of the present invention;
Fig. 4 is a level diagram for explaining the present invention;
Fig. 5 is a circuit diagram of a detailed example of the supervisory system shown in Fig. 3, and;
Figs. PA, 6B, and 6C depict timing charts of signals appearing at main portions in the circuit of Fig. 5.

DESCRIPTION OF THE PREFERRED EMBODIMENTS
Before describing the preferred embodiments, a prior art supervisory system will be explained for reference.
Figure 1 is a general block diagram of a convent tonal digital radio communication system. In Fig. 1, a data signal to be transmitted is supplied ox an input terminal 14 of a transmitter part and transmitted from a radio transmitting unit 13, via a transmission line TO
which includes antennas, to a receiver part, in which the data signal received, at a radio receiving unit 15, is output from an output terminal 20 and demodulated by a demodulator (DIM) to reproduce the original data.
Reference numeral 11 represents a Sudanese (PUN) code generator which generates a PUN code signal for detecting an error rate of the data signal. The PUN code signal for detecting the error rate is contained in the data signal in the form of frame synchronization bits or parity bits. The PUN code signal is inserted at a line supervisory unit 12 into the data signal, and then transmitted from the radio transmitting unit 13 to the remote receiver. In the line supervisory unit 12, the PUN code bits of the PUN code signal are inserted into the data bits of the data signal at a ratio of, for example, 1:100. That is, a transmission speed ratio is converted from 100 to 101 in the unit 12.
In the receiver part, the data and PUN code signals are received at the radio receiving unit 15 and the PUN
code signal is selectively detected therefrom in a line supervisory unit 16, in which the transmission speed ratio is converted from 101 to 100. The thus detected PUN code and a reference PUN code are applied to an error detection unit 17. The reference PUN code signal is given from a reference PUN code generator 18. This reference PUN code is equivalent to that generated in the transmitter part at the generator 11. In the error detection unit 17, both the received PUN code and the ~23;~0~3 reference code are compared bit by bit. Due to the comparison in the unit 17, every time both bits do not coincide, an error pulse is produced and output at a supervisory terminal 30. Then a counter unit 19 con-netted to the supervisory -terminal 30 at its input operates to count the number of the thus produced error pulses. The counter unit 19 further operates to determine whether or not each count number exceeds a predetermined threshold value every time a predetermined constant observation term for each error pulse is completed. If the count number is exceeded, the counter unit 19 discriminates whether the error rate of the transmission line has become worse than, for example, 10 3. After this discrimination is completed, the line switching command is issued from a decision terminal 21.
The above-recited error rate value of 10 3 signifies that one error bit occurs per 1000 bits of the data signal. In accordance with international standards, a transmission line with an error rate value of 10 3 can no longer be put into commercial use, in that an error rate value of 10 7 through 10 12 is considered suitable for a commercially available transmission line.
In summation, in the prior art, the line switching command is issued in accordance with each average error rate. However, as previously mentioned, in the prior art supervisory system the problem occurs wherein it often takes too long to issue the line switching command when the position of the observation time slot is unfavorable with respect to the time at which the error rate starts deteriorating.
The present invention is specifically directed to a supervisory system comprised of members other than said main data processing members 13, 14, 15, 20, and DIM
shown in the system of Fig. 1.
Figure 2 depicts pulse patterns for explaining the relationship in time between the location of the obser-ration time slots and the issuance of the line switching ., .

1232~23 command. The pulse patterns exemplify a case where the line switching command is issued for a changeover from the normal line to the emergency line, when the error rate in the normal line exceeds a value of 10 3. The error pulse for defining the error rate are illustrated in row (a) of Fig. 2. As shown in Figure, when ten or more error pulses in each observation time slot of term t indicate the occurrence of an error rate value of 10 3 or more, the line switching command must be issued just at the time when ten successive error pulses are detected. In row (b) of Fig. 2, numerals enclosed by quotation marks denote a number of measured error pulses in each observation time slot of the term t regarding the arbitrary error pulse pattern shown in row (a) and row (c) of Fig. 2. According to the observation time slots of row (b), ten or more error pulses in each time slot are measured in the third time slot, as shown by the number "11". Accordingly, the line switching command can be issued at the end of the third time slot, i.e., at the time To. Alternatively, according to the observation time slots of row (c), ten error pulses are measured in the second time slot as shown by the number "10". Accordingly, the line switching command can be issued at the end of the second time slot, i.e., at the time To A comparison of the two modes of rows (b) and I will show that, if the observation time slot are located in the arrangement of row (b), the issuance timing of the line switching command for row (b) is delayed over that of row (c) by To - To As mentioned above, in the prior art, the trays-mission line quality is estimated in accordance with variation in the average error rate, i.e., the number of the error pulses contained in each time slot. Therefore, by the time the line switching command is actually issued, the quality deterioration in the transmission line has become serious; for example, if a quick change in the fading occurs in which the number of error pulses ~L2320;~3 is quickly increased, the difference in time between To and To may be as much as 10 my through 100 my. However, in a typical digital radio communication system, the synchronization necessary in the receiver part not be achieved if the data signal is not supplied at a normal term of over 10 my.
The present invention is based on the above-men-toned observations and has as its object the provision of a supervisory system in which the line switching command is issued in accordance with a changing ratio of the error pulses, which ratio is determined by detecting the variation ox the time intervals of the error pulses.
Figure 3 is a general block diagram of a supervisory system of the digital radio transmission line according to an embodiment of the present invention. The super-visor system of Fig. 3 is mounted in the receiver part of the digital radio communication system. In part-cuter, it is located at the output stage of the error detection unit 17 shown in Fig 1 and connected thereto via the supervisory terminal 30 (shown in both Figs. 1 and 3) for producing the error pulses (refer to row (a) of Fig. 2). In Fig. 3, reference numeral 31 represents an OR circuit, 32 a clock generator, 33 a time interval counter unit, 34 an overflow detection unit, 35 a time comparison unit, 36 a counter unit, 37 an up/down counter unit, 38 a comparison unit, and 39 a memory unit. The decision terminal 21 is also shown in Fig. 1, and the line switching command is issued therefrom if there is an increase in the error pulses.
The supervisory terminal 30 is connected to the comparison unit 38 via a first input of the up/down counter 37, and to the time interval counter unit 33 via a first input of the OR circuit 31. The time interval counter unit 33 is driven by the clock generator 12, and connected to the comparison unit 38 via the time come prison unit 35 and the memory unit 39. The overflow detection unit I is connected, to both a second input 1~20;~

of the OR circuit 31 and a second input of the up/down counter unit 37, and to the -time comparison unit 35 via the counter unit 36. The counter unit 36 is connected to the memory unit 39, and to the comparison unit 38~
The comparison unit 38 is provided, at its output stage, with the decision terminal 21. The operation of the above system is as follows.
First, error pulses are supplied one by one from the supervisory terminal 30. Each supplied error pulse resets the time interval counter unit 33 via the first input of the OR circuit 31. After this reset operation by the error pulse, the time interval counter unit 33 starts courting the clock pulses output from the clock generator 32 and continues the count operation until the time interval counter unit 33 is reset by the following error pulse. Thereby measuring one desired time interval between two successive error pulses. The thus measured time interval is then supplied to the time comparison unit 35. If the following error pulse is not supplied from the supervisory terminal 30 within a predetermined constant term, the time interval counter unit 33 cannot be reset to zero, and therefore, the count number will overflow. The overflow is detected by the overflow detection unit 34, which resets the time interval counter unit 33 via the second input of the OR circuit 31.
The above count number of the time interval counter unit 33 is supplied to the time comparison unit 35. The unit 35 contains a read only memory (ROMP in which pro-determined changing ratio information sets are stored, such as shown below in Table I; where it denotes the time interval in my at which each two successive error pulses are generated, and n denotes a number of core-sponging error pulses.

ox Table I
Wit (no)_ _ n The above values it and n are shown only an example;
the values can be freely determined according to the intention of the system designer.
The changing ratio information sets defined by it and n indicate that (see top row of table) if two successive error pulses occur with a time interval of 1 through 2 my, the transmission line quality is assumed to have an error rate value of over 10 3, and accord-tingly, the line switching command must be issued mime-doughtily. Referring to the bottom row of the table, this indicates that if ten successive error pulses occur with a time interval value of over 32 my, the transmission line quality is also assumed to have an error rate value of below 10 3, and the related command must soon be issued. The above also applies to the remaining lnfor-motion sets (it, n). Thus, the time comparison unit 35 operates first to receive input data concerning the time interval it sent from the unit 33, and then to produce output data concerning the number n of the error pulses corresponding to the thus received input data of it, in accordance with the above recited Table I. The core-sponging number n determined by the unit is supplied to the memory unit 39 and stored -therein. In this case, the memory unit 39 operates in such a manner that the number n to be sorted therein is always renewed to a value smaller than any stored before. That is, the least value n is always maintained. This least value n is then transferred from the memory unit 39 to the comparison unit 38.
The comparison unit 38, receives data concerning N
supplied from the up/down counter unit 37. The unit 37 integrates the number of error pulses applied with its first input (up count input U) and produces the into-grated number de-fined by the above value N. If the comparison unit 38 determines that the number from the unit 37 is larger than the number n from the unit 39, it decides that the related transmission line quality may have an error rate value of over 10 3, and accordingly, produces, at the decision terminal 21, the line switching command.
The counter unit 36 counts the number of overflow detection signals sent from the unit 34 and resets the gist of the memory unit 39 every time the count number no reaches, for example, 2. This number "2" is used for assuring a trend indicating that the error rate is gradually decreasing, which is schematically shown by and so on (skipping the intermediate numbers 85, 70, 60 there between) with reference to row (c) of Figs. 6B and 6C. Alternatively, if an overflow detection signal is not supplied to the counter unit 36, the gist of the memory unit 39 is not reset, but is renewed by the least value n from the time comparison unit 35, and the value n is maintained in the unit 39.
The counter unit 36 transfers a discrimination signal, via a line L, to the time compairons unit 35.
This discrimination sisal reveals whether the reset operation for the time interval counter unit 33 is caused by the error pulse sent to the first input of the OR circuit 31 or by the overflow detection signal sent to the second input -thereof. If -the former, the time comparison unit 35 produces the number n corresponding to the time interval it from the unit 33 in accordance - 10 - ~;~320Z3 with the information sets shown in Table I. If the latter, the time comparison unit 35 produces the largest value n, i.e., 10 in Tale I, since an error pulse is not generated within at least the aforesaid predetermined constant term.
The uptown counter unit 37 works as a down counter and the integrated number N is decreased if the overflow detection signal is supplied from the unit 34 to the second input (down count input Do of the up/down counter unit 37, since no error pulse is supplied from the OR
circuit 31 to the time interval counter unit 33 at least within the predetermined constant term. In this case, the down count operation is stopped when the decreased number N reaches zero.
Thus, the system of Fig. 3 performs the supervising operation for the digital radio transmission line in terms of both the number N of the error pulses and the time interval at thereof in relation to the predetermined value n corresponding to each time interval it. This enables a rapid issuance of the line switching command, particularly when rapid changes occur in the fading, and accordingly, the digital radio communication system can be still maintained at normal operation through line switching even when such rapid changes occur in the fading.
Figure 4 is a level diagram explaining the present invention. In the level diagram, the ordinate indicates a level having an electric field strength En at the receiver part and the abscissa indicates an elapsed time. The level change at En is apparently induced by the fading. If serious fading occurs, the level of En is greatly reduced. When the level of En is reduced, due to fading, below a critical level Two the line switching command is actually issued. Contrary to the above, in the present invention, the time taken to issue the line switching command is shorter, as denoted by to in Fig. 4, even though the rapid change of the fading F

takes place. This is because the supervisory system of Fig. 3 has an inherent capability to forecast an occurrence of the state shortly before the level of En reaches the threshold level Thy. This forecasting function is derived from a characteristic wherein the line switching command is issued in accordance with the changing ratio of the error pulses. Conversely, the emergency line is also quickly changed over back to the normal line once the transmission line quality of the normal line is restored to an available error rate value of below 10 3.
As can be understood from the above description, an disadvantage exists in that an erroneous line switching comma may be issued if fading such as that shown by F2 takes place; since the supervisory system of the present invention has the forecasting function, the system will forecast that the digital radio communication system can no longer be maintained, if the related ratio trays-mission line, deteriorated by the fading, is used, as a large number of error pulses have been generated at this point. Another threshold level Thy is defined above the level Two and the line switching operation must be started at this level Thy. In the prior art, as pro-piously mentioned, it takes a long time to issue the line switching command, as denoted by if in Fig. 4.
Therefore, the digital radio communication system can often no longer be operated normally at the time when the error pulses increase along a broken line EN and will soon produce the line switching command. However, in actuality, the level of En does not reach the thresh-old level Thy but increases, as shown by F2 in Fig. 4, and therefore line switching would not be needed.
However, this disadvantage wherein an erroneous line switching command is produced is not considered a serious problem for digital radio communication systems, since line switching per so always has a good effect on the radio communication system. In addition, such a disadvantage can be completely offset by the advantage of the present invention, i.e., a rapid response to the fading.
Figure 5 is a circuit diagram of a detailed example of the supervisory system shown in Fig. 3. In Fig. 5, the members identical to those of Fig. 3 are represented by the same reference numerals. In the present invent lion, the up/down counter 37, the time interval counter units 33-l, 33-2, and the counter unit 36 can be a known lo product such as the "4029" series; the comparison unit 38, and a part of the time comparison unit 35, i.e., 35-l, 35-2, can be a known product such as the "4585"
series; and the other part of the time comparison unit 35, i.e., 35-3, 35-4, can be a known product such 15 as the "4035" sons. -The operations of the circuit shown in Fig. 5 will be clarified with reference to the timing charts of signals shown in Figures PA, 6B, and 6C.
Figures PA, 6B and 6C depict timing charts of 20 signals appearing at main portions in the circuit of Fig. 5. The time chart of Fig. PA represents a case where the radio transmission line is in good condition, and therefore the error pulses are not frequently generated.
The time chart of Fig. 6B represents a case where the condition of the radio transmission line is detent-orated by the fading, and therefore, the error pulses are generated very frequently, thereby causing the line switching command to be issued.
The time chart of Fig. 6C represents a case where the fading disappears and the line switching command is released. In each of the Figs. PA, 6B and 6C, rows pa), (b), I --- I show pulses appearing at portions a, b, c --- k in Fig. 5, respectively. With reference to 35 Figs. 5 and PA, the error pulses appear at the portion a with the pulse pattern of row C The error pulses at a are sequentially applied to the up/down counter unit 1~32~3 37, and therefore, the count number thereof at the portion b changes with 1 or 0, as shown in row b. The preset value at the portion c is decrement Ed with the timing shown in row (c). The time interval counter unit 33-1 produces, at the portion d, the overflow detection signal as shown in row d. The overflow detection signal resets the count number of the counter unit 37 to 0, while the count number is incremented by one (b) every time the error pulse (a) is generated. A
signal at the portion h rises to "H" (high) level every time the error pulse is generated, and then falls to "L"
(low) level. The "H" level signal of the portion h gives an instruction to hold the count number at the portion c, but in Fig. PA, the held number is zero, since the error pulses are not frequently generated.
Thus, no line switching command is issued at the portion k as shown in row (k).
With reference to Figs. 5 and 6B, the line switching command is raised as shown by CUD in row (k), since the error pulses are generated very frequently due to fading.
The count number of the unit 37 at the portion b is incremented one by one, e.g., 1, 2, 3, and so on, in response to the error pulses at the portion a. Each error pulse of row (a) induces the signal of row (g).
This signal g gives an instruction to hold the count number of the unit 33-1, 33-2 appearing at the portion (c), since the signal of row (h) is now changed to "H"
level by the first error pulse of row (a). According to each signal of row (g), the count numbers, such as 30, 40, 50, and so on, are held at the portion e, as shown in row (e). In this case, a preset value "100" is given from the switches SW to the unit 33-1, 33-2, every time the error pulse of row (a) is generated. If the count number at the portion c is 30, the ROM 35-5 of the unit 35 produces the number 8 at the portion j as shown in row (j). Similarly, the ROM produces 7, 6, 5, and so on as shown in row (j) in response to the count number ~232~Z3 40, 50, 60, and so on as shown in row (e). In this case the ROM 35-5 has the information sets as shown by Table II.

Table II
Row (c) Row (j) 70 ' 4 ' 10 When the unit 38 detects that the count number of row (b) coincides with the number of row (j), it produces the line switching command CUD as shown in row (k), where the number is "5".
With reference to Figs. 5 and 6C, the error pulses are not frequently generated, as in Fig. PA. Therefore, the hold number of row (e) is decrement Ed by e.g., 80, 70, 50, and 40. These numbers 80, 70, 50 and 40 are read in accordance with signals of row (i). At this time, the ROM 35-5 produces a fixed number 6 as shown in row (j), and the count number of row (b) is decrement Ed by e.g., 9, 10, 9,'10, 9, 8, 7, 6, 7, 6, 5, and so on.
If the count number of row (b) becomes smaller than the number "6" produced from the ROM 35-5, the co line switch command is released, as shown by CUD in row (k).
As explained above in detail, the present invention can provide quick response to the fading due to the aforesaid forecasting function. Accordingly, the line switching can be achieved without error even though a quick change of the fading takes place.

Claims

The embodiments of the invention in which an exclusive property or privilege is claimed are defined as follows:

1. An apparatus for supervising a digital radio transmission path between a transmitter and a receiver, in which the transmitter transmits a message including a data signal and a predetermined bit pattern portion, the portion being used for error rate detection, the receiver receives the portion, produces error pulses from the portion by comparing the portion to a reference and supervises the transmission path quality based on the produced error pulses to issue a path switching command when the detected error rate exceeds a predetermined threshold value, said apparatus comprising:
means for producing a first number equal to the number of received error pulses by counting the received errors in the portion and for producing a second number proportional to a time interval between two successive error pulses; and a comparison unit which receives the first number obtained by counting the received error pulses and the second number produced in accordance with the time interval between two successive error pulses, and produces the line switching command when the first number is greater than or equal to the second number.

2. An apparatus as set forth in claim 1, wherein said means comprises:
an up/down counter unit, having an up count input, for achieving an up count every time the error pulse is applied to the up count input;

a time comparison unit, operatively connected to said up/down counter unit, which computes the time interval, and produces a new second number and outputs the new second number if the new second number is less than a previously produced second number; and a memory unit, operatively connected to said time comparison unit, for storing the second number updated as the least value produced by said time comparison unit, said time comparison unit including a ROM which stores information defining a relationship between each time interval of the error pulses and the corresponding second number.

3. An apparatus as set forth in claim 2, further comprising a clock generator connected to said time interval counter unit and producing clock pulses and wherein the second number is produced by the ROM in response to data concerning the time interval produced by said time interval counter unit which is driven by the clock pulses supplied from the clock generator.

4. An apparatus as set forth in claim 3, further comprising an overflow detection unit connected to said time interval counter unit and said up/down counter, and which detects an overflow of the count by said time interval counter unit and produces an overflow detection signal which is supplied to reset the count number of said time interval counter unit and supplied to the down count input of the up/down counter unit.

5. An apparatus as set forth in claim 4, further comprising an OR circuit connected to the overflow detection unit and to receive the error pulses and wherein the overflow detection signal is supplied, via said OR circuit, to said time interval counter unit, when said OR circuit receives the error pulses, said time interval counter unit is reset every time one of the error pulses is received or when the overflow detection signal is generated.

6. An apparatus as set forth in claim 5, further comprising a counter unit connected to said overflow detection unit, and said time interval comparison unit, for counting the number of the overflow detection signals produced from the overflow detection unit and resetting said time interval comparison unit each time the count number of the overflow detection signals reaches a predetermined number suitable for assuring a trend indicating that the error rate is gradually decreasing, and when the overflow detection signal is not supplied to the counter unit, said time interval counter unit is not reset and said memory unit is updated by the least value from the time comparison unit, the counter further operates to transfer a discrimination signal via a line, to the time comparison unit, which discrimination signal indicates whether the reset operation for the time interval counter unit is caused by the error pulse provided to a first input of the OR circuit or the overflow detection signal given to a second input thereof, wherein, in the former case, the time comparison unit produces the number corresponding to the time interval from the time interval counter unit in accordance with the information stored in the ROM, and, in the latter case, the time comparison unit produces the largest second number stored in the ROM.

7. An apparatus for producing a transmission line switching command, said apparatus comprising:
error detection means for detecting error in a received digital radio signal; and change detecting means for detecting changes in a number of the errors detected by said error detection means and producing the switching command when the number is increasing, said change detecting means comprising:
projection means for producing an estimated count number in dependence upon the time between successive errors or a minimum count number when the time between the successive errors exceeds a predetermined value as a projected number;
count means for counting an actual number of errors; and comparison means for comparing the actual number with the projected number and producing the switching command when the actual number is greater than or equal to the projected number.