JPH0732414B2 - Correction method of total error - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial field of application] The present invention relates to a method of correcting a totalization error, more specifically, a system for connecting a control device and a large number of terminal devices via a communication line to transmit data and perform totalization on the control device side. When a transmission error occurs, it relates to a means for preventing it from becoming a counting error.

[Conventional technology]

The control device and the terminal device are often connected by a line interface, but in this case, a transmission error due to a line error is considered. When high reliability and high processing capacity like a totalizer system for voting tickets (horse tickets) etc. are required, the transmission must be a simple procedure to achieve fast turnaround time, but line error You must take sufficient measures against.

FIG. 3 shows an outline of the totalizer system. In the case shown in the figure, it has a control side having a host CPU and a totalizer controller, and a large number of terminal devices connected to the totalizer controller via a private line, and also outside the satellite field Has a plurality of terminals connected to the totalizer controller through a communication (public corporation) line. Data on voting tickets from the terminal, for example, x race, consecutive winning ceremony, 3-5
The data such as the number, voting amount x yen, etc. is sent, and the totalizer control device side performs the registration and aggregation, but this data transmission is performed by the polling method.

Referring to FIG. 4, it is assumed that both the control device and the terminal start in the idle state, and the control device periodically issues an inquiry signal.
Send ENQ and wait for its response. When the terminal receives a request RQ (issues a ticket), it receives an ENQ, sends attention data (voting content) AD, and waits for a response. When the control device receives the data AD, it sends it to itself or to the host and stores and aggregates it in the host. After this processing is completed, the control device sends response data (ticket issuable) RD and waits for a response. Upon receiving the data RD, the terminal returns a signal ACK to that effect and waits for cancellation. When the control device receives the ACK normally, it does nothing to the terminal for a predetermined guarantee time, and if the terminal does not cancel even after a certain period of time, it issues a ticket. When this completes one cycle, the controller goes into the idle state, and the controller again sends ENQ after a certain period of time.

The normal situation is as described above, the ticketing contents are stored and tabulated, and the tabulation result and the ticketing status match, but there is a case where the ticket is tabulated too much or too much due to an error. That is, the fifth
As shown in the figure, ENQ, AD, AD are stored and totaled, and RD, ACK proceed to the terminal, and the terminal goes into a wait for cancellation.If the ACK did not reach the control device due to a line failure, the control device Cancel and send EOT (End of Text) to the terminal. However, because of a line failure, this does not reach the terminal, and the terminal will issue a ticket without cancellation there. However, in this case, the counting result is less than the actual ticketing state.

Also, as shown in Fig. 6, ENQ, AD, AD memory and total, RD, AC
When I proceeded to K and entered the waiting list for cancellation at the terminal, noise was generated on the line and the terminal sometimes received it as EOT. Actually, the ACK is normally received by the control device, and the control device does not perform a predetermined guarantee time there, and then enters the idle state. EOT came at the terminal, so I accepted it as cancellation,
Does not issue tickets. Therefore, in this case, the totalized result becomes larger than it actually is.

[Problems to be solved by the invention]

As described above, in a totalizer system in which the control device and the terminal device are connected by a long-distance transmission line such as a telephone line, there is a risk that the line is disconnected or noise on the line causes under- or over-counting. The present invention is intended to automatically correct such a counting error when it occurs.

[Means for solving problems]

According to the present invention, a terminal device detects an abnormality in a totaling error correction method in a system in which a control device and a large number of terminals are connected by a communication line, and the control device totalizes data sent from each terminal by a polling method. Occasionally, the recovery processing mode is entered by itself, the recovery inquiry data is returned as a response to the inquiry signal periodically sent from the control device, and the control device sends the final processing data as a response to the recovery inquiry data from the terminal device. The terminal device compares the final processing data that the terminal device has with the final processing data from the control device, and if they do not match, the terminal device performs a totaling or canceling process on the difference data. is there.

[Work]

By providing such an automatic error correction function, manual correction is no longer required when a failure occurs, which is very effective for quick recovery and labor saving.

〔Example〕

First, the recovery process will be described. This recovery process can take an appropriate method. For example, the control device that returned to the polling mode sends ENQ, and when the terminal receives this, it issues the serial number to the ticket issuing data immediately before the recovery processing mode, specifically, the ticket issuing data, instead of the normal ticket issuing data. The number (process number) is sent to the control device. The control device compares the latest processed data with the data sent from the terminal device, and checks whether they match or not.

In the above example, the processing number is also used in the control device, and the processing number is canceled in this example. Therefore, if the processing number of the terminal is i, the processing number of the control device is i-1.
Therefore, the i-th data is stored and tabulated again.
Alternatively, the control device instructs to send the data of terminal i,
In response to this, the terminal sends the i-th data. For this purpose, of course, in the former case, the control device stores the cancel data at least until the next AD storage and totalization is normally completed, and in the latter case, the terminal device waits until the next data is normally received by the control device. The data of is stored. The control device has a large number of terminals under its control, for example, 400 terminals. The current processing number is stored for each of the terminals, and the cancellation data is stored in the former case. The processing number can be used cyclically, and may be a fractional bit, for example, 1 byte, which can cover a counting error. When the terminal sends the recovery data (process number, etc.), it goes into a normal idle state and issues a ticket request RQ.
Send AD if there is and the controller has ENQ.

FIG. 1 shows the processing when an abnormality is detected on the terminal side.

In FIG. 1, when the terminal detects an abnormality, the terminal enters the recovery processing mode and cancels the operation related to the ticket issuing at that time. The controller is in polling mode,
Send me an ENQ. Upon receiving this, the terminal makes a recovery inquiry A
When D is sent and the control device receives the AD, the processing number of the last processed data is sent to the terminal by recovery inquiry RD, and when the terminal receives it, it returns ACK and enters inquiry data check, which it has Compare with the final treatment number. After that, the control device sends ENQ again, but at this time, the terminal device sends the result of the above comparison to the control device as the recovery process AD. Recovery processing AD is, for example, 2 bits,
Each means the following:

0: Data match (recovery process not required) 1: Data mismatch (cancel aggregated data) 2: 〃 (Re-aggregate canceled data) 3: Uncheckable (alarm output because automatic correction is impossible,
Manual correction) When 0 is the normal counting state, 1 is over-counting, and 2 is under-counting. In the case of line failure, most of the time is 0 to 2, and automatic correction is possible. No. 3 can be considered when the terminal itself has a failure, but in this case, it is considered to be a small number of cases, and there is no problem so much as it is dealt with by manual tabulation. Also in this case, the control device stores the previous data for each terminal at least until the next normal processing of the data.

When the control device receives the recovery processing AD, it performs that processing,
Send the recovery process RD to the terminal to notify the end of the process and do nothing during the specified guarantee time. When the terminal device receives the recovery RD, it returns ACK and waits for cancellation. If there is nothing, both will be in an idle state, and the control device will restart the periodic ENQ transmission.

In the recovery processing mode, the recovery processing is performed prior to the normal processing, and automatic correction is performed when there is an error in the aggregation in the previous processing. Also, when the terminal itself is reset due to a failure or when the power is turned on again, the terminal similarly enters the recovery processing mode and is started after the recovery processing procedure. The conditions for the recovery processing mode are collectively indicated as follows: when instructed by the control device side, when the operation is canceled while waiting for the cancellation, and when the power is turned on / reset.
Is.

Although the inquiry data check is performed on the terminal side in FIG. 1, this may be performed on the control device side. FIG. 2 is a diagram for explaining this, where (a) shows the former and (b) shows the latter.

〔The invention's effect〕

As described above, according to the present invention, conventionally, when there is a possibility of a counting error, it was necessary to manually correct by stopping the alarm, but this can be corrected automatically. , A highly reliable totalizer system (voting ticket vending machine control system) can be built. In particular, even in the case of a failure that exceeds the censored range in the procedure, the following procedure guarantees it, which is highly effective in providing a highly reliable system even with a simple procedure.

[Brief description of drawings]

1 and 2 are explanatory views of an embodiment of the present invention, FIG. 3 is an explanatory view of a totalizer system, and FIGS. 4 to 6
The figure is an illustration of a data transmission procedure. In the drawing, AD is attention data, RD is response data, ENQ is an inquiry signal, and ACK is an acknowledgment signal.

Claims (1)

[Claims] 1. A method for correcting an error in a system in which a control device and a large number of terminals are connected by a communication line, and the control device totalizes data sent from each terminal by a polling method. When it does, it enters the recovery processing mode by itself and returns the recovery inquiry data as a response to the inquiry signal periodically sent from the control device at normal times, and the control device sends the final processing data as a response to the recovery inquiry data from the terminal device. To the terminal device, the terminal device compares the final processing data that it has with the final processing data sent from the control device, and returns the comparison result as a response to the inquiry signal, after which the control device If the content of the comparison result returned from the terminal device indicates a mismatch, the difference data A method of correcting a totalization error, characterized by performing totalization or cancellation processing.