JPH0572859A - Electrophotgraphic process controller - Google Patents

Abstract

PURPOSE:To provide a controller capable of optimally controlling an electrophotographic process mechanism. CONSTITUTION:This electrophotographic process controller is an image forming process controller determining an optimum manipulated variable for forming an image, composed of a measuring part measuring the states of each part inside an image forming device, a preprocessing part 1 converting information obtained from the measuring part into a parameter displaying a prescribed state, and a neural network having prescribed learning in advance, and has a state predicting part 2 inputting the parameter and the information on the manipulated variable to the input layer of the neural network and outputting the predictive state of a specific part inside the image forming device from the output layer of the neural network, a state comparing part 4 comparing the output of the state predicting part 2 with the target state of the specific part, and a manipulated variable determining part 3 making a difference in the manipulated variable from the present one, based on the result of the comparison of the state comparing part 4, sending it to the state predicting part 2, and determining a quantity obtained by adding the difference to the present maniplated variable, as the optimum manipulated variable, when the result of the comparison is converged.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a copying machine, a printer, etc. for copying or printing by an electrophotographic process mechanism.
The present invention relates to an electrophotographic process control device for controlling each part of an electrophotographic process and each part of the image forming device in an image forming apparatus such as a facsimile and its image forming apparatus.

[0002]

2. Description of the Related Art In a conventional development control system in an electrophotographic process, a surface potential and a toner adhesion amount on a photosensitive drum are measured by a surface potential meter, a photo sensor, etc., and a table prepared in advance by experiments ( Hereinafter referred to as a "table")
A method of controlling by the toner supply amount or the operation amount of the developing bias corresponding to each measured value obtained by referring to
A method in which the operation amount of each part in the device is changed and the device state is fed back by a sensor or the like to find an optimum operation amount using a known method such as PID control, and further, a fuzzy inference operation is performed. By mounting the device, a method of operating each control target based on a comprehensive judgment from many state parameters that are intricately and entangled is devised.

Reference patent: [1] Japanese Patent Application Laid-Open No. 2-311860
Fuzzy control system for copying machine [2] Japanese Patent Laid-Open No. 63-85769: Image forming apparatus

[0004]

However, in the above-mentioned conventional control method for the electrophotographic process, any control method has the following problems.

First, the method of controlling by the operation amount corresponding to each measured value obtained by referring to a table created in advance by experiments or the like cannot obtain the optimum operation amount for all cases of the apparatus.

That is, it is necessary to carry out an experiment assuming all cases of the apparatus in consideration of environmental conditions and to store all the operation value data in the apparatus as a table in the labor of the experiment and the cost of the apparatus. Is not realistic considering.
Further, in order to cover all cases of the device, a huge amount of experiments are required, and the amount of data of the table held in the device becomes very large. Therefore, in this type of system, as a result, the values that are considered safe only in a typical case of the device are retained in the device as a table, and it is difficult to perform optimum control in all the cases of the device. Furthermore, in consideration of machine differences (meaning differences in characteristics of each image forming device due to manufacturing variations, etc.) and changes over time, it is necessary to include a certain amount of margin (error) in the operating values. It is very difficult to perform control with respect to the optimum operating value. Secondly, among the parameters that represent the state of the image forming apparatus, there are limits to the measurement such as the surface potential of the photosensitive drum and the toner adhesion amount. Therefore, in practice, feedback using PID control or the like is repeated. It is difficult to control.

In the conventional control method, the surface potential of the photoconductor, the toner adhesion amount, etc. are measured, and the value (operation value) of the voltage / current or the like given to the charger or the exposure device is measured while feeding back the value.
Is operated by a method such as PID control. Then, in order to determine the measurement-operation value, the measurement is repeated until the surface potential of the photoconductor or the toner adhesion amount converges to a target value. However, in practice, the surface potential of the photoconductor, the toner adhesion amount, and the like cannot be repeatedly measured because of the following limitations.

For example, in order to measure the surface potentials of a charging portion and an exposing portion of an electrostatic latent image (hereinafter, simply referred to as a latent image) formed on a photosensitive drum, a latent image pattern serving as a measurement reference is exposed. An image is formed on the body drum, and the surface potential of the charged portion and the exposed portion (some halftone portions if necessary) of this pattern is measured. However, the latent image pattern generated on this photoconductor drum is inevitably developed (toner is attached to the photoconductor drum) in the process, so the toner image of this pattern is not transferred to transfer paper or the like. As far as possible, the cleaning part of the device is extremely burdened, and in some cases cleaning cannot be performed. Actually transferring the toner image of this pattern onto a transfer paper or the like causes waste of the transfer paper and an increase in processing steps.
In an apparatus using such a control method, in order to protect the cleaning portion, it is required to reduce the generation frequency of this pattern as much as possible. Further, when such a pattern is created, wasteful consumption of toner is also increased.

Further, in the conventional control method, even when the amount of toner adhered to the photoconductor is measured, a toner image having the same pattern as that described above is created, and the same problem as cleaning of this pattern image occurs. Because
The measurement frequency is limited. In addition to the above-mentioned problems, it takes time to perform the measurement because the process of charging-exposure-developing-cleaning is performed, and the repeated measurement deteriorates the performance of the apparatus itself, that is, the copying speed and the printing speed of the apparatus. And a problem that causes an increase in the time from first copy to print output.

Furthermore, since each parameter for determining the manipulated variable depends on each other in a complicated manner, it is difficult to converge the control by the conventional PID control or a simple rule.

Such a limitation that the surface potential of the photoconductor cannot be measured forever is naturally a problem even in the conventional method using the fuzzy calculation which is premised on that the surface potential of the photoconductor is known.

The present invention has been made in view of the above-mentioned circumstances, and in an electrophotographic process, a neural
By using a network-based image forming apparatus state predicting means to estimate and generate parameters with limited measurement frequency such as surface potential on the photoconductor and toner adhesion amount and parameters difficult to measure, To provide an electrophotographic process control device capable of solving a problem and realizing a copying machine, a printer, a facsimile or the like which outputs a higher quality image by optimally controlling each part of the image forming device. With the goal.

[0013]

In order to solve the above-mentioned problems, the invention according to claim 1 is an image forming process control apparatus for determining an optimum operation amount for image formation, which is provided in the image forming apparatus. The measurement unit for measuring the state of each part of the above, a preprocessing unit for converting the information obtained from the measurement unit into a parameter representing a predetermined state, and a neural network which has been subjected to predetermined learning in advance. , A state prediction unit that inputs the information about the parameter and the operation amount to the input layer and outputs the predicted state of a specific portion in the image forming apparatus from the output layer, an output of the state prediction unit and the specific portion Based on the comparison results of the state comparison unit that compares the target state with the state comparison unit, a difference amount of the operation amount from the current operation amount is generated and sent to the state prediction unit, and the comparison result converges. It is configured to have an operation amount determination unit that determines the amount of addition of the difference amount to the current operating value as the optimum operation amount when.

According to a second aspect of the present invention, in the electrophotographic process control apparatus according to the first aspect, a fuzzy arithmetic unit for performing a fuzzy arithmetic operation is provided, and the fuzzy arithmetic operation is performed to operate each part in the image forming apparatus. It is configured to determine the amount.

According to a third aspect of the present invention, in an image forming apparatus of an image forming apparatus such as a copying machine, a printer and a facsimile having an electrophotographic process mechanism, the surface potential of a photoconductor as an electrostatic latent image carrier and toner adhesion. Measuring means for measuring the internal / external state of the image forming device such as quantity, temperature and humidity with a sensor,
The internal / external state of the image forming apparatus and / or the operation amount information inside the image forming apparatus such as the charge amount such as the electric potential of the charger or the grid and the exposure amount obtained by the measuring means are stored in the image forming apparatus. The pre-processing means for converting into a parameter representing the state and the parameter representing the state of the image forming apparatus are input, and the parameter representing the state of the image forming apparatus previously obtained by an experiment or the like is used as a teacher value to form the image forming apparatus. State prediction means for estimating the state of the image forming apparatus using a neural network in which the characteristics of the system have been learned, and the state of the image forming apparatus estimated by the state estimating means and the control target. A state comparing means for comparing the state of the image forming apparatus, and the operation amount of each part of the image forming apparatus is calculated from the comparison result output by the state comparing means,
An operation amount determining means for determining an optimum operation amount of each part of the image forming apparatus is provided.

The invention according to claim 4 is the same as claim 3
The electrophotographic process control apparatus described above has a fuzzy operation means for performing a fuzzy operation, and the operation amount of each part of the image forming apparatus is determined by performing the fuzzy operation.

According to a fifth aspect of the present invention, in an image forming apparatus having an electrophotographic process mechanism, a measurement for measuring the surface potential of the photosensitive drum, the image density, the internal state of the apparatus and the external state of the apparatus. Preprocessing means for converting the information obtained by the means into parameters representing the state,
State prediction means using a neural network in which the characteristics of the system of the device are learned by using a parameter representing the state obtained in advance by experiments as a teacher value, and the prediction result of this state prediction means and / or the internal state of the device Fuzzy calculation is performed by using the parameter indicating the external state of the apparatus and / or information from the operation panel of the apparatus and / or the current operation amount, and each section of the apparatus such as voltage and current of the charger and the exposure unit of the apparatus. And an operation amount determining means for determining an operation amount to be sent to each control unit for controlling the above.

[0018]

According to the present invention, the state of the photoconductor is predicted by using the parameters representing the internal state of the image forming apparatus and the external state of the apparatus and the state predicting means using the neural network, and the optimum state of each unit of the apparatus is predicted. The amount of operation is determined by fuzzy calculation, and fine-tuned control suitable for each case is possible. Further, forming an image on the photoconductor and measuring the surface potential and the toner adhesion amount are performed once per control.
Since it is performed only once, the number of times the developed reference pattern is cleaned is reduced, and the burden on the cleaning portion is reduced. In addition, even if a system change that is difficult to predict occurs due to changes over time, and the predicted state and the actual state are inconsistent with each other, the neural network is trained at any time, and the system change is detected in the state prediction means. Can be dealt with. In addition, the generalization capability of neural networks allows state prediction tools to be built with less experimentation.

[0019]

Embodiments of the present invention will now be described in detail with reference to the drawings.

However, regarding the configuration and operation of the range that can be clearly recalled from the description of this specification, illustration and disclosure thereof will be omitted or simplified in order to avoid complication of description. First, an embodiment of the invention described in claims 1 and 3 will be described.

FIG. 1 is a diagram showing the basic construction of an electrophotographic process control apparatus according to the present invention. First, the basic configuration of this device will be described.

In FIG. 1, the electrophotographic process control apparatus according to the present invention takes in the internal state of the apparatus (for example, the operation value of the charger or the exposure unit) and the external state (for example, the external temperature and humidity) from a sensor or the like. , A preprocessing unit 1 for generating parameters representing the internal / external state of the apparatus, a state prediction unit 2 for predicting / estimating the state of each unit of the image forming apparatus using a neural network having a learning function,
An operation amount determination unit 3 that determines the operation amount to be sent to each element driving unit of the image forming apparatus such as a charger or an exposure device, and a target state and a predicted state are compared, and a state comparison is performed to generate the difference. It consists of part 4.

Further, the current state having a parameter representing the current state of the image forming apparatus, the operation amount and the operation amount difference currently given to the image forming apparatus are input into the state predicting section 2, and the operation is performed in the current state. A neural network for predicting and outputting the state of the image forming device when the amount is changed by the operation amount difference and given to the image forming device is incorporated.
In this neural network, a parameter representing a state obtained in advance by an experiment or the like is used as a teacher value, and it has been learned by a method described later.

Next, the operation of the electrophotographic process control apparatus according to the present invention will be described.

(1) Control Operation FIG. 2 shows the state of the apparatus at time t. At this time, the state of the image forming apparatus generated by the preprocessing unit 1 is S (t), and the operation amount sent to each element driving unit of the image forming apparatus is E (t).
The state prediction unit 2 determines the operation amount E (t) at the current time (time t).
And the system of the electrophotographic process is emulated from the current state S (t) of the image forming device, and the manipulated variable E (t) is calculated as the manipulated variable difference ΔE.
It is assumed that the state s (t + 1) of the image forming device at the time of changing only (t) (time t + 1) is predicted.

In the initial state, since the manipulated variable difference ΔE (t) is 0, the predicted state s output from the state prediction unit 2
(t + 1) becomes the same as the current state S (t) (it is expected that the state does not change because the operation amount is not changed).

For example, when a sudden temperature change occurs and it is necessary to reset the operation amount to be sent to each element drive unit of the image forming device due to a change in internal / external state of the image forming device (control is performed). If it is done), follow the procedure below (see FIG. 3).

The internal / external state of the image forming apparatus is fetched from each sensor, and the state S (t) is generated by the preprocessing unit 1. The generation of this state means, for example, the analog output from the sensor.
Data can be converted to digital data, neural
It refers to the process of generating parameters that normalize values for input to the network.

In the manipulated variable determiner 3, the manipulated variable difference ΔE (t) =
The value is set to 0 (initial value), and the value is sent to the state prediction unit 2.

The state predictor 2 calculates the operation amount E from the current state S (t), the operation amount E (t) and the operation amount difference ΔE (t).
The state s (t + 1) of the image forming apparatus when (t) changes by the operation amount difference ΔE (t) is predicted and output.

The state comparing section 4 compares the predicted state s (t + 1) with a preset target state Si and sends the difference to the manipulated variable determining section 3 as a state difference e.

The manipulated variable determiner 3 changes the manipulated variable difference ΔE (t) so that the absolute value of the state difference e becomes small, and sends the value to the state predictor 2.

The above-described operation is repeated until the state difference e becomes smaller than the minimum or allowable value.

The manipulated variable determiner 3 uses ΔE (t) when the above operation converges (when the state difference e becomes the minimum or becomes smaller than the allowable value) by using the following (time t +
New manipulated variable E (t + 1) = E (t) + ΔE (t) in 1)
Is generated and sent to each element driving unit of the image forming device (see FIG. 4).

For example, the operation of the present invention will be described by taking the case where the present invention is applied to the control of the surface potential of the charging portion of the photoconductor as an example. The operation target is the voltage applied to the grid of the charger.
FIG. 7 is an example of a surface potential state predicting unit of a charging unit configured by using a three-layer hierarchical neural network. FIG. 8 is a flowchart for determining the manipulated variable.

First, a latent image pattern as a reference is formed on the surface of the photoconductor, and the charging potential in the current state is measured. When the absolute value of the difference (state difference) between the charging potential of the current state and the target charging potential of the target state is larger than the allowable range, the operating amount of the charger grid potential is adjusted while adjusting the difference of the operating amount of the charger grid potential. The state prediction unit is made to predict the charging potential when the difference is changed. The adjustment of the manipulated variable difference and the prediction of the charging potential are repeated until the absolute value of the difference (state difference) between the predicted charging potential in the predicted state and the target charging state in the target state becomes smaller than the allowable range, and the difference is allowed. The difference between the manipulated variable of the charger grid when it reaches the range and the current potential of the charger grid is sent to each element drive unit as the potential of the new charger grid.

(2) Learning Operation Next, the neural network in the state prediction unit 2
Follow the steps below to learn.

Learning the current state (see FIG. 5).

When the manipulated variable difference ΔE is 0, the predicted state s must be equal to the current state S regardless of the manipulated variable.

Therefore, at time t, the current state S
(t), the manipulated variable E (t), and the manipulated variable difference ΔE (t) = 0 are input, the learning value T (t) = S (t) is set, and learning of the neural network is performed by, for example, a back propagation learning rule. To do.

Learning prediction state (see FIG. 6).

Next, the learning of the prediction of the state when the manipulated variable is changed by ΔE (t) is performed as follows. First, the state S (t) of the image forming state at time t, the operation amount E (t) to each element driving unit of the image forming apparatus, and the operation amount difference ΔE (t) are fixed as inputs. Next, a new manipulated variable E (t + 1) = E
(t) + ΔE (t) is generated by the manipulated variable determiner 3 and sent to each element driver of the image forming apparatus. Then, each part of the apparatus (charger grid,
Control the exposure device). The internal / external state of the new image forming device is fetched from the sensor etc., and the state S (t + 1) is set in the preprocessing unit 1.
After generating, the learning of the neural network is performed with the teacher value T (t + 1) = S (t + 1).

Next, an embodiment of the invention described in claims 2 and 4 will be described.

FIG. 9 is a diagram showing a basic configuration of the inventions described in claims 2 and 4. In the control device according to the present invention, the operation amount determining unit 3 and the state comparing unit 4 in the basic configuration (see FIG. 1) of the invention according to claim 1 are provided in the operation amount determining unit 3 having a fuzzy arithmetic unit therein. Use the replaced configuration. The manipulated variable determiner 3 includes a state comparator 4 and a fuzzy calculator 5.

Next, the operation of the electrophotographic process control apparatus according to the present invention will be described.

(1) Control Operation The basic operation of the control device is the same as the invention described in claim 1 except for the following points. That is, the manipulated variable determiner 3 performs a fuzzy calculation using the state difference e and / or the predicted state s (t + 1) and / or the target state Si to reduce the absolute value of the state difference e. ΔE
(t) is generated and sent to the state prediction unit 2 (the same as the invention according to claim 1). In the fuzzy calculator 5, the manipulated variable determiner 3 receives the state difference e and / or the predicted state s (t + 1) and / or the target state Si as input, and the manipulated variable difference ΔE
It is assumed that the rule for generating (t) and the membership function for fuzzifying / defuzzifying each input / output parameter are created by experimentation or the like and are incorporated in advance.

(2) Learning operation The learning method of the neural network in the state predicting section 2 is the same as the invention described in claim 1.

Next, an embodiment of the electrophotographic process control apparatus according to the present invention will be described.

FIG. 10 is a block diagram showing the basic arrangement of an electrophotographic process control apparatus according to the fifth aspect of the invention.
That is, the electrophotographic process control apparatus according to the present invention,
The sensor unit 100, the preprocessing unit 10, the state prediction unit 20, and the operation amount determination unit 30 are included.

Here, the sensor unit 100 includes a timer 10
1, thermometer 102, hygrometer 103, toner concentration meter 10
4, copy / print number counter 107, image densitometer 1
08, measuring means such as photoconductor surface potential meter 109, and
It has operation panel information 105, document information 106, and the like.

In FIG. 10, the preprocessing unit 10 converts each measurement value and information fetched from the sensor unit 100 into a form that can be input to the state prediction unit 20 and the operation amount determination unit 30. Examples of this conversion method include analog / digital conversion (A / D conversion).

A fuzzy arithmetic unit is incorporated in the manipulated variable determining unit 30, and the fuzzy arithmetic unit performs fuzzy arithmetic operations to control the state predicting unit 20 and each unit of the image forming apparatus such as a copying machine. Generate an operation amount to be sent to the control unit of.

The state predicting section 20 has a current state, an operation amount,
A neural network that inputs the operation amount difference and outputs a state predicted when the operation amount is changed by the operation amount difference in the current state is incorporated, and the state is predicted.

Here, the neural network has already learned the characteristics of the system of the apparatus by using a parameter representing a state obtained in advance by experiments as a teacher value.

Next, the operation of the electrophotographic process control apparatus according to the present invention constructed as described above will be described with reference to FIG.

Each measurement value and information from the sensor unit 100 is processed (converted) into a form that can be input to the state prediction unit 20 and the operation amount determination unit 30 in the preprocessing unit 10 including an A / D converter. It

The parameters converted from the sensor unit 100 by the preprocessing unit 10 are the timer 101, the thermometer 102, the hygrometer 103, the toner densitometer 104, the panel information 105, the document information 106 (printer and digital type). (For copiers only), Copy / print counter 10
7 parameters of group A capable of real-time measurement (measurement without stopping the process), and image densitometer 108 incapable of real-time measurement because it is necessary to interrupt the process to generate a pattern or the like, and It is divided into the parameters of the B group of the photoconductor surface potential meter 109.

The parameters of the group A are constantly measured in real time, and the values of the parameters of the group B, which have been measured last time, are held and input to the manipulated variable determiner 30. The operation amount determining unit 30 calculates the operation amount to be sent to the control unit of each unit of the image forming apparatus such as a copying machine by using a predetermined membership function based on each parameter sent from the preprocessing unit 10.

On the other hand, as shown in FIG. 12, when it is necessary to change the manipulated variable due to various conditions (changes in the state of the device), (1). The manipulated variable determiner 30 issues a request to the central processing unit that controls the entire apparatus, stops the process, and causes the parameters of the group B to be measured.

(2). Next, the manipulated variable determiner 30 causes the state predictor 20 to send a manipulated variable difference and a state prediction request to predict the state.

(3). In the state prediction unit 20, each parameter from the preprocessing unit 10, the current operation amount and the operation amount difference parameter from the operation amount determination unit 30 are used by using a neural network in which the system of the device has been learned in advance by experiments or the like. Is used to predict the state of the device when the current operation amount is moved by the operation amount difference, and a state parameter is generated and sent to the operation amount determination unit 30.

(4). In the manipulated variable determiner 30, the target state is compared with the state parameter sent from the state predictor 20, and if the target state deviates, the state difference parameter is changed, Again (2),
The operation of (3) is performed.

(5). In this way, the above operation is performed until the predicted state converges to the target state.
At this time, a fuzzy operation is performed using a predetermined membership function using the manipulated variable difference and the change in the predicted state with respect to the manipulated variable difference and / or the parameter representing the state inside the device and the state outside the device. Make changes.

Then, a new manipulated variable is determined by using the manipulated variable difference when the predicted state converges to the target state as the change amount of the actual manipulated variable.

On the other hand, the neural network in the state predicting unit 20 performs the learning in the following procedure when it is necessary to re-learn due to various conditions (deviation of prediction, user's designation, etc.).

"1". When the learning operation amount difference in the current state is “0”, the predicted state must be equal to the current state regardless of the operation amount. Therefore, the current state, the manipulated variable, and the manipulated variable difference (= 0) are input, and the neural network is learned using the current state as a teacher value.

"2". Learning of preparatory state Learning of state prediction when the operation amount is changed by the operation amount difference is performed as follows. First, the state, the operation amount, and the operation amount difference are fixed as inputs. Next, each unit (charger grid, exposure device, etc.) of the apparatus is controlled by a new operation amount (= current operation amount + operation amount difference). And
The sensor unit 10 detects the internal state and the external state of the new device.
The neural network is learned by using the state at this time, which is taken in from 0 and processed by the preprocessing unit 10, as a teacher value.

[0068]

The electrophotographic process control device according to the present invention described above has the following operational effects.

(1) Finer control of each part of the apparatus becomes possible.

The electrophotographic process control apparatus according to the present invention enables finer control in accordance with each state of the image forming apparatus, as compared with the conventional method of determining the operation amount using the table reference. As a result, it is possible to expect a higher quality image than before, and it is possible to avoid the occurrence of abnormal images such as background stains, white spots in black solid portions, and insufficient density.

(2) The load on the cleaning portion can be reduced.

In the electrophotographic process control apparatus according to the present invention, the surface potential, the toner density and the like are measured only once for each control, so that the operation amount is determined while the repeated reference pattern is created. As a result, the number of times the developed reference pattern is cleaned is reduced, and the burden on the cleaning portion is reduced. This extends the life of the device and improves its reliability.

(3) It is possible to perform control in accordance with machine differences and changes with time of the photosensitive drum and the like.

In the electrophotographic process control apparatus according to the apparatus of the present invention, there are variations in the characteristics of the image forming apparatus due to machine differences, changes in the system that are difficult to predict due to changes over time, and the predicted state and the actual state. Even if the two differ from each other, it can be dealt with by re-learning the neural network at any time and incorporating the characteristic change of the system into the state prediction unit. As a result, the control device can be made to follow changes over time, and optimum control can always be performed.

(4) The process speed of the apparatus itself is improved.

In the conventional algorithm such as PID control,
It was difficult to converge the prediction state of the state prediction unit to the target value by performing a plurality of operation amount difference operations that depend on each other in a complicated manner.
In the method of the present invention, since the state difference is created by using the fuzzy calculation, it is possible to expect a faster convergence of the predicted state and improve the process speed of the device itself.

(5) The development period and cost can be reduced.

In the electrophotographic process control apparatus according to the present invention, the generalization ability of the neural network
The state predictor can be constructed with fewer experiments.
In other words, the state of the image forming device due to environmental factors such as temperature and humidity,
And, it means that the state prediction function can be realized from a smaller number of combinations of data regarding parameters such as the operation amount to each element drive unit of the device. This is to realize the same function by a table reference method or the like. In this case, the development period and the cost are greatly reduced as compared with the case where a huge amount of tables must be held in the apparatus in addition to the huge experiment for every combination of each parameter.

[Brief description of drawings]

FIG. 1 is a block diagram showing a basic configuration of an electrophotographic process control apparatus according to the first and third aspects of the present invention.

FIG. 2 is an explanatory diagram of a control operation of the electrophotographic process control apparatus according to the first and third aspects of the invention.

FIG. 3 is an explanatory diagram of a control operation of the electrophotographic process control apparatus according to the first and third aspects of the invention.

FIG. 4 is an explanatory diagram of a control operation of the electrophotographic process control device according to the first and third aspects of the invention.

FIG. 5 is an explanatory diagram of a learning operation of the electrophotographic process control apparatus according to the first and third aspects of the invention.

FIG. 6 is an explanatory diagram of a learning operation of the electrophotographic process control device according to the first and third aspects of the invention.

FIG. 7 is an example of a surface potential state predicting unit of a charging unit configured by using a three-layer hierarchical neural network of the electrophotographic process control apparatus according to the first and third aspects of the invention.

8 is a flow chart for determining an operation amount of the electrophotographic process control apparatus of FIG.

FIG. 9 is a block diagram showing a basic configuration of an electrophotographic process control apparatus according to the inventions of claims 2 and 4.

FIG. 10 is a block diagram showing a basic configuration of an electrophotographic process control apparatus according to the invention of claim 5.

FIG. 11 is a block diagram for explaining the operation of the electrophotographic process control apparatus according to the fifth aspect of the invention.

FIG. 12 is a block diagram for explaining another operation of the electrophotographic process control apparatus according to the invention of claim 5;

[Explanation of symbols]

 1, 10 Pre-processing unit 2, 20 State prediction unit 3, 30 Manipulation amount determination unit 4 State comparison unit 5 Fuzzy calculation unit 100 Sensor unit 101 Timer 102 Thermometer 103 Hygrometer 104 Toner density meter 105 Operation panel information 106 Original information 107 Copying / printing counter 108 Image densitometer 109 Photoconductor surface potential meter

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁵ Identification code Internal reference number FI technical display location G06G 7/60 7368-5B

Claims (5)

[Claims] 1. An image forming process control apparatus for determining an optimum operation amount for image forming, comprising: a measuring section for measuring the state of each part in the image forming apparatus; and information obtained from the measuring section. It is composed of a pre-processing unit for converting into a parameter representing a predetermined state, and a neural network which has been subjected to predetermined learning in advance, and inputs information on the above parameters and manipulated variables into its input layer, and from the output layer to the above Based on the comparison result of the state prediction unit that outputs the predicted state of the specific portion in the image forming apparatus, the state comparison unit that compares the output of the state prediction unit and the target state of the specific portion, and the comparison result of the state comparison unit. Generates a difference amount of the operation amount from the current operation amount and sends it to the state prediction unit, and when the comparison result converges, the difference amount is added to the current operation value as the optimum operation amount. Electrophotographic process control device characterized by having an operation amount determination unit constant to. 2. The electrophotographic process control apparatus according to claim 1, further comprising a fuzzy operation section for performing fuzzy operation, wherein the operation amount of each portion in the image forming apparatus is determined by performing the fuzzy operation. Characteristic electrophotographic process control device. 3. In an image forming apparatus of an image forming apparatus such as a copying machine, a printer and a facsimile having an electrophotographic process mechanism, the surface potential of a photoconductor as an electrostatic latent image carrier, the toner adhesion amount,
Measuring means for measuring the internal / external state of the image forming apparatus such as temperature and humidity by a sensor, etc., and the internal / external state of the image forming apparatus and / or the potential of the charger or grid obtained by the measuring means. Preprocessing means for converting the operation amount information inside the image forming device such as the charge amount and the exposure amount into a parameter representing the state of the image forming device, and a parameter representing the state of the image forming device are input, and experiments etc. are performed in advance. State predicting means for estimating the state of the image forming device using a neural network in which the characteristics of the system of the image forming device have been learned using the parameter representing the state of the image forming device obtained as a teaching value. And a state comparing means for comparing the state of the image forming apparatus estimated by the state predicting means with the state of the image forming apparatus targeted for control, and output by the state comparing means. An electrophotographic process comprising: an operation amount determining means for calculating an operation amount of each part of the image forming device from the comparison result and determining an optimum operation amount of each part of the image forming device. Control device. 4. The electrophotographic process control apparatus according to claim 3, further comprising fuzzy operation means for performing fuzzy operation, wherein the operation amount of each part of the image forming apparatus is determined by performing fuzzy operation. And an electrophotographic process control device. 5. In an image forming apparatus having an electrophotographic process mechanism, information obtained by a measuring means for measuring the surface potential, image density, temperature and humidity of the photosensitive drum, and the external state of the apparatus is used. A pre-processing means for converting into a parameter representing the state, a state predicting means using a neural network in which the characteristics of the system of the apparatus are learned by using a parameter representing the state obtained by an experiment or the like in advance as a teacher value, A fuzzy calculation is performed by using the prediction result of the prediction means and / or the parameter indicating the internal state of the apparatus and the external state of the apparatus and / or the information from the operation panel of the apparatus and / or the current operation amount, and the charger of the apparatus. And operation amount determining means for determining the operation amount to be sent to each control unit for controlling each unit of the apparatus such as voltage and current of the exposure device Electrophotographic process control apparatus characterized by comprising a.