JPH04284465A - Electrophotographic process controller - Google Patents

Abstract

PURPOSE:To offer a process controller for controlling the respective parts of an electrophotographic process based on information inputted from many sensors set at the respective parts in an image forming device. CONSTITUTION:Measuring instruments such as a voltmeter, an ammeter, a temperature sensor, a humidity sensor and a toner quantity meter, etc., are respectively provided in the respective parts of electrostatic charging/image exposing/developing/transfer/fixing in the image forming device 1. Then, a control value decision part 6 is provided, which decides a control value for a process control part 4 by fetching sensor output information from the respective measuring instruments of a sensor part 2 while changing the state of the process, performing the learning of neural net in a state where the control value for the process control part 4 previously obtained by experimentation is set as an educator value, and inputting the sensor output information outputted from the respective measuring instruments of the sensor part 2 after the learning. The relation of a sensor output pattern to a control parameter is previously learned in a learning mode in the control value decision part 6, so that optimum process control is performed in real time in an evaluation mode.

Description

[Detailed description of the invention]

0001

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

0002

2. Description of the Related Art In an image forming apparatus having an electrophotographic process mechanism such as a copying machine, a conventional control method for the electrophotographic process is based on a table (1) that can refer to appropriate control values for the output of a surface electrometer or a photosensor. It has been common practice to create a table (table) in advance experimentally, and to control each parameter by referring to this table during actual control.

0003

[Problem to be Solved by the Invention] However, the above method requires a huge amount of experimentation to create the table.
Another problem is that the amount of data required for control is large. In addition, since the well-known control method such as the PID method has a simple device configuration and easy adjustment of control parameters,
used for all sorts of purposes. However, it is difficult to apply this control method to complex controlled objects or systems whose dynamic characteristics frequently change. The present invention has been made in view of the above circumstances, and is based on information input from a large number of sensors installed in various parts inside the image forming apparatus in an image forming apparatus having an electrophotographic process mechanism such as a copying machine. Another object of the present invention is to provide an electrophotographic process control device that controls each part of an electrophotographic process.

0004

[Means for Solving the Problems] In order to achieve the above object, an electrophotographic process control device according to claim 1 of the present application is a device for controlling each part of an electrophotographic process in an image forming apparatus having an electrophotographic process mechanism. The image forming apparatus has measuring instruments such as voltmeters, ammeters, temperature sensors, humidity sensors, and toner amount meters in charging, image exposure, development, transfer, fixing, etc., to change the state of the process. The neural network is trained by taking in sensor output information from each measuring device and using the control values for the process control unit obtained through experiments as teacher values. After learning, the sensor output information output from each measuring device is The method is characterized in that it has a control value determination unit that determines a control value to be input to the process control unit.

Further, the electrophotographic process control device according to claim 2 of the present application is a device for controlling each part of the electrophotographic process in an image forming apparatus having an electrophotographic process mechanism, and which controls charging/image exposure/ Each developing/transfer/fixing area has measuring instruments such as a voltmeter, ammeter, temperature sensor, humidity sensor, toner amount meter, etc., and sensor output information is captured from each measuring instrument while changing the process status. Differentiation/integration/normalization/frequency analysis/
It has a pre-processing unit that generates a characteristic pattern from the sensor output information pattern by performing processing such as noise removal,
This feature pattern is used as input, and the neural network is trained using control values for the process control unit obtained in advance through experiments as teacher values. After learning, the sensor output information output from each measuring instrument is used as input, and the The present invention is characterized by having a control value determining section that determines a control value to be applied to the control section.

[0006]

[Operation] In the electrophotographic process control device of the present invention, the control value determination section learns the relationship between the sensor output pattern and the control parameters in advance in the learning mode, so that the optimal process can be determined in real time in the evaluation mode. Control becomes possible.

[0007]

Embodiments Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. FIG. 1 is a diagram showing an example of the configuration of an electrophotographic process control device according to the present invention, FIG. 2 is a diagram showing an example of the configuration of a copying machine in which the present invention is implemented, and FIG. 3 is a diagram showing an example of the configuration of a control value determining section. It is. The process control device having the configuration shown in FIG. 1 includes an electrophotographic processing unit 3 that reads an image and reproduces the input image on plain paper via a photosensitive drum, and a sensor that captures various status values in the electrophotographic process of the copying machine 1. Section 2, grid voltage and laser diode (LD)
The configuration includes a control unit 4 that controls power, etc., and a control value determination unit 6 that determines a control value to be input to the control unit 4. The control value determination unit 6 controls the process of the copying machine 1. Sensor output information is captured from each measuring device of the sensor unit 2 while changing the state, and the neural network is trained using the control value for the process control unit 4 obtained through experiments etc. as a teacher value, and after learning, each measurement The sensor output information output from the device is input, and the control value to be sent to the process control unit is determined. Further, when the preprocessing unit 5 is provided as shown in the figure, the preprocessing unit 5 takes in sensor output information from each measuring device of the sensor unit 2 while changing the process state of the copying machine 1, and differentiates this information. /integral/
A feature pattern is generated from the sensor output information pattern by performing processing such as normalization, frequency analysis, and noise removal. Then, the control value determination unit 6 inputs this characteristic pattern and uses the process control unit 4 obtained in advance through experiments etc.
The neural network is trained using the control value for the process control unit as a teacher value, and after learning, the sensor output information output from each measuring device is used as input to determine the control value to be sent to the process control unit.

Next, the copying machine having the configuration shown in FIG.
A photosensitive drum 11 and a charging grid 12 for performing an electrophotographic process of image exposure/development/transfer/fixing;
A grid voltage control section 27 that controls the voltage applied to the grid, a laser diode (LD) 13, a developing device 14, a toner control section 19 that controls the amount of toner supplied to the developing device 14, and a developing bias voltage. a development bias control section 15 that generates a transfer current, a transfer grid 16, a transfer current control section 21 that controls a transfer current, and a separation grid 17.
and a separation current control section 22 that controls the separation current,
In addition, a surface electrometer 18 that detects the surface potential and current of the photosensitive drum 11, a drum current sensor 23, a toner concentration sensor 20 that detects the toner concentration in the developing device 14, a photo diode (P sensor) 24, and a temperature sensor. 25, a sensor section consisting of a group of measuring instruments such as a humidity sensor 26. Note that each control section (controlled object) corresponds to the control section 4 in FIG. Furthermore, illustrations of a fixing device, a cleaning device, etc. are omitted.

Next, the control value determination unit shown in FIG. 3 inputs the characteristic parameters extracted by the preprocessing unit 5 shown in FIG. 1, and outputs control values for process control of each control unit 4. It is a neural network. This neural
The network has a hierarchical structure consisting of an input layer, a middle layer, and an output layer, and the units are connected from the input layer to the output layer, and there are no connections within each layer. Therefore, the input pattern is propagated from the input layer to the output layer. On the other hand, learning proceeds from the output layer to the input layer. Note that learning means changing the strength of the connections so as to reduce the difference between the actual output and the desired output value. For example, this network is trained as follows. When input data is given to each unit in the input layer, this signal is converted in each unit, transmitted to the intermediate layer, and finally comes out from the output layer. Therefore, the output value and
Compare the desired output values and change the coupling strength to reduce the difference. Learning is then repeated until the output value from the output layer converges to a desired output value.

Now, the internal configuration of the copying machine is shown in FIG. 2, and the components such as the grids 12, 16, 17 and the photosensitive drum 11 are different from each other depending on the number of copies, temperature and humidity, toner concentration, and developer fatigue. As a result, the characteristics change or deteriorate due to factors such as background stains, density unevenness, etc., and stable output image quality cannot be obtained. The electrophotographic process control device according to the present invention selects appropriate control values for each control unit based on output information from various sensors installed inside the copier in a copier that fluctuates due to the influence of internal and external environments. , which aims to stabilize the output image, and has a <learning mode> and <evaluation mode> depending on the operation mode of the control value determination unit 6 shown in FIGS. 1 and 3. These operating modes will be explained below.

<Learning mode of control value determining unit> As described above, learning is performed in the neural network of the control value determining unit 6, and learning is performed using the input image and appropriate control value results of grid voltage/current as teacher values. . The learning method in the control value determining unit 6 includes, for example, the document ``Using neural nets for pattern recognition, signal processing, and knowledge processing.''
, Nikkei Electronics 1987.8.10 (No.4
27) PP. 115-124" may be used. As input values during learning, values obtained by converting sensor output information output from the sensor section 2 in the copying machine 1 into feature parameters in the preprocessing section 5 are used. Here, the preprocessing section 5 is a section that performs processing to convert sensor output information from the sensor section into input information for the neural network, such as changes (differential values) and integrals with respect to the time axis. Performs processing such as determining values and frequency components, normalizing numerical values, and applying noise removal and correction functions.

<Process control after learning (evaluation mode)>
Control value determination is learned so that a teacher value given in advance during learning is output for every output pattern of each sensor. Therefore, by inputting characteristic parameters from the copying machine to the control value determining section 6 that has completed learning, each control section can be controlled using the optimum control value at that time. Here, the quality estimated from the feature parameters includes quality related to abnormal images such as the above-mentioned background stains, uneven density, and decreased density. As described above, by learning the relationship between sensor output patterns and control parameters in advance in the learning mode, it is possible to perform optimal process control in real time in the evaluation mode.

[0013]

[Effects of the Invention] As explained above, according to the present invention,
An electrophotographic process control device that controls each part of the electrophotographic process based on information input from a large number of sensors installed in each part inside the image forming apparatus in an image forming apparatus such as a copying machine that has an electrophotographic process mechanism. can be provided. In the electrophotographic process control device of the present invention, the control value determination unit previously learns the relationship between the sensor output pattern and the control parameter in the learning mode, so that optimal process control can be performed in real time in the evaluation mode. It becomes possible. Furthermore, in the electrophotographic process control apparatus of the present invention, 1. The image quality of the output image of an image forming device such as a copying machine can be estimated from the internal state of the image forming device. 2. Real-time process control is also possible for electrophotographic process conditions that change over time. 3. Through learning, it is possible to adapt to process control of different types of image forming devices. 4. The development period and cost of image forming devices such as copying machines can be reduced. The following effects can be obtained.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing an example of the configuration of an electrophotographic process control apparatus according to the present invention.

FIG. 2 is a schematic configuration diagram showing an example of the internal configuration of a copying machine in which the present invention is implemented.

FIG. 3 is a diagram showing a configuration example of a neural network of a control value determining section.

[Explanation of symbols]

1 Copy machine 2 Sensor part 3 Electrophotographic process department 4 Control section 5 Pre-processing section 6 Control value determination section 11 Photoreceptor drum 12 Charging grid 13 Laser diode 14 Developing device 15 Developing bias control section 16 Transfer grid 17 Separation grid 18 Surface electrometer 19 Toner control section 20 Toner concentration sensor 21 Transfer current control section 22 Separation current control section 23 Drum current sensor 24 P sensor 25 Temperature sensor 26 Humidity sensor 27 Grid voltage control section

Claims (2)

[Claims] 1. A device for controlling each part of the electrophotographic process in an image forming apparatus having an electrophotographic process mechanism, wherein a voltmeter is installed at each of charging/image exposure/developing/transfer/fixing parts of the image forming apparatus. It has measuring instruments such as an ammeter, temperature sensor, humidity sensor, toner amount meter, etc., and captures sensor output information from each instrument while changing the process state, and generates control values for the process control unit obtained through experiments etc. in advance. The neural network is trained by using the as a teacher value, and after learning, it is characterized by having a control value determination unit that receives sensor output information output from each measuring device as input and determines a control value to be sent to the process control unit. Electrophotographic process control equipment. 2. A device for controlling each part of the electrophotographic process in an image forming apparatus having an electrophotographic process mechanism, wherein a voltmeter is provided at each of charging/image exposure/developing/transfer/fixing parts of the image forming apparatus. It has measuring instruments such as ammeter, temperature sensor, humidity sensor, toner amount meter, etc. It captures sensor output information from each measuring instrument while changing the process state, and performs differentiation/integration/normalization/frequency analysis on this information. / It has a pre-processing unit that generates a feature pattern from the sensor output information pattern by performing processing such as noise removal, and uses this feature pattern as input and controls the control value for the process control unit obtained in advance through experiments etc. An electronic photograph characterized in that it has a control value determination unit that performs neural network learning as a value, and after learning, inputs sensor output information output from each measuring device and determines a control value to be sent to a process control unit. Process control equipment.