Heater controlling unit using a fuzzy neural network

• US 5,683,605 A
• Filed: 12/14/1995
• Issued: 11/04/1997
• Est. Priority Date: 12/27/1994
• Status: Expired due to Term

• Current Assignee: Sharp Kabushiki Kaisha (Hon Hai Precision Industry Co., Ltd.)
• Original Assignee: Sharp Kabushiki Kaisha (Hon Hai Precision Industry Co., Ltd.)
• Inventors: Matsuoka, Teruhiko
• Primary Examiner(s): Paschall, Mark H.
• Application Number:
  US08/572,079
• Time in Patent Office: 691 Days
• Field of Search: 219/411-413, 219/486, 219/492, 219/497, 219/501, 219/506, 219/505, 219/508, 340/588, 340/589, 374/102, 374/107, 395/3, 395/900
• US Class Current: 219/497
• CPC Class Codes: G03G 15/2003 using heat
• CPC Class Codes: G03G 2215/00118 using fuzzy logic
• CPC Class Codes: G05D 23/1917 using digital means

Abstract

A heater controlling unit used in a heat fusing device of an image forming apparatus, such as a copying machine, comprises a temperature detecting unit for outputting a detected temperature, a temperature change rate computing unit for computing the surface temperature of a heat fusing roller using the output of the temperature detecting unit, and an on-time computing and controlling unit for computing a heater on-time within a predetermined period by means of a fuzzy neural network using the above temperature and a temperature change rate found by the temperature-change rate computing unit, a heater controlling circuit for controlling the on/off action of the heater, a predictive computing unit for predicting the surface temperature, and a comparing and adjusting unit for comparing the actual temperature with the predicted temperature, and, when the difference is greater than a predetermined value (e.g., ±5° C.), for adjusting the weights of the links within the network. Accordingly, the roughly-set parameters can be amended to the ones for an optimal on-time by the sequential learning. As a result, the program can be generated in a simpler manner, and the program can be changed easily for individual units depending on aged distortion and environments thereof.

Claims

1. A heater controlling unit for controlling an on/off action of a heater to control a surface temperature of heat releasing means which releases heat generated by said heater through conduction, said heater controlling unit comprising:

temperature detecting means for detecting an actual temperature of a surface of said heat releasing means;

temperature change rate computing means for computing a temperature change rate within a predetermined period using the actual temperature detected by said temperature detecting means; and

on-time computing and control means including a fuzzy neural network, said fuzzy neural network receiving the actual temperature detected by said temperature detecting means and the temperature change rate computed by said temperature change rate computing means and computing a period of time during which said heater is turned on using a fuzzy logic to output a computed value, said on-time computing and control means controlling said on/off action based on the output computed value from said fuzzy neural network,

said fuzzy neural network including a plurality of layers each having a node for outputting an output signal in response to an input signal, and a plurality of links that interlink the node of said each layer to transmit a signal among layers,

a weight of said each link, which indicates signal transmission efficiency among the nodes, being adjusted based on a difference between a target value and the output computed value from said fuzzy neural network.

2. The heater controlling unit as defined in claim 1 further comprising:

predictive computing means for, when said heater is turned on by said on-time computing and control means, computing a predictive temperature of the surface of said heat releasing means for a next time said heater is turned on, said predictive temperature being computed using the actual temperature detected by said temperature detecting means and the temperature change rate computed by said temperature-change rate computing means; and

comparing and adjusting means for computing a difference between the actual temperature and the predictive temperature, and, when said difference is greater than a predetermined value, said comparing and adjusting means adjusting the weight of said each link within said fuzzy neural network in such a manner to minimize a difference between said target value and said period of time during which said heater is turned on computed by said on-time computing and control means.

3. The heater controlling unit as defined in claim 2, wherein said fuzzy neural network comprises an input layer, a membership input layer, a membership output layer, a rule layer, and an output layer, each of links among said layers having their respective weights.

4. The heater controlling unit as defined in claim 3, wherein:

each of said input layer and said membership input layer has a structure that divides each of the actual temperature and the temperature change rate into three areas; and

said rule layer outputs ANDs of all possible links of an output value from a node corresponding to each area of the actual temperature and an output of a node corresponding to each area of the temperature change rate in said membership layer to said output layer.

5. The heater controlling unit as defined in claim 3, wherein said comparing and adjusting means adjusts a weight of a link from said input layer to said membership input layer, a weight of a link from said membership input layer to said membership output layer, and a weight of a link from said rule layer to said output layer.

6. The heater controlling unit as defined in claim 5, wherein said input layer includes a first node for receiving the actual temperature, a second node for receiving the temperature change rate, and a third node for receiving a fixed value,

a weight of a link from said third node to said membership input layer being a center value of a membership function,

a weight of a link from said membership input layer to said membership output layer being a slope of the membership function.

7. The heater controlling unit as defined in claim 2, wherein:

said temperature change rate computing means computes a first temperature change rate in a period during which said heater is turned on, and a second temperature change rate in a period during which said heater is turned off; and

said predictive computing means computes the predictive temperature by adding up (1) a current actual temperature, (2) a product of a preceding first temperature change rate and a following period during which said heater is turned on, and (3) a product of a preceding second temperature change rate and a following period during which said heater is turned off.

8. The heater controlling unit as defined in claim 1, wherein said fuzzy neural network includes:

a rule layer having a plurality of nodes each for processing a signal based on their respective fuzzy rules; and

an output layer having a node for outputting said period of time during which said heater is turned on, the node of said output layer being connected to an output side of said each node of said rule layer.

9. The heater controlling unit as defined in claim 8, wherein said fuzzy neural network further includes:

an input layer into which the actual temperature and the temperature change rate are inputted; and

a membership input layer and a membership output layer, provided between said input layer and rule layer, for converting the actual temperature and the temperature change rate into a fuzzy variable using a membership function.

10. The heater controlling unit as defined in claim 1 further comprising:

predictive computing means for, when said heater is turned on by said on-time computing and control means, computing a predictive temperature of the surface of said heat releasing means for a next time said heater is turned on, said predictive temperature being computed using the actual temperature means and the temperature change rate computed by said temperature-change rate computing means; and

comparing and adjusting means for computing a difference between the actual temperature detected by said temperature detecting means and the predictive temperature computed by said predictive computing means, and, when the difference is greater than a predetermined value, for adjusting a weight of a link within said fuzzy neural network of said on-time computing and controlling means to minimize the difference.