JPH04297723A - Foodstuff heating device - Google Patents

Abstract

PURPOSE:To provide a microwave oven capable of realizing the substantially non-staged finish of the cooking of foodstuffs without having multiple cooking sequences as a memory. CONSTITUTION:A neural network for coefficient recognition is formed upon learning based on parameters corresponding to cooking sequences and on a coefficient at each stage of cooking sequences. When different parameters are then inputted, cooking sequences are formed by the aforementioned neural network on the basis of both a neuro-unit 1 for outputting the recognition of a coefficient at each stage and the coefficient outputted from the neuro-unit 1. The title foodstuff heating device comprises a sequence control section 2 for controlling cooking operation on the basis of the cooking sequences thus formed.

Description

[Detailed description of the invention]

0001

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a cooking sequence structure for a microwave oven or the like.

0002

[Employee skills] The cooking sequence in conventional microwave ovens only has a few adjustment functions such as standard, strong, and weak, and even those with correction functions simply change the cooking time and microwave output. It belonged to

0003

[Problems to be Solved by the Invention] However, with such conventional methods, it is not possible to perform detailed cooking that matches the user's preferences. One possible way to solve this problem is to provide a large number of cooking sequences, but there is a limit to the memory capacity.

[0004] The present invention was made in view of the above-mentioned circumstances, and an object thereof is to provide a microwave oven or the like that can achieve virtually stepless cooking without having a large number of cooking sequences in its memory. shall be.

[0005]

[Means for Solving the Problems] In the present invention, a food heating device performs learning based on parameters corresponding to a cooking sequence and coefficients of each stage of the cooking sequence to form a neural network for coefficient recognition. Next, when different parameters are input, the neuro unit recognizes and outputs the coefficients of each stage using the neural network, creates a cooking sequence based on the coefficients from the neuro unit, and cooks based on the cooking sequence. The configuration includes a sequence control section that controls the operation.

[0006]

According to the present invention, a cooking operation is performed based on a stepless cooking sequence created based on coefficients from a neuro unit.

[0007]

[Embodiment] Fig. 1 is a block diagram of a microwave oven as a food heating device of the present invention. 1 is a block diagram of a microwave oven as a food heating device of the present invention. 1 is a block diagram of a microwave oven in which learning is performed based on a parameter (α) corresponding to a cooking sequence and coefficients of each stage of the cooking sequence. Neuro unit 2 forms a neural network for coefficient recognition, and then recognizes and outputs the coefficients of each stage by the neural network when different parameters are input. Neuro unit 2 executes a cooking sequence based on the coefficients from neuro unit 1. 3 is a heater, 4 is a magnetron, and the heater 3 and magnetron 4 are connected to a heating chamber 5.
The heater 3 is provided with a cooking operation control signal from the sequence control unit 2 via a heater drive circuit 6, and the magnetron 4 is provided with a cooking operation control signal via a magnetron drive circuit 7.

Reference numeral 8 denotes a memory for storing parameters corresponding to the cooking sequence, and stores parameters ranging from 0.0 to 1.0 for the finished state of cooking, ranging from weak to strong. Reference numeral 10 denotes a correction unit for the parameters; 11 is a fuzzy unit that fuzzy infers the correction amount from the variation amount of the pick; and 12 is an operation unit that performs the cooking start operation and provides the fuzzy unit 11 with data of the variation amount of the pick.

[0009] The neuro unit 1 forms a neural network for coefficient recognition based on the parameters of standard, weak, and strong cooking states and the coefficients of each stage at that time, and the neural network is connected to the input layer. It consists of three layers: , intermediate layer, and output layer, and the learning algorithm is backpropagation.

The operation of the microwave oven will be explained below with reference to the flowchart shown in FIG.

First, when a cooking start operation is performed on the operation unit 12, a cooking course is selected according to the operation contents, and the corresponding parameters are called up (
Steps 1, 2, 3). Then, in neuro unit 1, the coefficient (
For example, heater 3 and magnetron 4) are recognized and output to complete the sequence, and operation control signals are given to each of heater 3 and magnetron 4, and heating cooking is performed based on the sequence (steps 4 and 5).
.

[0012] When the cooking is finished, if the finished state does not intuitively match the selected parameters, turn the knob on the operating section 12 to obtain the parameter correction amount, and add the correction amount. The modified parameters are stored in the memory 8 and the operation ends (steps 6, 7, 8).
, 9, 10, 11).

[0013] When obtaining the correction amount of the above parameters,
Since the relationship between the sense of adjustment and the amount of variation in the knob is ambiguous, a fuzzy rule is set: [If the amount of variation in the knob is negatively large (N1), the amount of parameter correction is negatively large (N2)]
). ] [If the amount of variation in the knob is exactly large (P1),
Make the parameter correction amount exactly large (P2). ], and by using the membership of the antecedent part and the consequent part shown in FIGS. 3 and 4, by fuzzy control,
Calculate the amount of correction using the indirect method.

By modifying the parameters in this manner, a sequence that satisfies the user can gradually be obtained. In the above example, a neural network was formed by learning using data when the finishing condition was standard, strong, and weak, but a neural network was formed by performing more reliable learning using more data. It is also possible to do so.

[0015]

[Effects of the Invention] With the above-described configuration, the cooking operation is performed based on a stepless cooking sequence created based on the coefficients from the neuro unit.
As a result, it has become possible to provide a microwave oven or the like that can achieve virtually stepless cooking without having a large number of cooking sequences in its memory.

[Brief explanation of drawings]

FIG. 1 is a block diagram of a microwave oven according to the present invention.

FIG. 2 is a flowchart for explaining operation.

FIG. 3 is an antecedent membership function diagram.

FIG. 4 is a consequent membership function diagram.

[Explanation of symbols]

1 Neuro unit 2 Sequence control section

Claims (1)

[Claims] 1. A neural network for coefficient recognition is formed by performing learning based on parameters corresponding to a cooking sequence and coefficients of each stage of the cooking sequence, and then when different parameters are input, the neural network a neuro unit that recognizes and outputs the coefficients of each stage, and a sequence control unit that creates a cooking sequence based on the coefficients from the neuro unit and controls the cooking operation based on the cooking sequence. Food heating equipment.