JP2000211338A - Estimating method of cabin interior rear temperature and device of the same - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a method of obtaining an estimate value of an interior rear temperature which is exceedingly close to the actual interior rear temperature, by using a temperature detected by one cabin interior temperature detecting means, and a device of the same. SOLUTION: This is a cabin interior rear temperature estimating method in which an estimate value of an interior rear temperature is estimated by a neural network 12 receiving a measurement temperature from a front seat temperature sensor and an environmental factor of a vehicle and a state of a rear air conditioner as input signals, and outputting an estimate value of the cabin interior rear temperature as an output value. Weighting means 20 which weights so as to reduce a signal variation width in an input state to the neural network with respect to the detected temperature is disposed. A signal weighted by the weighting means 20 is inputted to the neural network 20, and the neural network learns.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an air conditioner for a vehicle for properly controlling the temperature in a passenger compartment to a target temperature, and more particularly to an air conditioner for a vehicle of a one-box twin type vehicle. The present invention relates to an apparatus for estimating a rear temperature of a vehicle interior and a method for estimating a rear temperature of the vehicle interior.

[0002]

BACKGROUND ART FIG. 6 is a diagram showing a control flow of a conventional vehicle air conditioner, air conditioning system, the target temperature T _z set by the temperature setting unit 4 is installed in the vehicle interior control logic 11 that the feedback value of the measured temperature T _i measured by the inside air temperature sensor 6 as the vehicle interior temperature detection means and
Thus, the temperature, air volume, and the like of the conditioned air sent into the vehicle interior are controlled. However, since the inside air temperature sensor 6 is usually installed below the front panel, the inside air temperature sensor 6 is affected by the outside air temperature, the insolation, and the like, so that the measurement temperature _{Ti is used.}
Is different from the temperature (vehicle interior temperature) T ₀ near the occupant's seating position. Therefore, in order to bring the feedback value T _i closer to the actual vehicle interior temperature T ₀ , the vehicle is provided with an outside air temperature sensor or a solar radiation sensor, and the measured temperature T _i is corrected based on the output of each sensor. Corrected the measured temperature T _i based on the state of the air-conditioning equipment such as the air-conditioning air blowing mode.

[0003] However, it was measured by the internal temperature sensor 6.
Measurement temperature T_iAnd the actual vehicle interior temperature T₀To compensate for the difference
For this reason, there are many parameters used in the above control logic.
The matching work takes time.
Was. Japanese Patent Application Laid-Open No. 6-195323 discloses FIG.
As shown, the target temperature T_z， Car interior temperature (here,
Temperature measured by air temperature sensor) T_i, Outside temperature T_a, Solar radiation T _sWhen
As an input signal, and for example, a new
Air supply of conditioned air using a neural network type additional learning device
Techniques for controlling the air volume and the like have been disclosed. This is
In the neural network, target outlet temperature, outlet
Learning and calculating each arithmetic expression such as mode state and blower air volume
It controls the final air flow, but the neural network
Teacher signal used to learn the input / output relationship of the work
Calculation time is long because there are many
There was such a problem. In particular, the front and rear
In each case, the target temperature T_z1, T_z2And air volume W_z1, W_z2
One-box twin type vehicle air conditioning that can be set
In the device, the condition of the conditioned air at the front and rear in the passenger compartment
Learning will be reduced because each will affect the other conditioned air.
It is difficult to bundle and it will increase the learning time
There was a problem. In addition, conventional one-box twin
For vehicle air conditioning systems of the type, the front and rear
Each unit has an internal temperature sensor,
In addition to being a backup, the internal temperature sensor for the rear
Harness is long because it is installed far from the knit
And the cost increases.
Was.

Accordingly, the present applicant has filed Japanese Patent Application No. Hei 10-169.
In No.017, the number of the vehicle interior temperature detecting means is one, and the temperature detected by the vehicle interior temperature detecting means is used to estimate the front and rear vehicle interior temperatures, which are the main elements of the control logic of the air conditioner, by a neural network. By controlling the temperature of the vehicle interior based on the estimated temperature values of the front part and the rear part, it is possible to perform control in consideration of the influence of the conditioned air in the front part and the rear part of the vehicle. The device is proposed. FIG. 8 is a diagram showing the configuration of the air conditioner for a vehicle, wherein 1 is a control device, 2A is an inside / outside air switching door 2a, a blower 2b, an evaporator 2c, a heater 2d, a mix door 2e, an upper outlet 2f, and a lower outlet. A front air duct having an opening 2g and an outlet switching door 2h, a rear air duct 2B having a configuration similar to that of the front air duct 2A, and a setting panel 3 for setting the temperature and air volume of the conditioned air. , 4a and 4b _denote target temperatures T _z1 and T at the front and rear of the vehicle compartment input from the setting panel 3.
A temperature setter for setting each of _Z2 and outputting to the controller 1;
Reference numerals 5a and 5b denote a plurality of driving devices for adjusting the degree of opening of the inside / outside air switching door 2a and the mixing door 2e of each of the air ducts 2A and 2B, and reference numeral 6 denotes a temperature detecting means for the front part of the vehicle compartment installed below the front panel. An internal air temperature sensor (hereinafter referred to as a front seat temperature sensor), 7 is an external air temperature sensor installed near a bumper of the vehicle, and 8 is a solar radiation sensor installed on an upper part of a front panel. F1 is the front seat, R1
R1 in the first row indicates a rear seat in the second row.

[0005] The control device 1 controls the opening degree of the front or rear mixing door 2e in accordance with environmental factors such as the outside air temperature and the states (control factors) of the front and rear air conditioners such as the conditioned air blowing mode. a seat control logic 11a and the rear-seat control logic 11b prior to output a control signal for controlling the like to the driving device 5a or the drive device 5b, from the measured temperature T _i and the control logic 11a, 11b from the front seat temperature sensor 6 Information on environmental factors and air conditioners is used as input signals,
A front-seat temperature estimator 12a and a rear-seat temperature estimator 12b which are formed by neural networks and output the estimated values T _N1 and T _N2 of the vehicle interior temperature to be output to the control logics 11a and 11b, respectively. I have. The control factors include, for example, outputs from the outside air temperature sensor 7 and the solar radiation sensor 8, the degree of opening and closing of the inside / outside air switching door 2a, the mix door 2e, and the outlet switching door 2h, and the drive voltage of the blower 2b.

FIG. 9A is a diagram showing the configuration of a neural network (Inc.s model for front) in the front seat temperature estimator 12a. The neural network for front 12F includes an input layer, a hidden layer, and an output layer. The three-layer hierarchical network includes a rear neural network 12 in a rear seat temperature estimator 12b shown in FIG.
R has the same layer structure as the front neural network 12F. The input signal of the front neural network 12F is a front seat temperature sensor (In).
c. s) The measured temperature T _i (° C.) from 6 and the outside air temperature sensor 7
Outside temperature T _a (° C.), solar radiation T _s from the solar radiation sensor 8
(Kcal / m ² hour), the blow mode set on the setting panel 3, the opening degree (%) of the front mix door 2e, and the front blower motor duty ratio (%) corresponding to the drive voltage of the front blower 2b, and output. The value is an estimated value T _N1 (° C.) of the temperature in the front part of the vehicle compartment. The input signal of the rear neural network 12R is the same as the input signal of the front neural network 12F.
In the input signal of 2F, the control factor of the front air conditioner is replaced by the control factor of the rear air conditioner, and the output value is an estimated value T _N2 (° C.) of the rear temperature of the vehicle compartment. However, since no temperature detecting means is provided at the rear part of the vehicle interior, the measured temperature T _i from the front seat temperature sensor 6 is used as the measured temperature.
(° C.) is used as an input signal.

[0007] The neural networks 12F, 12F
In R, a sigmoid function as shown in the following equation (1) is used.
It is. y_j= 1 / (1 + exp (-| x_j|) ‥‥‥ (1) where x_jIs the input signal v to layer i_iWeight w_ij
(X)_j= Σ
v_i・ W_ij-B_j), Y_jIs the output signal from layer i (layer j
Input signal). Note that this sigmoid function is
Outputs 0 to 1 when the input variable is -∞ to + ∞
You. Here, the temperature estimator 12a at the front is a neural network.
The front seat is used as a teacher signal when learning the network 12F.
(Driver's seat and passenger seat) Each head and foot of F1
Using the average temperature of the four points at the corresponding position,
Estimated temperature T_N1For each input signal
Weight w_ijAnd bias b_jFind the value of by learning,
At the time of control, the temperature at the front of the vehicle
Degree estimate T _N1Is output to the control logic 11a. one
On the other hand, the rear temperature estimator 12b outputs
Right, center and left heads and legs of seats R1 and R2
Using the average temperature of 12 points at the position corresponding to the
Of the temperature T at_N2Each input signal of the operation expression
Weight w_ijAnd bias b_jLearning the value of
At the time of control, the above
Front temperature estimate T_N2Is output to the control logic 11b.
You. Each input state is input to the neural network.
The force signal changes the minimum value to the maximum value of the measurement data from 0 to 1.
Normalized, weight w for each of the above input signal types_ij
To be evaluated the same.

FIG. 10 is a control flow chart of the vehicle air conditioner.
In the figure, the front seat control logic 11 a
Control factor and estimated value T from front seat temperature estimator 12a_N1And
The temperature T at the front of the passenger compartment₁Is set by the temperature setting device 4a.
Target temperature T_Z1Controls the front air duct 2A so that
I do. Similarly, the rear seat control logic 11 b
Control factor and estimated value T from rear seat temperature estimator 12b_N2And
The temperature T at the rear of the cabin_TwoIs set by the temperature setting device 12b.
Target temperature T_Z2Control the rear air duct 2B so that
Control. That is, front seat control logic 11a, rear seat control
Logic 11b indicates the temperature at the front or rear of the vehicle cabin.
T₁, T_TwoTemperature estimator 12a, 1
Estimate of the temperature in the front or rear of the cabin from 2b
T_N1, T_N2Control with the feedback value as
Accurate and quick target temperature at front and rear of vehicle compartment
T _z1, T_z2Can be

[0009]

However, in a one-box car, since the volume at the rear of the passenger compartment is generally larger than the volume at the front of the passenger compartment, the influence of the temperature at the front of the passenger compartment on the temperature at the rear of the passenger compartment is not so large. Not big. Therefore, the influence of the temperature detected by the front seat temperature sensor 6 inputted to the neural network 12R on the rear seat temperature estimator 12b is too strong, and the estimated value of the estimated rear compartment temperature is changed to the actual vehicle temperature. It did not fully reflect the temperature in the rear of the room.

The present invention has been made in view of the conventional problems, and estimates the rear compartment temperature which is very close to the actual rear compartment temperature by using the detected temperature from one compartment temperature detecting means. It is an object of the present invention to provide a method and an apparatus for obtaining a value.

[0011]

According to a first aspect of the present invention, there is provided a method of estimating a rear temperature of a vehicle interior, comprising: detecting an estimated value of the rear temperature of the vehicle interior by detecting one temperature of the vehicle interior; When the temperature detected by the means and the environmental factors of the vehicle and the state of the air conditioner at the rear are input signals, and the estimated value of the rear temperature of the vehicle interior is estimated by a neural network, the neural network of the detected temperature is used. It is characterized in that weighting is performed so as to reduce the signal fluctuation width of the input state to the network, and learning of the neural network is performed.

[0012] According to a second aspect of the present invention, there is provided a device for measuring the temperature in the rear of the vehicle compartment, wherein the temperature detected by one vehicle interior temperature detecting means for detecting the temperature in the vehicle compartment, the environmental factors of the vehicle and the state of the air conditioning equipment in the rear portion are measured. For each of the input signals, an estimating device for the rear temperature of the vehicle interior formed of a neural network having an output value of the estimated value of the rear temperature of the interior of the vehicle, the signal fluctuation width of the input state to the neural network with respect to the detected temperature. Weighting means for performing weighting to reduce the size is added,
The signal weighted by the weighting means is input to the neural network to perform learning of the neural network.

According to a third aspect of the present invention, the input state of the detected temperature from the vehicle interior temperature detecting means to the neural network is normalized from the minimum value to the maximum value of the measured data from 0 to 1. Then, weighting is performed so that the width from the minimum value to the maximum value of the measurement data is smaller than 1 while maintaining the center value at 0.5.

According to a fourth aspect of the present invention, in the apparatus for estimating a rear temperature of a vehicle interior, the normalized detected temperature x is weighted by the following equation. y = K _w (x − ／) + ／ where K _w is a weight amount and y is a weighted detected temperature.

[0015] estimating device in the vehicle interior rear temperature according to claim 5, fuzzy inference for fuzzy inference the weight amount K _w on the basis of a deviation between the actual room temperature estimate of the vehicle interior rear temperature and vehicle interior rear Temperature The weight is estimated by means.

A device for estimating the rear temperature of a vehicle compartment according to a sixth aspect of the present invention uses a deviation between an estimated value of the rear temperature of the vehicle compartment and an actual room temperature of the rear temperature of the vehicle compartment as an input signal, and uses the weight _Kw as an output value. by the weight amount estimating means constituted by a neural network which is obtained so as to estimate the weight amount K _w.

According to a seventh aspect of the present invention, there is provided a device for estimating a rear temperature of a vehicle interior, wherein the vehicle interior temperature detecting means is a front interior temperature sensor.

According to an eighth aspect of the present invention, there is provided an apparatus for estimating a rear temperature of a vehicle interior, wherein the environmental factors are an outside air temperature and an amount of solar radiation, and the condition of the rear air conditioner is a blowing mode, a mixed door opening, and a blowing. Either any combination of the air volume information or all the information.

[0019]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a diagram showing a neural network (rear seat temperature estimation model) 12 constituting a rear seat temperature estimator 2b which is a device for estimating a rear temperature of a vehicle interior according to an embodiment of the present invention. Note that the configuration and control flow of the vehicle air conditioner provided with the device for estimating the temperature in the rear of the vehicle compartment of the present invention are the same as those in FIGS. Rear seat temperature estimator 12
b indicates the temperature at the rear part of the vehicle interior as the teacher signal at the time of learning of the neural network 12, using the 12-point average temperature at the positions corresponding to the right, center, and left heads and feet of the rear seats R1, R2. Determination by learning the value of the estimated value T _N2 to determine calculation formula weights w _ij and the bias b _j for each input signal, at the time of control for each input signal, the estimated value T _N2 in the vehicle interior rear temperature Output to rear seat control logic 11b. The input signal of the neural network 12 is measured by the weighting means 20 using the measured temperature T from the front seat temperature sensor (Inc. s) 6 weighted using the weight K _W calculated by the weight amount estimating means 21 in advance.
_i (° C.), the outside air temperature T _a (° C.) from the outside air temperature sensor 7, the solar radiation amount T _s from the solar radiation sensor 8 (Kcal / m ² hour),
Rear blow mode, rear mix door 2e which is the state of the rear air conditioner corresponding to the set temperature and set air volume set on the setting panel 3 (or the rear setting panel not shown)
And the output signal is an estimated value T _N2 (° C.) of the temperature at the rear part of the vehicle interior, which is the opening degree (%) of the rear blower motor duty ratio (%) corresponding to the drive voltage of the rear blower 2b. Here, as an input state to the neural network 12, for each input signal, the maximum value from the minimum value of the measurement data is normalized from 0 to 1 so that the weight w _ij for each type of the input signal can be evaluated in the same manner. ing. However, the measured temperature T _i (° C.) from the front seat temperature sensor (Inc.s) 6 is normalized from the minimum value to the maximum value of the measurement data from 0 to 1, and then weighted by the weighting means 20. The signal width from the minimum value to the maximum value of the measurement data is set to a value smaller than 1, for example, 0.1 to 0.9.

Next, weighting will be described. Normal,
The input state to the neural network is
Use the minimum to maximum value normalized from 0 to 1
Is the measured temperature T from the front seat temperature sensor 6. _i(° C) is front
Despite the temperature near seat F1, neural
Since this is an important factor of the network 12,
The site is big. Therefore, to reduce its effect,
The fluctuation range of the input state to the neural network 12
Needs to be smaller. Therefore, the normalized detected temperature
The degree x (variation range is 0 to 1) is determined by the following equation (2).
The detected temperature y found by the neural network 12
As an input signal, the estimated value T of the rear compartment temperature_N2(° C)
Confuse. y = K_w(X − ／) + ／ (2) where K_wIs the weight, 0 <K_w<1. FIG.
K_w= 0.8, the normalized detected temperature x,
FIG. 7 is a diagram illustrating an example of a relationship with a weighted detected temperature y,
For example, when x = 0.9, y = 0.82, x = 0.1
Y = 0.18, and the variation range of x is 0 to 1 (variation range
Is 1), whereas the variation range of y is 0.1 to 0.
9 (the fluctuation width is 0.8). The above equation (2) is
As shown in equation (3), the fluctuation range is simply proportional to the input value.
Unlike weighting to make it smaller, weighted
The center value of the output temperature y is 0.5, and the center value does not change after weighting.
Does not work. Y = pX ‥‥‥‥ (3) That is, the weighting in the present embodiment is such that the center value is
Weighted so that only the range of fluctuation is small without fluctuation
Therefore, change the fluctuation width symmetrically with respect to the center value.
Can be. That is, the variation of the input signal is
Fluctuations in the input signal.
Can be accurately expressed, and using the above-mentioned comparison formula (3).
Output drift, such as would occur when using
This has the advantage that no damage occurs. Note that the weight
K_wIs manually changed by the designer, and the optimal K_wSet
You.

FIG. 3A shows a rear seat temperature estimator 12b composed of a conventional neural network 12R (see FIG. 9B), and FIG. 3B shows the neural network 12R of the present embodiment. seats temperature estimator 12b after being constituted (see Fig. 1), it was sufficiently learned by the learning conditions below, in a graph estimating the temperature T ₂ of the vehicle interior rear enter the outside air temperature data, (a) The solid line in the figure is the estimated value of the rear temperature in the conventional example, and the solid line in the figure (b) is the estimated value of the rear temperature in the present invention. 3 (a) and 3 (b), the broken line indicates the temperature measured by the front seat temperature sensor 6.
The circles indicate the 12-point average temperatures (actual room temperature) of the measured temperatures measured at the positions corresponding to the right, center, and left heads and feet of the rear seats R1, R2. In the learning of the neural network 12, the weight was set to K _w = 0.8. Learning conditions Outside air temperature -10 to 35 (° C) Solar radiation ‥‥ 0 to 660 (Kcal / m ² hour) Vehicle speed ‥‥ Idle to 40 (equivalent to Km / h) It is clear from FIGS. 3 (a) and (b). As described above, the estimated value () of the rear temperature estimated by the conventional neural network 12 </ b> R is
It can be seen that the estimated value () of the rear temperature estimated in (2) substantially follows the temperature change () of the actual room temperature at the rear of the vehicle compartment, instead of the temperature change () of the temperature measured by the front seat temperature sensor 6. . However, in the case where the temperature change of the temperature measured by the front seat temperature sensor 6 is large as in the region surrounded by the round frame A in FIGS. 3A and 3B, the estimation is performed by the conventional neural network 12R. The estimated value of the rear temperature is different from the change of the actual room temperature under the influence of the temperature measured by the front seat temperature sensor 6, whereas the estimated value of the rear temperature is estimated by the neural network 12 of the present embodiment. The estimated value almost follows the temperature change of the actual room temperature. As described above, when the detected temperature of the front seat temperature sensor 6 is input to the neural network 12, learning of the neural network 12 is performed using a signal weighted such that the signal fluctuation width of the input signal is reduced. By doing so, the influence of the temperature detected by the front seat temperature sensor 6 can be suppressed, and the estimated value T _N2 of the rear temperature can be made very close to the actual room temperature T ₂ .

In the above embodiment, the weight K
_{The w} was manually changed by the designer to set the optimum K _w , but as shown in FIG. 4, an estimated value of the rear temperature estimated by the neural network 12 (rear indoor temperature estimation the deviation of the values) and the actual room temperature and a membership function of the input side, the fuzzy inference means 21F in which the weight quantity K _w and membership function of the output side is provided, if so to determine the weight amount K _w, it can automatically determine the weight amount K _w. Alternatively, as shown in FIG.
The deviation between the rear indoor temperature estimated value and the actual room temperature as an input signal, the weighting amount estimating means 21N provided by the neural network to an output signal of the weights K _w, may be estimated the weight amount K _w.

[0023]

As described above, according to the first aspect of the present invention, the neural network which uses the temperature detected from the cabin temperature detecting means as an input signal and the estimated value of the rear compartment temperature as an output value. When estimating the rear cabin temperature, weighting is performed so as to reduce the signal fluctuation width of the input state of the detected temperature to the neural network,
Since the learning of the neural network is performed, the influence of the detected temperature can be suppressed, and the estimated value of the rear compartment temperature can be made sufficiently close to the actual room temperature of the rear compartment.

According to the second aspect of the present invention, the apparatus for estimating the rear temperature of the vehicle cabin, which is constituted by a neural network for estimating the rear temperature of the vehicle cabin, supplies a signal fluctuation width to the temperature detected by the vehicle temperature detecting means. Weighting means for performing weighting to reduce the weight is added, and learning of the neural network is performed by using the detected temperature weighted by the weighting means as an input signal of the neural network. Can be sufficiently brought close to the temperature of the rear part of the passenger compartment, and the temperature of the rear part of the passenger compartment can be accurately controlled.

According to the third aspect of the present invention, the input state of the detected temperature from the vehicle interior temperature detecting means to the neural network is normalized from the minimum value to the maximum value of the measured data from 0 to 1. Since the weighting is performed such that the width from the minimum value to the maximum value of the measurement data is smaller than 1 while maintaining the center value at 0.5, the fluctuation width is simply reduced in proportion to the input value. Unlike the weighting, accurate weighting can be performed regardless of the level of the detected temperature.

According to the invention described in claim 4, the normalized detected temperature x, the K _w weight amount, and the detected temperature is weighted y, computing equation _{y = K w (x-1} /2 ) + 1 /
Since weighting is performed by 2, it is possible to perform weighting such that the width from the minimum value to the maximum value of the measurement data is made smaller than 1 using a simple conversion formula while maintaining the center value at 0.5.

According to the fifth aspect of the present invention, the weight is estimated by the fuzzy inference means for performing fuzzy inference based on a deviation between the estimated value of the rear temperature of the vehicle interior and the actual room temperature of the rear temperature of the vehicle interior. it can be automatically and accurately determine the weights K _w.

According to the sixth aspect of the present invention, the weight is used as an input signal and a deviation between the estimated value of the rear temperature of the vehicle interior and the actual room temperature of the rear temperature of the vehicle is used as an input signal, and the weight is used as an output value. Since the weight is estimated by the neural network,
_w can be determined automatically and accurately.

According to the seventh aspect of the present invention, since the vehicle interior temperature detecting means is a front interior temperature sensor, the estimated value of the rear temperature of the vehicle interior can be improved without moving the position of the conventional sensor. be able to. Further, since the position of the sensor remains unchanged, there is an advantage that it is not necessary to change the design of the temperature estimator in the front seat.

According to the eighth aspect of the invention, the environmental factors are the outside air temperature and the amount of solar radiation, and the state of the air conditioner is any combination of any of the following information: the blow mode, the mix door opening, and the blow air volume at the rear of the passenger compartment. Since all or all are set, the estimated value of the temperature at the rear of the vehicle compartment can be obtained with high accuracy.

[Brief description of the drawings]

FIG. 1 is a diagram showing a configuration of a rear seat temperature estimation model according to an embodiment of the present invention.

FIG. 2 is a diagram for explaining weighting according to the embodiment.

FIG. 3 is a diagram illustrating a relationship between an estimated value of a vehicle interior temperature and an actually measured value by a temperature estimator of the present embodiment.

FIG. 4 is a diagram illustrating a method of estimating a weight amount according to the embodiment;

FIG. 5 is a diagram illustrating another example of a weight amount estimation method according to the embodiment;

FIG. 6 is a diagram showing a control flow of a conventional vehicle air conditioner.

FIG. 7 is a diagram showing a configuration of a conventional neural network of a vehicle air conditioner.

FIG. 8 is a diagram showing a configuration of a conventional one-box twin-type vehicle air conditioner.

FIG. 9 is a diagram showing a configuration of a temperature estimation model of a conventional one-box twin type vehicle air conditioner.

FIG. 10 is a diagram showing a control flow of a conventional one-box twin type vehicle air conditioner.

[Explanation of symbols]

1 control device, 2A, 2B air duct, 3 setting panel, 4, 4a, 4b temperature setting device, 5a, 5b driving device, 6 inside temperature sensor (front seat temperature sensor), 7 outside temperature sensor, 8 sunlight sensor, 11a front Seat control logic,
11b rear seat control logic, 12a front seat temperature estimator,
12b Rear seat temperature estimator, 20 weighting means, 21
Weight estimation means.

Claims (8)

[Claims] An input value of an estimated value of a rear temperature of a vehicle interior is input to a detected temperature from one vehicle interior temperature detecting means for detecting a temperature of the vehicle interior, an environmental factor of the vehicle, and a state of a rear air conditioner. In the method of estimating the rear cabin temperature by using a neural network that outputs a signal and an estimated value of the rear cabin temperature as an output value, weighting is performed so as to reduce a signal variation width of an input state of the detected temperature to the neural network. A method for estimating a rear temperature in a vehicle interior, wherein learning of the neural network is performed using the performed signal. 2. The temperature and air volume of the front and rear portions of the vehicle compartment can be set independently of each other, and the front and rear portions of the vehicle compartment can be set based on the detected temperature of the vehicle compartment detected by one vehicle compartment temperature detecting means. And an air conditioner for controlling the two air conditioners provided in the vehicle. The detected temperature, the environmental factor of the vehicle, and the state of the air conditioner at the rear are input signals respectively, and the estimated value of the rear temperature of the vehicle interior is provided. Weighting means for weighting the detected temperature to reduce the signal fluctuation range of the input state to the neural network, the weighting means comprising: An apparatus for estimating a rear temperature in a vehicle interior, wherein a signal weighted by weighting means is used as an input signal of the neural network. 3. The input state of the detected temperature from the vehicle interior temperature detecting means to the neural network is normalized from the minimum value to the maximum value of the measured data from 0 to 1, and then the center value is maintained at 0.5. The apparatus for estimating a rear temperature of a vehicle interior according to claim 2, wherein weighting is performed such that a width from a minimum value to a maximum value of the measurement data is smaller than 1. The 4. normalized detected temperature signal x, the weight amount as K _w, weighted by the following equation, according to claim 3, characterized in that to calculate the weighted signal y For estimating the rear temperature of the vehicle interior. y = K _w (x − ／) + ／ 5. The fuzzy inference means according to claim 4, wherein said weighting amount is estimated by fuzzy inference means for performing fuzzy inference based on a deviation between an estimated value of the rear temperature of the vehicle interior and an actual room temperature of the rear temperature of the vehicle interior. A device for estimating the temperature inside the cabin rear. 6. A weighting amount estimating means configured by a neural network having a weighting amount as an input signal and a deviation between an estimated value of the rear temperature of the vehicle interior and an actual room temperature of the rear temperature of the vehicle interior as an input signal. The device for estimating the temperature in the rear of a vehicle compartment according to claim 4, wherein the estimation is performed by: 7. An apparatus for estimating a rear temperature of a vehicle interior according to claim 2, wherein said vehicle interior temperature detecting means is a front interior temperature sensor. 8. The environmental factors are the outside air temperature and the amount of solar radiation, and the state of the rear air conditioner is a combination of any one of the following information of the air outlet mode, the mix door opening, and the air flow of the rear of the vehicle compartment, or each information. 3. The apparatus for estimating the rear temperature of a vehicle interior according to claim 2, wherein