JPH08166818A - Situation recognition device - Google Patents

Abstract

PURPOSE: To adequately recognize the state of an object system. CONSTITUTION: This device is equipped with a dust sensor 107 which detects sucked dust and dirt, a fractal-dimension calculation part 1 which calculates a fractal dimension of the output of the dust sensor 107, a pulse frequency detection part 2 which detects the frequency of the output of the dust sensor 107, and a state decision part 3 which decides the state of the sucked dust and dirt on the basis of the obtained fractal dimension and frequency and controls a motor fan 106 according to the decision result.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a situation recognizing device for recognizing a situation necessary for controlling a control device such as home electric appliances, traffic and air conditioning.

[0002]

2. Description of the Related Art A vacuum cleaner is taken as an example of a control device using a situation recognition device, and a conventional technique will be described.

The structure of a conventional vacuum cleaner is shown in FIG.
101 is a nozzle for sucking dust and dirt on the floor or tatami mat, 102 is a tube that collects dust and dust to the main body by the nozzle 101, 103 is a handle attached to the tube 102, and 104 is the same as the tube 102 A hose for guiding dust to the main body, 105 a paper pack for collecting dust and dust, 106 a motor fan for generating a suction force, 107 a dust sensor for detecting dust passing through the tube 102 (In the figure, the dust sensor 107 is installed near the handle 103), and 108 is a control unit that controls the motor fan 106 according to the detection result of the dust sensor 107.

The structure of the above dust sensor 107 is shown in FIG. The dust sensor 107 is the infrared light emitting diode 10 arranged on the pipe 1074 of the handle 103 portion.
71, a phototransistor 1072, and their control circuits 1073. Dust sensor 107
Then, each time a small particle of dust or dust crosses the light beam emitted from the infrared light emitting diode 1071, a pulse signal as shown in the figure is output. Therefore, if a large amount of dust or dust is inhaled, pulses are frequently generated (frequency becomes high), and if not inhaled too much, pulses decrease and the generation interval becomes long (frequency becomes low). In this way, the frequency of the output signal of the dust sensor 107 corresponds to the amount of dust and dirt inhaled. Utilizing this, the conventional cleaner uses the configuration shown in FIG. 7B, and the control unit 108 drives the motor fan 106 in proportion to the frequency of the output signal of the dust sensor 107.
By such control, for example, when there is a large amount of dust and dust and the frequency of the output signal of the dust sensor 107 becomes high, the motor fan 106 is strongly driven to increase the suction force.

[0005]

However, in the above-mentioned conventional vacuum cleaner, the amount of dust and dust is large when inhaling a large dust or when simultaneously sucking dust and dust whose sizes and weights are not uniform. Can not be estimated well,
There is a problem that proper suction force control cannot be performed. For example, when a relatively large dust (dust) is inhaled, the air flow in the tube 102 constantly changes during the inhalation process of dust, so that the dust sensor 107 outputs a signal such that the frequency changes irregularly. Is output. In addition, when heavy dust and light dust are inhaled at the same time, the dust sensor 10 has different speeds of dust and dust in the tube.
The output of 7 is a signal in which high frequency pulses and low frequency pulses are mixed. Therefore, the amount of dust and dust cannot be sufficiently estimated only by the frequency of the output signal of the dust sensor 107.

As described above, the structure of the conventional dust sensor 107 cannot sufficiently estimate the amount of inhaled dust or dust. Generally, in a system in which a large number of particles can be regarded as moving, when recognizing the status of the system, there is a problem that the status of the target system cannot be properly recognized only by the moving amount.

An object of the present invention is to provide a situation recognition device capable of appropriately recognizing the situation of a target system in consideration of such problems in conventional situation recognition.

[0008]

SUMMARY OF THE INVENTION The present invention provides a status sensor for detecting the status of a predetermined element in a target system, and at least one of the predetermined elements from a detection signal of the status sensor.
A feature amount detection unit for detecting one of the predetermined feature amount, the fluctuation of the detection signal of the state sensor, a fluctuation detection unit for detecting by calculating the fractal dimension or Lyapunov exponent,
The situation recognition device includes a situation determination unit that determines the situation of the target system based on the detection result of the fluctuation detection unit and the detection result of the feature amount detection unit.

[0009]

According to the present invention, the state sensor detects the state of the predetermined element in the target system, and the feature amount detection section detects at least one predetermined feature amount of the predetermined element from the detection signal of the state sensor, Fluctuation detection unit, the fluctuation of the detection signal of the state sensor, by detecting the fractal dimension or Lyapunov exponent, the situation determination means, based on the detection result of the fluctuation detection unit and the detection result of the feature amount detection unit, Determine the status of the target system.

[0010]

DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be described below with reference to the drawings showing its embodiments.

FIG. 1 is a diagram showing a configuration of a vacuum cleaner using a situation recognition apparatus according to the first embodiment of the present invention, and shows a control diagram of the vacuum cleaner. In FIG. 1, 106 is a motor fan, 107 is a dust sensor as a state sensor, and these are the same as the structure of FIG. 7 of a prior art example. On the other hand, 1 is a fractal dimension calculation unit as a fluctuation detection unit that obtains a fractal dimension of the signal output from the dust sensor 107, and 2 is a pulse frequency as a feature amount detection unit that detects the frequency of the signal output from the dust sensor 107. The detection unit 3 is a situation determination unit that controls the motor fan 106 by estimating the amount and state of inhaled dust and dust according to the outputs from the fractal dimension calculation unit 1 and the pulse frequency detection unit 2. The situation determination unit 3 includes a situation determination unit and a situation recognition control unit.

Next, the operation of the vacuum cleaner using the situation recognition device of the first embodiment will be described with reference to the drawings.

Here, the fractal dimension calculated by the fractal dimension calculation unit 1 is an index representing the self-similarity and complexity of the behavior of the target system, and the fluctuation of the system and the degree of the complexity of the behavior are quantified. You can know it.

First, several methods for obtaining the fractal dimension have been proposed by academic societies. Here, one of the methods is shown (T.Higuchi: Physica D, Vol.31, p.
p.277-283, 1988).

It is assumed that the output signal from the target system at time t is represented as x (t), and the data from time 0 to time M-1 is obtained. At this time, from time k to time interval t
If the length of the broken line of the data in is represented by Δ _k (t) / t,
Δ _k (t) is

[0016]

[Equation 1] Can be expressed as However, L _t = [(Mk) / t] t, and [z] is an integer not exceeding z. At this time, the average of Δ _k (t) with respect to k is represented by Δ (t). If the following relation holds, the index D (1 ≦ D ≦ 2) at that time indicates the fractal dimension of the data.

[0017]

[Equation 2] To obtain the fractal dimension D, log from the actual signal
Calculate (Δ (t) / t) and log t, linearly approximate the relationship between them, and calculate the proportional constant. From the relationship of (Equation 2),
The fractal dimension can be obtained from the calculated proportional constant value.

The fractal dimension thus obtained is
In general, if the state of the target system changes smoothly like a linear system, D = 1, and if the state is maximally disordered,
It is known that D = 2. By obtaining such a fractal dimension, it is possible to distinguish whether the state of the target system is a state that undergoes a simple linear change or is a very disordered state, and the degree of this can also be determined by the value of D. I can know.

Therefore, with the configuration as shown in FIG. 1, by calculating the fractal dimension for the pulse signal output from the dust sensor 107, it is possible to evaluate the irregular change (fluctuation) of the frequency of the pulse signal. it can.
As a result, it is possible to know the state in which a large dust (dust) is inhaled, a heavy dust and a light dust are simultaneously inhaled, and the like.

In this embodiment, the fractal dimension calculating unit 1 calculates the fractal dimension D from the pulse signal output from the dust sensor 107, and at the same time, the pulse frequency detecting unit 2
To detect frequency F. The situation determination unit 3 determines the current inhalation state of dust and dust from the fractal dimension value D calculated by the fractal dimension calculation unit 1 and the frequency F detected by the pulse frequency detection unit 2, and The motor fan 106 is controlled based on the rules. 1) The value of fractal dimension D is close to 1, and the frequency F of the pulse signal is low.

-> The output of the motor fan 106 is reduced. 2) The value of the fractal dimension D is close to 1, and the frequency F of the pulse signal is high.

-> Increase the output of the motor fan 106. 3) The value of the fractal dimension D is close to 2, and the frequency F of the pulse signal is low.

-> Increase the output of the motor fan 106. 4) The value of the fractal dimension D is close to 2, and the frequency F of the pulse signal is high.

-> Maximize the output of the motor fan 106. The situation determination unit 3 executes the above rules by fuzzy inference. As a result, suction input control can be performed according to various types and states of dust and dust sucked from the nozzle, and power consumption of the cleaner can be reduced.

In the present embodiment, the dust sensor 107
The fractal dimension calculation unit 1 was used to evaluate the irregular change in the signal output from, but as shown in FIG.
Instead of the fractal dimension calculation unit 1, the Lyapunov exponent calculation unit 11 may be used. The Lyapunov exponent is an index for evaluating the stability of the target system, and in particular, the maximum Lyapunov exponent, which is the maximum value of the Lyapunov exponent, well represents the characteristics of the target system. The maximum Lyapunov exponent has a negative value when the state of the target system converges to a certain state, zero when it changes regularly with a fixed period, and a positive value when it changes irregularly. Become. By obtaining this maximum Lyapunov exponent in the Lyapunov exponent calculation unit 11, the same effect as when the fractal dimension is used can be obtained. However, when the Lyapunov exponent calculation unit 11 is used, the fuzzy inference performed by the situation determination unit 3 ′ must be changed to the following rules. 1) Maximum Lyapunov exponent is negative, pulse signal frequency F
Is low.

-> The output of the motor fan 106 is reduced. 2) Maximum Lyapunov exponent is negative, pulse signal frequency F
Is high.

-> The output of the motor fan 106 is increased. 3) Maximum Lyapunov exponent is positive, pulse signal frequency F
Is low.

-> The output of the motor fan 106 is increased. 4) Maximum Lyapunov exponent is positive, pulse signal frequency F
Is high.

-> Maximize the output of the motor fan 106.

Many methods for obtaining the maximum Lyapunov index have been proposed by academic societies (TS Parker, LOChua: Pract.
ical Numerical Algorithm for Chaotic System, Sprin
ger-Verlag, 1989). A simple calculation method is shown below as an example.

First, from the signal x (t) detected by the dust sensor 107, the time series vector X (i) = {x (i), x (i +
T), x (i + 2T), ...., x (i + (d-1) × T)}. Where d
Indicates the dimension of the time series vector, and T indicates the time delay amount. Set d and T to appropriate values. At this time, choose a suitable plane in the d-dimensional space, and a vector that crosses this hyperplane.
Calculate X (i) -X (i + 1). Let X (i) be the coordinates of the intersection on the hyperplane
And X (i + 1) as the interior division point, and the set of points on the plane {X
Make p1, Xp2, ..., Xpk, ....}. In this set
Select all pairs whose distance is equal to or less than the predetermined threshold E, and express the two points in them as Xpk, Xpk '. At this time, the maximum Lyapunov exponent λ can be calculated by the following equation.

[0032]

(Equation 3) However, Np indicates the total number of data pairs whose distance is less than or equal to the threshold E. In (Equation 3), if the value of τ is increased, λ
(τ) converges. Λ (τ) when it converges is the maximum Lyapunov exponent. Note that other methods for calculating the maximum Lyapunov number have been proposed, and other methods may be used for the calculation.

FIG. 3 is a block diagram of a situation recognition apparatus according to the second embodiment of the present invention, and specifically shows a traffic monitoring system.

In FIG. 3, reference numeral 1 is a fractal dimension calculation unit, and 2 is a pulse frequency detection unit, which is similar to that of the first embodiment. Reference numeral 21 is a vehicle detection sensor for detecting a vehicle, 3 '' is a situation determination unit for determining a traffic situation according to the outputs of the fractal dimension calculation unit 1 and the pulse frequency detection unit 2, and 2
Reference numeral 2 denotes a display unit as a traffic condition display unit that displays the determination result of the condition determination unit 3 ″.

Various methods are known to grasp the traffic condition of the road. As one of the easy methods,
As shown in FIG. 4 (a), there is a method of setting a pillar from the side of the road and installing a vehicle detection sensor 21 on the road to detect the flow of the vehicle. The vehicle detection sensor 21 irradiates a sound wave such as an ultrasonic wave or an electromagnetic wave such as an infrared ray downward and detects the presence or absence of the vehicle by detecting the reflection thereof. With such a sensor, a pulsed output signal as shown in FIG. 4B is obtained. Here, a signal of high level output is output when the vehicle has passed, and has a low level when the vehicle has not passed.

In general, when the vehicle is flowing smoothly on the road, the output of the vehicle detection sensor 21 becomes a regular and substantially constant high frequency pulse signal, and when the vehicle is congested, the low frequency signal. Is output. In the conventional traffic monitoring system, the flow of vehicles on the road is estimated by detecting the frequency of the vehicle detection sensor 21.

However, in the conventional system, since only the frequency of the sensor output is detected, only basic information about the flow of the vehicle is known, and other traffic information cannot be obtained. The present embodiment proposes a device that extracts more detailed traffic information than before by calculating the fractal dimension from the output signal of the vehicle detection sensor 21.

The fractal dimension can evaluate the irregular fluctuations of the state of the target system. The fractal dimension calculation method described in the first embodiment is known to be 1 when the system state changes smoothly like a linear system, and 2 when the state is maximally disordered.

Generally, when an accident occurs before the detection place or when an obstacle falls on the road,
Since there is a difference between the vehicles in the driving operation for avoidance, the output pulse interval of the vehicle detection sensor 21 changes at a relatively large level, and the frequency does not become a constant frequency but fluctuates. In such a case, by calculating the fractal dimension of the output signal, it is possible to grasp the fluctuation component of the signal,
You can see that there is some problem on the road.

Therefore, the traffic condition can be known in more detail by making the following judgment in the condition judging section 3 ''. 1) The value of fractal dimension D is close to 1, and the frequency F of the pulse signal is low.

-> Congestion. 2) The value of the fractal dimension D is close to 1, and the frequency F of the pulse signal is high.

-> The car is flowing smoothly. 3) The value of fractal dimension D is close to 2, and the frequency of the pulse signal is low.

-> There is a considerable problem in traffic conditions.

(It is highly possible that the running condition of the vehicle is disturbed due to an accident, etc.) 4) The value of the fractal dimension D is close to 2, and the frequency of the pulse signal is high.

-> There is a slight problem in traffic conditions, and there is a possibility that there are obstacles on the road.

The display section 22 displays the judgment result of the situation judging section 3 ''. As described above, by providing the fractal dimension calculation unit 1, the fractal dimension is calculated for the output signal of the vehicle detection sensor 21, and the fractal dimension calculation unit 1 evaluates the fractal dimension together with the frequency information obtained from the pulse frequency detection unit 2. I can know.

Similar to the first embodiment, the same determination can be made by using the Lyapunov exponent calculation unit instead of the fractal dimension calculation unit 1.

FIG. 5 is a block diagram of a situation recognition device according to the third embodiment of the present invention, and specifically shows a commercial air conditioning system using images. The commercial air conditioning system here is an air conditioning system for a relatively large space such as a department store floor or a station premises.

In FIG. 5, reference numeral 1 is a fractal dimension calculation unit, which is the same as that in the first embodiment. Reference numeral 31 is a video camera for photographing the air-conditioned space, 32 is a temperature sensor for measuring the representative temperature of the air-conditioned space, and 3 '''is the estimation of the thermal condition based on the output of the fractal dimension calculation unit 1 and the output of the temperature sensor 32. The condition determination unit 33 for performing the operation is an air conditioner that controls the air conditioning based on the condition determination of the condition determination unit 3 ′ ″. Here, part of the situation determination unit 3 ′ ″ and the air conditioner 33 constitutes a situation determination control unit.

In the conventional air-conditioning system, constant value control for keeping the room temperature at the target temperature is often performed. However, recently, more advanced air-conditioning control research has been done, estimating the number and position of people in the room from thermal images obtained by pyroelectric sensors and visible images from video cameras,
Air conditioning has been realized accordingly (see Japanese Patent Application No. 4-254302).

However, in order to realize a higher degree of air conditioning, it is necessary to detect not only the rough information on the number and position of people but also the movement of people. In a relatively narrow space such as a home or an office, since the upper limit of the number of people who exist is not so large, it is possible to track the movement of each person from the image information and detect the movement in some cases. However, in a relatively large space such as a department store floor or a station yard, many people gather, so it is practically impossible to detect individual movements from image information. Need to be caught.

The fractal dimension is suitable for capturing macroscopic human movement in space from image information obtained by a video camera. The fractal dimension can evaluate the complexity and irregular fluctuation of the state of the target system as described in detail in the first embodiment. The method of calculating the fractal dimension from the image information can be basically calculated by expanding the method described in the first embodiment into a two-dimensional method. In the paper, Hiroshi Kaneko'Fractal feature and texture analysis' (electronic information IEICE Transactions D Vol. J70-D No.5 pp.964--972)
It is described in detail in.

The situation recognition from the image information by the fractal dimension will be described with reference to the concrete example of FIG. Figure 6
(A) is a case where there are few people in the room. In this case, there are naturally few people in the image taken by the video camera 31,
It appears sparsely. If the fractal dimension is calculated from such an image, a relatively small value can be obtained because the complexity of the image is not great. In the case of FIG. 6B, there are many people, but they are moving side by side and regularly. This is the case, for example, when it is crowded with cashiers at department stores and ticket purchases at stations. In such a case, the video camera 3
When the fractal dimension is calculated for the image obtained from 1, a small value is obtained because the human flow is a regular movement. FIG. 6C shows a situation in which there are many people and it is very crowded. In this case, the motion of each person is disjointed, and the video has complicated motions, so the fractal dimension becomes large.

In the above three cases, the most power is required for air conditioning in the case of (c). For example, during cooling, it is necessary to increase the air volume and lower the target temperature. In the case of (b), although there are many people, they move in a regular line, so the wind is good, and it is comfortable even if the cooling is not strengthened too much.
In the case of (a), naturally, it is not necessary to increase the degree of air conditioning.

As described above, by calculating the fractal dimension from the image information obtained from the video camera 31,
Macro information about the movement of a person in the image can be detected, and the air conditioning can be controlled based on the detected macro information. As a result, more comfortable air conditioning control than in the past can be realized.

In order to realize such air conditioning control based on the fractal dimension, the configuration shown in FIG. 5 is required.
The image captured by the video camera 31 is input to the fractal dimension calculation unit 1, and the fractal dimension for the captured image is calculated. Further, temperature information can be obtained from the temperature sensor 32 installed indoors. The situation determination unit 3 ′ ″ controls the air conditioning of the air conditioner 33 by making the following determinations based on the obtained fractal dimension and room temperature. However, in this case, cooling is taken as an example. 1) The value of fractal dimension D is close to 1.

-> Normal room temperature constant control. 3) The fractal dimension D is close to 2 and the room temperature is close to the target temperature.

-> The target temperature is slightly lowered. 4) The fractal dimension D is close to 2, and the room temperature is higher than the target temperature.

-> Increase the air volume and lower the target temperature.

As described above, the fractal dimension calculation unit 1 is provided, the fractal dimension is calculated for the output image of the video camera 31, and the air conditioning is performed based on the calculated fractal dimension.
It is possible to perform more detailed air conditioning control than before.

As described above, by calculating the fractal dimension or Lyapunov exponent for the signal obtained from the sensor, the fluctuation and complexity of the target system can be evaluated. Alternatively, by controlling the target system based on the Lyapunov exponent, it is possible to perform finer control than before.

Similar to the first embodiment, the same judgment can be made by using the Lyapunov exponent calculation unit instead of the fractal dimension calculation unit 1.

Further, in the third embodiment, the video camera 31 is used to detect the indoor condition, but instead of this, a pyroelectric sensor is used to detect a thermal image, and the thermal image is detected. On the other hand, the same effect can be obtained by calculating the fractal dimension.

Further, in the third embodiment, the information such as the number of persons that can be detected from the image is not used, but the number of persons and the humidity and temperature distribution can be detected by a conventional sensor. Information or the like may be input to the situation determination unit 3 '''so that it can be used for situation determination. In this case, more detailed control can be realized.

Further, in the above embodiment, an example in which the situation recognition device is applied to a vacuum cleaner, a traffic monitoring system, and an air conditioner has been described, but the present invention is not limited to this, and any system in which a large number of particles can be regarded as moving Applicable.

[0066]

As is apparent from the above description, the present invention is characterized by detecting at least one predetermined characteristic amount of a predetermined element from the detection signal of the state sensor that detects the state of the predetermined element in the target system. Amount detection unit, the fluctuation of the detection signal of the state sensor, a fluctuation detection unit to detect by calculating the fractal dimension or Lyapunov exponent, based on the detection result of the fluctuation detection unit and the detection result of the feature amount detection unit, Since the status determination means for determining the status of the target system is provided, it has an advantage that the status of the target system can be appropriately recognized.

Further, there is an advantage that fine control can be performed by controlling the target system based on the judgment result of the status judging means.

[Brief description of drawings]

FIG. 1 is a diagram showing a configuration of a vacuum cleaner using a situation recognition device according to a first embodiment of the present invention.

FIG. 2 is a diagram showing a configuration in which a Lyapunov exponent calculation unit is used as a fluctuation detection unit in the first embodiment.

FIG. 3 is a diagram showing a configuration of a traffic monitoring system according to a second embodiment of the present invention.

FIG. 4A is a diagram showing an installation situation of a vehicle detection sensor in the second embodiment, and FIG. 4B is a diagram showing an output signal of the vehicle detection sensor.

FIG. 5 is a diagram showing a configuration of an air conditioning system according to a third embodiment of the present invention.

6A, 6B, and 6C are diagrams for explaining the relationship between the image and the fractal dimension in the third embodiment.

FIG. 7A is a diagram showing a schematic structure of a conventional vacuum cleaner, and FIG. 7B is a diagram showing its configuration.

FIG. 8 is a diagram illustrating an operation of a conventional dust sensor.

[Explanation of symbols] 1 fractal dimension calculation unit 2 pulse frequency detection unit 3 situation determination unit 11 Lyapunov exponent calculation unit 21 vehicle detection sensor 22 display unit 31 video camera 32 temperature sensor 33 air conditioner

─────────────────────────────────────────────────── ───

[Procedure amendment]

[Submission date] April 3, 1995

[Procedure Amendment 1]

[Document name to be amended] Statement

[Correction target item name] 0053

[Correction method] Change

[Correction content]

The situation recognition from the image information by the fractal dimension will be described with reference to the concrete example of FIG. Figure 6
(I) is a case where there are few people in the room. In this case, there are naturally few people in the image taken by the video camera 31,
It appears sparsely. If the fractal dimension is calculated from such an image, a relatively small value can be obtained because the complexity of the image is not great. In the case of FIG. 6 (ii), there are many people, but they are moving side by side and regularly. This is the case, for example, when it is crowded with cashiers at department stores and ticket purchases at stations. In such a case, the video camera 3
When the fractal dimension is calculated for the image obtained from 1, a small value is obtained because the human flow is a regular movement. FIG. 6 (iii) shows a situation where there are many people and it is very crowded. In this case, the movements of each person are different,
The fractal dimension increases because the image has complicated movements.

[Procedure Amendment 2]

[Document name to be amended] Statement

[Correction target item name] 0054

[Correction method] Change

[Correction content]

In the above three cases, it is the case of (iii) that the most power is required for air conditioning. For example, during cooling,
It is necessary to increase the air volume and lower the target temperature. In the case of (ii), although there are many people, they move in a regular line, so the wind is good, and it is comfortable even if the cooling is not strengthened too much. In the case of (i), naturally, it is not necessary to increase the degree of air conditioning.

[Procedure 3]

[Document name to be amended] Statement

[Name of item to be corrected] Figure 6

[Correction method] Change

[Correction content]

FIGS. 6 (i), (ii), and (iii) are diagrams for explaining the relationship between the image and the fractal dimension in the third embodiment.

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ Identification code Internal reference number FI Technical display location G05B 24/02 G06T 1/00 G08G 1/01 D

Claims (4)

[Claims] 1. A state sensor that detects a state of a predetermined element in a target system, a feature amount detection unit that detects at least one predetermined feature amount of the predetermined element from a detection signal of the state sensor, and the state. Fluctuation of the detection signal of the sensor, the fluctuation detection unit to detect by calculating the fractal dimension or Lyapunov exponent, based on the detection result of the fluctuation detection unit and the detection result of the feature amount detection unit, the situation of the target system A situation recognition device comprising a situation determination means for determining. 2. The situation recognizing device according to claim 1, wherein the state sensor is a dust sensor that detects passage of dust and dust, and the predetermined feature amount is a frequency of a signal output from the dust sensor. And a situation recognition control unit that controls the motor fan according to the determination result of the situation determination means of the situation recognition device. 3. A sensor for detecting a temperature characteristic in a room, an image capturing device for capturing at least one of a visible image and a thermal image in the chamber, and a fractal for a crowd in an image obtained by the image capturing device. A fluctuation detection unit that calculates a dimension or a Lyapunov exponent, and a situation determination control unit that controls at least one of an air volume, a wind direction, and a target temperature based on an output of the sensor and an output of the fluctuation detection unit. Characteristic air conditioning equipment. 4. The situation according to claim 1, wherein the state sensor is a vehicle detection sensor that detects passage of a vehicle on a road, and the predetermined feature amount is a frequency of a signal output from the vehicle detection sensor. A traffic monitoring system, comprising: a recognition device; and a traffic condition display unit that displays the traffic condition of the road according to the determination result of the condition determination means of the condition recognition device.