CN112051621A - Method and device for judging whether people are in room or not - Google Patents
Method and device for judging whether people are in room or not Download PDFInfo
- Publication number
- CN112051621A CN112051621A CN201910754512.XA CN201910754512A CN112051621A CN 112051621 A CN112051621 A CN 112051621A CN 201910754512 A CN201910754512 A CN 201910754512A CN 112051621 A CN112051621 A CN 112051621A
- Authority
- CN
- China
- Prior art keywords
- room
- probability
- person
- data
- people
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 67
- 230000007613 environmental effect Effects 0.000 claims description 127
- 230000000875 corresponding effect Effects 0.000 claims description 48
- CURLTUGMZLYLDI-UHFFFAOYSA-N Carbon dioxide Chemical compound O=C=O CURLTUGMZLYLDI-UHFFFAOYSA-N 0.000 claims description 46
- 229910002092 carbon dioxide Inorganic materials 0.000 claims description 23
- 238000005286 illumination Methods 0.000 claims description 21
- 239000001569 carbon dioxide Substances 0.000 claims description 20
- 238000004364 calculation method Methods 0.000 claims description 10
- 150000001875 compounds Chemical class 0.000 claims description 8
- 230000002596 correlated effect Effects 0.000 claims description 8
- 238000004590 computer program Methods 0.000 claims description 6
- 238000002474 experimental method Methods 0.000 description 11
- 238000010801 machine learning Methods 0.000 description 11
- 238000001514 detection method Methods 0.000 description 8
- 238000012549 training Methods 0.000 description 7
- 230000001960 triggered effect Effects 0.000 description 4
- 230000006399 behavior Effects 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 3
- 229930091051 Arenine Natural products 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 230000001276 controlling effect Effects 0.000 description 1
- 238000002790 cross-validation Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000005265 energy consumption Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
- G01V9/00—Prospecting or detecting by methods not provided for in groups G01V1/00 - G01V8/00
Abstract
The invention relates to a method and a device for judging the state of people in a room without people, belonging to the field of intelligent home furnishing, wherein the method comprises the steps of storing historical environment data of people and people in the room, wherein the historical environment data at least comprises data of one environment characteristic, calculating the conditional probability of each environment characteristic in the historical environment data of people and people in the corresponding room by respectively utilizing the historical environment data of people and people in the room, and calculating according to the positive correlation between the conditional probabilities of people in the room and the probability of people in the room to obtain the probability of people in the room at the moment corresponding to the current environment data; and calculating according to the positive correlation between the conditional probabilities when no person exists in the room and the probability of no person in the room to obtain the probability of no person in the room at the moment corresponding to the current environment data. And finally, whether a person exists in the room is judged by utilizing the larger probability of the person and the probability of no person, the judgment reliability is high, and the problem of failure judgment is solved.
Description
Technical Field
The invention belongs to the field of intelligent home furnishing, and particularly relates to a method and a device for judging the state of people in a room.
Background
At present, the method can accurately judge whether people are in a room, and plays a crucial role in determining that an intelligent control system automatically changes the control mode (such as a comfortable or energy-saving mode) of the room and automatically controls indoor equipment. For example, when the room is judged to be 'unmanned', the system can automatically turn off energy consumption equipment such as indoor air conditioners and lamplight, and unnecessary energy waste is avoided; when the 'people' in the room is judged, the control logic of the room is converted into a preset mode (such as a comfortable mode), and when indexes such as temperature, humidity, air quality or illumination and the like do not reach a set optimal value, indoor equipment can be controlled to be automatically started to operate, or the indoor equipment can be used for systems such as security alarm and the like.
In the prior art, the research on the judgment of the presence or absence of people in a room mainly focuses on the following aspects:
(1) the presence or absence of a person in a current room is determined by motion detection, and a sensor generally used includes: the intelligent door and window detection system comprises an infrared human body detector, an intelligent door magnetic sensor, an intelligent camera and the like, and can detect moving objects or door and window opening and closing actions so as to judge the state of people in a current room. The sensors belong to invasive sensors, are mostly deployed in places such as outdoors and doorways in personal home environments, and prevent the situation that the indoor conditions cannot be well judged due to the limited detection range of the sensors at fixed positions such as doorways and the like, detection dead angles exist, when people in a room are in the detection dead angles, the motion detection judgment method is invalid, and the judgment reliability is low.
(3) Whether people exist in a current room is judged through WIFI signals or GPS positioning detection, the method has certain limitation, a user is required to hold a mobile phone by hands or put the mobile phone in a pocket, and once the mobile phone is used for not carrying the mobile phone with the user, the method is invalid in judgment, so that the method is low in applicability and poor in reliability.
Disclosure of Invention
The invention aims to provide a method and a device for judging the presence or absence state of a person in a room, which are used for solving the problem of low reliability in the prior art for judging the presence or absence of the person in the room.
Based on the above purpose, a first technical scheme of a method for judging whether a person is in a room or not by calculating the probability of the person in the room and the probability of the absence of the person in the room is as follows:
storing historical environmental data of people and no people in a room, wherein the historical environmental data at least comprises data of one environmental characteristic, and the environmental characteristic comprises: carbon dioxide concentration, illumination intensity and noise decibel;
the method comprises the following steps of collecting current environment data in a room, and calculating the probability of people and the probability of no people in the room at the moment corresponding to the current environment data, wherein the calculation mode of the probability of people is as follows: taking the historical environment data of people as samples, respectively calculating the conditional probability of a certain environment characteristic in the current environment data when people exist in a room, and calculating the obtained conditional probability of each environment characteristic to obtain the probability of people in the room at the moment corresponding to the current environment data, so that the probability of people is positively correlated with each conditional probability;
the calculation mode of the unmanned probability is as follows: taking the historical environment data of nobody as a sample, respectively calculating the conditional probability of a certain environment characteristic in the current environment data when no person exists in a room, and calculating the obtained conditional probability of each environment characteristic to obtain the nobody probability in the room at the moment corresponding to the current environment data, so that the nobody probability is positively correlated with each conditional probability;
and judging whether the room is occupied at the moment corresponding to the current environment data according to the greater probability of the probability of being occupied and the probability of being unoccupied.
The beneficial effects of the above technical scheme are:
the invention utilizes the historical environment data of the presence and absence of people in the room and the historical environment data of the presence and absence of people in the room to calculate the conditional probability of each environment characteristic in the historical environment data of the presence and absence of people in the room, and the operation is carried out according to the positive correlation between the conditional probabilities and the probability of the presence of people in the room, thereby obtaining the probability of the presence of people in the room at the moment corresponding to the current environment data.
Meanwhile, the conditional probabilities of all the environmental features in the historical environment data when no person exists in the room are calculated by utilizing the historical environment data when no person exists in the room, and the operation is carried out according to the positive correlation between the conditional probabilities and the probability of no person in the room, so that the probability of no person in the room at the moment corresponding to the current environment data is obtained. And finally, whether a person exists in the room is judged by utilizing the higher probability of the person and the probability of the absence of the person, compared with the method in the prior art, the method has high reliability, and the problem of failure judgment does not exist.
In order to ensure the reliability of the determination method of the present invention, one embodiment of calculating the conditional probability is:
the conditional probability under a certain environmental characteristic in the current environmental data when the room is occupied is as follows:
in the formula (I), the compound is shown in the specification,
for environmental characteristics when a person is present in a room
The conditional probability of (a) of (b),
wherein
Respectively representing the values of carbon dioxide concentration, illumination intensity and noise decibel at the moment T, wherein the moment T is the moment corresponding to the current environmental data; mu.s1Corresponding environmental characteristics in historical environmental data for people in a room
Average value of (a)1Corresponding environmental characteristics in historical environmental data for people in a room
The variance of the value of (a).
The conditional probability under a certain environmental characteristic in the current environmental data when the room is unmanned is as follows:
in the formula (I), the compound is shown in the specification,
for environmental characteristics when no person is in the room
Conditional probability of (d), mu2Corresponding environmental characteristics in historical environmental data for no person in a room
Average value of (a)2Corresponding environmental characteristics in historical environmental data for no person in a room
The variance of the value of (a).
In order to improve the accuracy of judging whether a person is in the room or not, when the current environment data corresponds to the probability of the person in the room at the moment, the operation is carried out by multiplying the conditional probabilities by the prior probabilities of the person in the room and the person in the room in the historical environment data of the person in the room and the person not in the room.
Meanwhile, when the probability of no person in the room at the moment corresponding to the current environment data is obtained, the operation is performed in such a way that the conditional probabilities are multiplied and then multiplied by the prior probability of the presence of a person in the room and the absence of a person in the room in the historical environment data of the absence of a person, so that the misjudgment of the presence of a person in the room and the absence of a person in the room is reduced.
Based on the above purpose, a second technical scheme of the method for judging the presence or absence of a person in a room is to judge whether the person is in the room by calculating the probability of the person in the room, and the specific scheme is as follows:
storing historical environmental data of a person in a room, the historical environmental data at least comprising data of one environmental characteristic, the environmental characteristic comprises: carbon dioxide concentration, illumination intensity and noise decibel;
the method comprises the following steps of collecting current environment data in a room, and calculating the probability of people in the room at the moment corresponding to the current environment data, wherein the calculation mode is as follows: taking the historical environment data of people as samples, respectively calculating the conditional probability of a certain environment characteristic in the current environment data when people exist in a room, and calculating the obtained conditional probability of each environment characteristic to obtain the probability of people in the room at the moment corresponding to the current environment data, so that the probability of people is positively correlated with each conditional probability;
and comparing the probability of the person with a set probability threshold, and judging whether the person exists in the room at the moment corresponding to the current environment data according to the comparison result.
The beneficial effects of the above technical scheme are:
according to the invention, the condition probability of each environmental characteristic in the historical environment data when people exist in the room is calculated by collecting the historical environment data when people exist in the room and utilizing the historical environment data when people exist in the room, and the operation is carried out according to the positive correlation between the condition probabilities and the probability of people in the room, so that the current environment data corresponds to the probability of people in the room at the moment. And finally, judging whether a person exists in the room or not by utilizing a comparison result between the probability of the person and a set probability threshold, wherein compared with the judging method in the prior art, the judging method has the advantages that the problem of method failure does not exist, and the judging reliability is high.
In order to ensure the reliability of the judging method of the invention, the conditional probability under a certain environmental characteristic in the current environmental data is as follows:
in the formula (I), the compound is shown in the specification,
for environmental characteristics when a person is present in a room
The conditional probability of (a) of (b),
wherein
Respectively representing the values of carbon dioxide concentration, illumination intensity and noise decibel at the moment T, wherein the moment T is the moment corresponding to the current environmental data; mu.s1Corresponding environmental characteristics in historical environmental data for people in a room
Average value of (a)1Corresponding environmental characteristics in historical environmental data for people in a room
The variance of the value of (a).
In order to improve the accuracy of judging whether a person is in the room or not, the method further comprises the step of pre-storing historical environment data of the absence of the person in the room, and the operation is to multiply all the conditional probabilities after being multiplied and then multiply the prior probabilities of the presence of the person in the room and the absence of the person in the room in the historical environment data, so that the probability of the presence of the person in the room at the moment corresponding to the current environment data is obtained, and the misjudgment of the presence of the person in the room and the absence of the person in the room is reduced.
Based on the above purpose, a third technical scheme of the method for judging the presence or absence of a person in a room is to judge whether the person is present in the room by calculating the probability of absence in the room, and the specific scheme is as follows:
storing historical environmental data of the room when no person is present, wherein the historical environmental data at least comprises data of environmental characteristics, and the environmental characteristics comprise: carbon dioxide concentration, illumination intensity and noise decibel;
the method comprises the following steps of collecting current environment data in a room, and calculating the probability of no person in the room at the moment corresponding to the current environment data, wherein the calculation mode is as follows: taking the historical environment data of nobody as a sample, respectively calculating the conditional probability of a certain environment characteristic in the current environment data when no person exists in a room, and calculating the obtained conditional probability of each environment characteristic to obtain the nobody probability in the room at the moment corresponding to the current environment data, so that the nobody probability is positively correlated with each conditional probability;
and comparing the unmanned probability with a set probability threshold, and judging whether people exist in the room at the moment corresponding to the current environment data according to the comparison result.
The beneficial effects of the above technical scheme are:
according to the invention, the conditional probability of each environmental characteristic in the historical environment data when no person exists in the room is calculated by collecting the historical environment data when no person exists in the room and utilizing the historical environment data when no person exists in the room, and the operation is carried out according to the positive correlation between the conditional probabilities and the probability of no person in the room, so that the current environment data corresponds to the probability of no person in the room at the moment. And finally, judging whether a person exists in the room or not by utilizing a comparison result between the unmanned probability and a set probability threshold, wherein compared with the judging method in the prior art, the method has the advantages that the problem of method failure does not exist, and the judging reliability is high.
In order to ensure the accuracy of judgment of a person in a room, the conditional probability of a certain environmental characteristic in the current environmental data is as follows:
in the formula (I), the compound is shown in the specification,
for environmental characteristics when no person is in the room
The conditional probability of (a) of (b),
wherein
Respectively representing the values of carbon dioxide concentration, illumination intensity and noise decibel at the moment T, wherein the moment T is the moment corresponding to the current environmental data; mu.s2Corresponding environmental characteristics in historical environmental data for no person in a room
Average value of (a)2Corresponding environmental characteristics in historical environmental data for no person in a room
The variance of the value of (a).
In order to improve the accuracy of judging whether a person is in the room or not, the method further comprises the step of pre-storing historical environment data of the person in the room, wherein the operation is to multiply all the conditional probabilities after being multiplied, and then multiply the prior probabilities of the person in the room and the person in the unmanned historical environment data when the person is not in the room, so that the probability of the absence of the person in the room at the moment corresponding to the current environment data is obtained, and the misjudgment of the person in the room and the absence of the person in the room is reduced.
Based on the above purpose, a technical scheme of the device for judging the state that people are not in a room is as follows:
the system comprises a memory, a processor and a computer program which is stored on the memory and runs on the processor, wherein the processor is coupled with the memory, and when the processor executes the computer program, the method for judging the existence state of any one of the first technical scheme to the third technical scheme is realized.
Drawings
FIG. 1 is a flow chart of the present invention for determining whether a person is present in a room;
FIG. 2 is a schematic diagram of the present invention trained using a naive Bayes model;
FIG. 3 is a schematic representation of features in historical environmental data of the present invention.
Detailed Description
The following further describes embodiments of the present invention with reference to the drawings.
The first embodiment of the method comprises the following steps:
a corresponding number of sensors are provided at appropriate locations in the room for collecting environmental data, and one embodiment of the sensors is to provide a sensor C1 for collecting light intensity near a window in the room, a sensor C2 for collecting carbon dioxide concentration near a sofa or bed in the room, and a sensor C3 for collecting noise decibels.
Then, when the user acts in the room according to the behavior habit of the user, the sensors (C1, C2 and C3) are utilized to collect environmental data within a period of time to be used as historical environmental data of people in the room, wherein the historical environmental data comprises three environmental characteristics of carbon dioxide concentration, illumination intensity and noise decibel; when the environment data of no person in the room is collected by the sensors and used as historical environment data of no person in the room, the historical environment data of the person in the room and the historical environment data of the person in the room are stored in the memory.
According to historical environment data of people and no people in a room, a machine learning model of the relation between the people and all environment characteristic values in the room is established, and if the established naive Bayes model is as follows:
in the above model, P (Y ═ Y)1|X＝XT) Represented in the current environment data XTProbability of someone in the lower room, XTIs the environmental data at the time T,
wherein
Respectively representing carbon dioxide concentration and illumination at T momentIntensity, decibel of noise values; p (Y ═ Y)1) And P (Y ═ Y)2) Respectively representing the prior probability of a person and of an nobody,
i＝(1,2)，n1and n2Respectively representing the number of the historical environment data when a person exists in the room and the number of the historical environment data when no person exists in the room, wherein N is N1+n2；
Representing features of the environment when a person is present in the room
The conditional probability of (a) of (b),
representing features of the environment when nobody is in the room
J ═ 1,2,3), the expression is as follows:
wherein muiFor corresponding environmental characteristics
The average value of the environmental features of (1), namely the average value of the corresponding environmental features in the historical environmental data; sigmaiFor corresponding environmental characteristics
I.e. variance is calculated for the corresponding environmental eigenvalue in the historical environmental data.
Parameter mu in the naive Bayes modeli、σiAfter the historical environmental data are determined, a well-learned machine learning model is obtained, and each sensor is used for detectingAnd measuring current environment data in the room, substituting the environment characteristic value in the current environment data into the learned naive Bayes model to obtain the probability of the person in the room at the moment corresponding to the current environment data, and judging whether the room at the moment corresponding to the current environment data is the person according to the probability of the person. In one embodiment, a probability threshold (e.g., 0.75) of the probability of the presence of a person is set for comparison with the calculated probability of the presence of a person, and if the calculated probability of the presence of a person is greater than the probability threshold, it is determined that a person is present, and if the calculated probability of the presence of a person is not greater than the probability threshold, it is determined that no person is present.
After judging that people exist in the room at the moment corresponding to the current environment data, updating the current environment data into historical environment data according to the judgment result, and updating the parameter mu in the machine learning model according to the updated historical environment datai、σiTherefore, the naive Bayes model can predict the probability of someone in the room more accurately along with the increase of the collected historical environmental data, thereby further improving the reliability of the judging method.
Compared with the judging method in the prior art, the judging method has the advantages that the problem of method failure does not exist, and the judging reliability is high.
It should be noted that the environmental characteristics selected in this embodiment may be selected according to actual situations, for example, in the daytime, three environmental characteristics of carbon dioxide concentration, illumination intensity and noise decibel are selected, and in the nighttime, since the user in the room falls asleep, the environmental characteristic of noise decibel may not represent whether the room is occupied, so only two environmental characteristics of carbon dioxide concentration and illumination intensity may be selected, in this case,
the corresponding naive bayes model is also changed correspondingly to:
therefore, in this embodiment, the conditional probability of each environmental feature is obtained by using a naive bayes model, and then multiplied by a prior probability after a multiplication operation is performed, so as to calculate the probability of existence in a room, which is only an implementation manner. In another embodiment, the conditional probabilities of the environmental features may be logarithmized and then accumulated to obtain an accumulated value of the probability of presence and an accumulated value of the probability of absence, and then the accumulated values of the probability of presence and the probability of absence are compared to each other, and the larger one is used as the determination result.
The second method embodiment:
like the first embodiment of the method, the present embodiment also collects environmental data of people and nobody in the room over a period of time by using the sensors (C1, C2, C3) as historical environmental data of people and nobody in the room. The difference is that a machine learning model of the relationship between nobody in a room and all environment characteristic values is established according to historical environment data of the nobody and the nobody in the room, and the established model is as follows by taking a naive Bayes model as an example:
in the above model, P (Y ═ Y)2|X＝XT) Represented in the current environment data XTThe first method embodiment is referred to for the unmanned probability in the lower room, and other symbolic representations and calculation methods in the model, which are not described in detail in this embodiment.
Similar to the first embodiment of the method, the parameter μ in the machine learning model is described abovei、σiAfter the historical environmental data are determined, a learned naive Bayes model is obtained, current environmental data in a room are detected by using each sensor, environmental characteristic values in the current environmental data are substituted into the learned naive Bayes model, and a current ring is obtainedAnd judging whether people exist in the room at the moment corresponding to the current environment data according to the probability of people.
One way to determine whether a person is present in the room is to set a probability threshold (e.g., 0.85) of a certain probability of absence for comparison with the calculated probability of absence, determine that no person is present if the calculated probability of absence is greater than the naive bayes model, and determine that a person is present if the calculated probability of absence is not greater than the naive bayes model.
The method comprises the steps that when a person is judged to be in a room at the moment corresponding to the current environment data, the current environment data are updated to historical environment data, and therefore the machine learning model is updated.
The third method embodiment:
the embodiment provides a method for judging the state of people in a room, which is a combined state judging method based on various sensors, and the method for judging the state of people in the room improves the accuracy of judgment of an environment sensor by using a machine learning technology to learn the behavior habits of users. The environment characteristics mainly adopted comprise carbon dioxide concentration, illumination intensity and noise decibel, and the environment data are measured by the sensor arranged in the first method embodiment. The judgment process is shown in fig. 1, and includes judgment of a manned mode and an unmanned mode, and as shown in fig. 1, the specific steps are as follows:
step (1), motion detection preliminary judgment:
in the manned mode, a 30-minute timer is started, and if an interrupt condition is triggered within 30 minutes, the timer is reset for 30 minutes; and (4) if the timer is up and the interrupt condition is not triggered, executing the step (2). As shown in fig. 1, the interrupt conditions include:
1) triggering a person signal when the infrared human body detector detects a person;
2) the control panel device of any intelligent device in the room is triggered manually;
3) detecting the movement of a human body by an intelligent camera in a room;
4) an infrared curtain detector detects the entry of a person.
And if the certain interruption condition is met, judging that people exist in the room, restarting timing according to the timing time, and re-judging the interruption condition.
And (2) performing secondary judgment by using the environment information (namely historical environment data and collected current environment data):
and judging whether the air box is in the unmanned mode at present (namely, no person in the room is in the unmanned mode, and someone in the room is in the manned mode) according to the environmental data (CO2 concentration, illumination intensity and noise decibel) detected by the air box. The behavior habits of the users are learned in an individualized mode through the machine learning model, and different judgment standards are established for each user. The machine learning model which can be adopted is a naive Bayes model for example, parameters in the historical environment data learning model are adopted, then the probability that a person is in a current room and the probability that the person is not in the current room are respectively calculated, and if the probability that the person is in the current room is judged to be in an unmanned state is higher, the mode is switched to an unmanned mode; and (4) if the probability of the person is judged to be larger, returning to the step (1). As shown in fig. 2, the training process of the model is as follows:
1) firstly, historical environment data of a user in a room and not in the room is collected to be used as training samples, then the prior probability of a manned mode and an unmanned mode is calculated, and Y is used1Representing manned mode, Y2Representing an unmanned mode, a total of N pieces of historical environment data are collected, wherein N1Bar of manned pattern data, n2Bar of unmanned pattern data, N ═ N1+n2. Then the prior probability calculation for presence and absence in the room is as follows:
2) x for each sample of historical environmental dataj＝{x1,x2,x3Y represents j ═ 1,2 … N. Wherein x1,x2,x3The data respectively represent the numerical values of carbon dioxide concentration, illumination intensity and noise decibel, and y represents the category of the sample data and is classified into human and unmanned. Assuming that the environmental data distribution follows GaussianDistribution, calculating the conditional probability of each environment characteristic respectively:
wherein, l ═ 1,2,3, muiIn the sample class YiIn (1), all xlIs determined by the average value of (a) of (b),
in the sample class YiIn (1), all xlThe variance of (c).
3) Using the trained parameter mui、
A determination is made for the current environmental parameter. Environmental parameter X at time TTFor the purpose of example only,
wherein
Respectively representing the values of carbon dioxide concentration, illumination intensity and noise decibel at the T moment. Based on the conditional independence assumption in naive Bayes theory, the category Y can be obtained according to the following formulaiConditional probabilities (i.e., probability of presence and probability of absence of presence in a room):
wherein the content of the first and second substances,
the conditional probability of the corresponding environmental characteristics is obtained.
4) And 3) calculating the probability of each category according to the step 3), and judging the current sample as the category with the maximum probability.
The logic of the above steps (1) to (2) is:
and (3) if all the interruption conditions in the step (1) are not met, judging whether people exist in the room at the moment corresponding to the current environment data according to the greater probability in the probability of people existence and the probability of no people in the step (2), restarting timing according to the timing time in the step (1) when people exist in the room, and judging the interruption conditions again.
Step (3), unmanned mode judgment:
when one of the following conditions is satisfied, the unmanned mode is switched to the manned mode.
1) Triggering a person signal when the infrared human body detector detects a person;
2) the control panel device of any intelligent device in the room is triggered manually;
3) detecting the movement of a human body by an intelligent camera in a room;
4) an infrared curtain detector detects that a person enters;
5) and detecting that the home wireless router has new access of the mobile phone.
The reliability and stability of the naive Bayes model in the step (2) are verified by adopting K-fold-cross verification as follows:
the environmental data are collected by utilizing an air box in the continuous time of 7 days and are used as historical environmental data, and 10080 training data are counted, wherein 6701 pieces of data are labeled by people, and 3379 pieces of data are not labeled by people.
Then, 10-fold cross validation is adopted, the training data are uniformly and randomly divided into 10 parts, and one part is selected as a test set every time and used for determining the probability of people and the probability of no people in the room through the step (2); the remaining 9 were used as training sets for historical environmental data. The performance of the model on the whole data set can be tested by carrying out 10 times of experiments, and the experimental results of the probability of human being or the probability of no human being in the room obtained by the method in the step (2) are shown as the following table:
| |
|
|
|
|
|
|
|
|
|
| 0.9454 | 0.9534 | 0.9583 | 0.9444 | 0.9524 | 0.9534 | 0.9544 | 0.9514 | 0.9504 | 0.9494 |
As can be seen from the experimental results, the judgment accuracy is about 95%, and the results are stable on the whole data set.
It should be noted that, in the present embodiment, X is used for each sample of the historical environmental dataj＝{x1,x2,x3Y represents wherein xlThe number and values of the features of (a) are not limited, and as another embodiment, features such as CO2 concentration, light intensity, maximum value, average value and standard deviation of noise decibels within the time interval of 1s may be added as needed, so that there are nine features, as shown in fig. 3, L _ MAX, L _ MEAN and L _ STD refer to the maximum value, average value and standard deviation of light intensity, C _ MAX, C _ MEAN and C _ STD refer to the maximum value, average value and standard deviation of CO2 concentration, N _ MAX, N _ MEAN and N _ STD refer to the maximum value, average value and standard deviation of noise decibels, respectively, and LABEL refers to a LABEL of the piece of history data, such as human or non-human. Thus xlThe number and value of the features can be selected according to specific needs.
In addition, in the present embodiment, the formula for determining the probability of presence and the probability of absence in the room is actually two formulas, which are equivalent to two machine learning models, one is a model for calculating the probability of presence, and the parameter obtained by training is μ1、σ1(ii) a The other is a model for calculating the probability of no person, and the parameter obtained by training is mu2、σ2。
The embodiment of the device is as follows:
the embodiment provides a device for determining a presence or absence state in a room, which includes a memory, a processor, and a computer program stored in the memory and running on the processor, where the processor is coupled to the memory, and when the processor executes the computer program, the determination methods in the three method embodiments are implemented, and details of the specific determination method are not repeated in this embodiment.
In the embodiment, the processor is connected with each sensor arranged in a room through collection and is used for acquiring environment data and storing the environment data in the memory, and corresponding historical environment data is acquired through the memory when parameters in the machine learning model need to be trained; the processor is also in control connection with a timer and is used for controlling the timer to start timing; the processor is also in communication connection with an infrared detector, a camera, a router and intelligent equipment which are arranged in the room, and is used for judging the interrupt condition.
After the processor judges whether a person is in the room by using one of the three method embodiments, the current state in the room is sent to the intelligent device (such as an intelligent mobile phone) carried by the user, so that the user can conveniently carry out remote control on the intelligent household appliances in the room.
In addition, the processor in this embodiment may be a computer, a microprocessor such as an ARM, or a programmable chip such as an FPGA, a DSP, or the like. In order to reduce the workload of one processor, as another alternative, a plurality of processors may be used to share the work content, such as a main processor and an auxiliary processor, two processors are communicatively connected to each other, and the two processors cooperate to implement the determination method in the above three method embodiments.
Claims (10)
1. A method for judging the state of people in a room is unmanned, which is characterized by comprising the following steps:
storing historical environmental data of people and no people in a room, wherein the historical environmental data at least comprises data of one environmental characteristic, and the environmental characteristic comprises: carbon dioxide concentration, illumination intensity and noise decibel;
the method comprises the following steps of collecting current environment data in a room, and calculating the probability of people and the probability of no people in the room at the moment corresponding to the current environment data, wherein the calculation mode of the probability of people is as follows: taking the historical environment data of people as samples, respectively calculating the conditional probability of a certain environment characteristic in the current environment data when people exist in a room, and calculating the obtained conditional probability of each environment characteristic to obtain the probability of people in the room at the moment corresponding to the current environment data, so that the probability of people is positively correlated with each conditional probability;
the calculation mode of the unmanned probability is as follows: taking the historical environment data of nobody as a sample, respectively calculating the conditional probability of a certain environment characteristic in the current environment data when no person exists in a room, and calculating the obtained conditional probability of each environment characteristic to obtain the nobody probability in the room at the moment corresponding to the current environment data, so that the nobody probability is positively correlated with each conditional probability;
and judging whether the room is occupied at the moment corresponding to the current environment data according to the greater probability of the probability of being occupied and the probability of being unoccupied.
2. The method according to claim 1, wherein the conditional probability of the current environmental data for a certain environmental characteristic is as follows:
in the formula (I), the compound is shown in the specification,
for environmental characteristics when a person is present in a room
The conditional probability of (a) of (b),
wherein
Respectively representing the values of carbon dioxide concentration, illumination intensity and noise decibel at the moment T, wherein the moment T is the moment corresponding to the current environmental data; mu.s1Corresponding environmental characteristics in historical environmental data for people in a room
Average value of (a)1Corresponding environmental characteristics in historical environmental data for people in a room
The variance obtained from the numerical values of (a);
the conditional probability under a certain environmental characteristic in the current environmental data when the room is unmanned is as follows:
in the formula (I), the compound is shown in the specification,
for environmental characteristics when no person is in the room
Conditional probability of (d), mu2Corresponding environmental characteristics in historical environmental data for no person in a room
Average value of (a)2Corresponding environmental characteristics in historical environmental data for no person in a room
The variance of the value of (a).
3. The method for determining a presence or absence in a room according to claim 1, further comprising: when the current environment data corresponds to the probability of the person in the room at the moment, the operation is performed, namely, after the conditional probabilities are multiplied, the conditional probabilities are multiplied by the prior probability when the person is in the room in the historical environment data of the person and the nobody in the room;
and when the probability of no person in the room at the moment corresponding to the current environment data is obtained, calculating to multiply the conditional probabilities after being multiplied, and then multiplying the conditional probabilities by the prior probability of the absence of the person in the room in the historical environment data of the presence of the person in the room and the absence of the person in the room.
4. A method for judging the state of people in a room is unmanned, which is characterized by comprising the following steps:
storing historical environmental data of a person in a room, the historical environmental data at least comprising data of one environmental characteristic, the environmental characteristic comprises: carbon dioxide concentration, illumination intensity and noise decibel;
the method comprises the following steps of collecting current environment data in a room, and calculating the probability of people in the room at the moment corresponding to the current environment data, wherein the calculation mode is as follows: taking the historical environment data of people as samples, respectively calculating the conditional probability of a certain environment characteristic in the current environment data when people exist in a room, and calculating the obtained conditional probability of each environment characteristic to obtain the probability of people in the room at the moment corresponding to the current environment data, so that the probability of people is positively correlated with each conditional probability;
and comparing the probability of the person with a set probability threshold, and judging whether the person exists in the room at the moment corresponding to the current environment data according to the comparison result.
5. The method according to claim 4, wherein the conditional probability of the current environmental data for an environmental feature is as follows:
in the formula (I), the compound is shown in the specification,
for environmental characteristics when a person is present in a room
The conditional probability of (a) of (b),
wherein
Respectively representing the values of carbon dioxide concentration, illumination intensity and noise decibel at the moment T, wherein the moment T is the moment corresponding to the current environmental data; mu.s1History ring for people in roomCorresponding environmental features in environmental data
Average value of (a)1Corresponding environmental characteristics in historical environmental data for people in a room
The variance of the value of (a).
6. The method for determining the presence or absence of a person in a room according to claim 4, further comprising: and storing the historical environment data of no person in the room, wherein the operation is to multiply the conditional probabilities after being multiplied, and then multiply the prior probabilities of persons in the room and persons in the room in the historical environment data of no person, so as to obtain the probability of persons in the room at the moment corresponding to the current environment data.
7. A method for judging the state of people in a room is unmanned, which is characterized by comprising the following steps:
storing historical environmental data of the room when no person is present, wherein the historical environmental data at least comprises data of environmental characteristics, and the environmental characteristics comprise: carbon dioxide concentration, illumination intensity and noise decibel;
the method comprises the following steps of collecting current environment data in a room, and calculating the probability of no person in the room at the moment corresponding to the current environment data, wherein the calculation mode is as follows: taking the historical environment data of nobody as a sample, respectively calculating the conditional probability of a certain environment characteristic in the current environment data when no person exists in a room, and calculating the obtained conditional probability of each environment characteristic to obtain the nobody probability in the room at the moment corresponding to the current environment data, so that the nobody probability is positively correlated with each conditional probability;
and comparing the unmanned probability with a set probability threshold, and judging whether people exist in the room at the moment corresponding to the current environment data according to the comparison result.
8. The method according to claim 7, wherein the conditional probability of the current environmental data for an environmental feature is as follows:
in the formula (I), the compound is shown in the specification,
for environmental characteristics when no person is in the room
The conditional probability of (a) of (b),
wherein
Respectively representing the values of carbon dioxide concentration, illumination intensity and noise decibel at the moment T, wherein the moment T is the moment corresponding to the current environmental data; mu.s2Corresponding environmental characteristics in historical environmental data for no person in a room
Average value of (a)2Corresponding environmental characteristics in historical environmental data for no person in a room
The variance of the value of (a).
9. The method for determining a presence or absence in a room according to claim 7, further comprising: and storing the historical environment data of the presence of the person in the room, and multiplying the conditional probabilities by the prior probability of the absence of the person in the room in the historical environment data of the presence of the person and the absence of the person in the room to obtain the absence probability of the person in the room at the moment corresponding to the current environment data.
10. A device for determining an occupancy state in a room, comprising a memory, a processor and a computer program stored in the memory and running on the processor, wherein the processor is coupled to the memory, and wherein the processor executes the computer program to implement the method for determining an occupancy state in a room according to any one of claims 1 to 9.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910754512.XA CN112051621B (en) | 2019-08-15 | 2019-08-15 | Method and device for judging whether room is occupied or not |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910754512.XA CN112051621B (en) | 2019-08-15 | 2019-08-15 | Method and device for judging whether room is occupied or not |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112051621A true CN112051621A (en) | 2020-12-08 |
| CN112051621B CN112051621B (en) | 2024-03-15 |
Family
ID=73609583
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910754512.XA Active CN112051621B (en) | 2019-08-15 | 2019-08-15 | Method and device for judging whether room is occupied or not |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112051621B (en) |
Citations (19)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US4514729A (en) * | 1982-08-16 | 1985-04-30 | Szarka Jay R | Environmental control system and method |
| JP2005172377A (en) * | 2003-12-12 | 2005-06-30 | Mitsubishi Electric Corp | Air conditioner having human body detecting sensor, and human body detecting method for air conditioner having human body detecting sensor |
| JP2009266502A (en) * | 2008-04-24 | 2009-11-12 | Panasonic Electric Works Co Ltd | Lighting system |
| CN101661268A (en) * | 2008-08-25 | 2010-03-03 | 李海青 | Method for integrally controlling living room and human body count detection energy-saving control system |
| JP2010256045A (en) * | 2009-04-21 | 2010-11-11 | Taisei Corp | Wide range/high accuracy human body detection sensor |
| JP2012233890A (en) * | 2011-04-21 | 2012-11-29 | Takenaka Komuten Co Ltd | Presence/absence management system |
| FR2985070A1 (en) * | 2011-12-21 | 2013-06-28 | Orme | System for detecting fall of e.g. person in retirement home, has programmed analysis unit to acquire images taken by sensors in synchronous manner, and to detect fall of person in scene from time sequence of pair of images of scene |
| US8630741B1 (en) * | 2012-09-30 | 2014-01-14 | Nest Labs, Inc. | Automated presence detection and presence-related control within an intelligent controller |
| WO2014051632A1 (en) * | 2012-09-30 | 2014-04-03 | Nest Labs, Inc. | Automated presence detection and presence-related control within an intelligent controller |
| CN103837906A (en) * | 2014-03-13 | 2014-06-04 | 三和智控(北京)系统集成有限公司 | Method for detecting whether people exist in room or not |
| JP2014190724A (en) * | 2013-03-26 | 2014-10-06 | Kddi Corp | Space state determination device |
| FR3005367A1 (en) * | 2013-05-06 | 2014-11-07 | Vence Innovation | COMPACT DETECTOR OF HUMAN PRESENCE |
| US20150380013A1 (en) * | 2014-06-30 | 2015-12-31 | Rajeev Conrad Nongpiur | Learning algorithm to detect human presence in indoor environments from acoustic signals |
| CN106772656A (en) * | 2015-11-19 | 2017-05-31 | 上海理工大学 | A kind of indoor human body detection method based on infrared sensor array |
| CN106842356A (en) * | 2017-01-17 | 2017-06-13 | 云丁网络技术(北京)有限公司 | There is nobody detection method and its detecting system a kind of interior |
| WO2018121330A1 (en) * | 2016-12-29 | 2018-07-05 | 阿里巴巴集团控股有限公司 | Detection method, and related apparatus and system |
| CN108256283A (en) * | 2016-12-28 | 2018-07-06 | 中国移动通信有限公司研究院 | A kind of occupancy recognition methods and device |
| CN109030715A (en) * | 2017-06-09 | 2018-12-18 | 精英电脑股份有限公司 | The method of indoor human body detecting |
| US10169975B1 (en) * | 2017-11-14 | 2019-01-01 | Vi-Enterprises, Llc | Detecting life by means of CO2 in an enclosed volume |
-
2019
- 2019-08-15 CN CN201910754512.XA patent/CN112051621B/en active Active
Patent Citations (19)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US4514729A (en) * | 1982-08-16 | 1985-04-30 | Szarka Jay R | Environmental control system and method |
| JP2005172377A (en) * | 2003-12-12 | 2005-06-30 | Mitsubishi Electric Corp | Air conditioner having human body detecting sensor, and human body detecting method for air conditioner having human body detecting sensor |
| JP2009266502A (en) * | 2008-04-24 | 2009-11-12 | Panasonic Electric Works Co Ltd | Lighting system |
| CN101661268A (en) * | 2008-08-25 | 2010-03-03 | 李海青 | Method for integrally controlling living room and human body count detection energy-saving control system |
| JP2010256045A (en) * | 2009-04-21 | 2010-11-11 | Taisei Corp | Wide range/high accuracy human body detection sensor |
| JP2012233890A (en) * | 2011-04-21 | 2012-11-29 | Takenaka Komuten Co Ltd | Presence/absence management system |
| FR2985070A1 (en) * | 2011-12-21 | 2013-06-28 | Orme | System for detecting fall of e.g. person in retirement home, has programmed analysis unit to acquire images taken by sensors in synchronous manner, and to detect fall of person in scene from time sequence of pair of images of scene |
| WO2014051632A1 (en) * | 2012-09-30 | 2014-04-03 | Nest Labs, Inc. | Automated presence detection and presence-related control within an intelligent controller |
| US8630741B1 (en) * | 2012-09-30 | 2014-01-14 | Nest Labs, Inc. | Automated presence detection and presence-related control within an intelligent controller |
| JP2014190724A (en) * | 2013-03-26 | 2014-10-06 | Kddi Corp | Space state determination device |
| FR3005367A1 (en) * | 2013-05-06 | 2014-11-07 | Vence Innovation | COMPACT DETECTOR OF HUMAN PRESENCE |
| CN103837906A (en) * | 2014-03-13 | 2014-06-04 | 三和智控(北京)系统集成有限公司 | Method for detecting whether people exist in room or not |
| US20150380013A1 (en) * | 2014-06-30 | 2015-12-31 | Rajeev Conrad Nongpiur | Learning algorithm to detect human presence in indoor environments from acoustic signals |
| CN106772656A (en) * | 2015-11-19 | 2017-05-31 | 上海理工大学 | A kind of indoor human body detection method based on infrared sensor array |
| CN108256283A (en) * | 2016-12-28 | 2018-07-06 | 中国移动通信有限公司研究院 | A kind of occupancy recognition methods and device |
| WO2018121330A1 (en) * | 2016-12-29 | 2018-07-05 | 阿里巴巴集团控股有限公司 | Detection method, and related apparatus and system |
| CN106842356A (en) * | 2017-01-17 | 2017-06-13 | 云丁网络技术(北京)有限公司 | There is nobody detection method and its detecting system a kind of interior |
| CN109030715A (en) * | 2017-06-09 | 2018-12-18 | 精英电脑股份有限公司 | The method of indoor human body detecting |
| US10169975B1 (en) * | 2017-11-14 | 2019-01-01 | Vi-Enterprises, Llc | Detecting life by means of CO2 in an enclosed volume |
Non-Patent Citations (4)
| Title |
|---|
| 党小超;邓琦研;郝占军;: "基于30°角同心圆环形取样的室内人员检测方法", no. 04, pages 56 - 61 * |
| 刘德峰;袁锁中;: "基于二氧化碳测量的室内人数估计算法", no. 02, pages 43 - 47 * |
| 王闯;燕达;孙红三;丰晓航;江亿;: "室内环境控制相关的人员动作描述方法", no. 10, pages 199 - 211 * |
| 韦春玲;王步飞;: "基于WLAN接收信号强度特征的室内活动识别", no. 05, pages 1326 - 1330 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112051621B (en) | 2024-03-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10535349B2 (en) | Controlling connected devices using a relationship graph | |
| EP3779306B1 (en) | Aroma releasing system | |
| JP2017528788A (en) | Air quality alarm system and method | |
| CN110131846B (en) | Intelligent air conditioner control method and air conditioner | |
| EP2985540B1 (en) | Air environment regulating system, and controlling device | |
| CN108427314A (en) | Intelligent home furnishing control method, device and computer readable storage medium | |
| CN106556111A (en) | Household electric appliance control method, device, server and air-conditioner | |
| US20230314037A1 (en) | Integrated air quality control system | |
| Kim et al. | Improved occupancy detection accuracy using PIR and door sensors for a smart thermostat | |
| Amayri et al. | Decision tree and parametrized classifier for estimating occupancy in energy management | |
| US11734932B2 (en) | State and event monitoring | |
| Taştan | An IoT based air quality measurement and warning system for ambient assisted living | |
| CN111433801A (en) | Data generation device, data generation method, data generation program, and sensor device | |
| US11544924B1 (en) | Investigation system for finding lost objects | |
| CN114063572A (en) | Non-sensing intelligent device control method, electronic device and control system | |
| CN112051621B (en) | Method and device for judging whether room is occupied or not | |
| CN110941189A (en) | Intelligent household system and control method thereof and readable storage medium | |
| CN208475570U (en) | Air purifier booting-self controller switch and its system | |
| CN110209059A (en) | There are unmanned detection method, device, electronic equipment and system | |
| US11624519B2 (en) | Carbon monoxide purge system for a property | |
| US11719677B1 (en) | Air quality sensors | |
| Zhang et al. | Inferring building occupancy based on statistical modeling of multi-sensor data | |
| WO2017149324A1 (en) | Occupancy sensing | |
| CN110195926A (en) | Air purifier booting-self controller switch and its system | |
| CN106765935B (en) | Energy-saving data acquisition method and device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |