Steady state prediction method and device for smell sensor after power-off recovery

Abstract

The invention discloses a steady state prediction method and device after power-off recovery of a smell sensor. The method comprises the following steps: collecting data information after power-off recovery of the smell sensor for feature extraction; the steady-state value information of the smell sensor before power failure is obtained, the steady-state value of the sensor is predicted by constructing the regression model in combination with the characteristics of the data information, and the refrigerator peculiar smell state is output, so that the time for recovering the balance state of the smell sensor is shortened to a great extent; a guarantee is provided for accurately predicting the peculiar smell in the refrigerator and the use experience of a user after the sensor recovers power supply. The method further comprises the steps of collecting and storing data information of the sensor in real time, and judging whether the sensor is in a balanced state or not. And when the sensor is in a balanced state, the steady state value is stored, so that the working state of the sensor can be known in time, and a reference is provided for keeping the sensor in the balanced state.

Description

Technical Field

[0001] The present invention relates to the technical field of sensors, and in particular to a method and device for predicting the steady state of an odor sensor after power failure and recovery. Background Art

[0002] Odor sensors have a wide range of applications in modern life, including but not limited to monitoring odors inside food preservation equipment (such as refrigerators). The sensor can detect odor substances in the air to provide information about the status of food inside the storage device.

[0003] However, sensors face specific technical issues in certain situations. One common situation is that after the sensor is connected to the power supply, it takes a period of time to stabilize and return to normal operation, which generally takes 20 minutes to 2 hours. During this recovery period, the sensor may not be able to accurately monitor or predict the changes in odor inside the food preservation equipment, which may lead to a poor user experience or inaccurate data output.

[0004] For a device like a refrigerator, the odor inside may change in many ways, including but not limited to the following: (1) Unchanged: When the refrigerator loses power unexpectedly and the power is restored within a short period of time, the odor detected by the sensor usually does not change significantly; (2) Deteriorated: If perishable food is stored inside the refrigerator and the power is cut off for a long time, the odor may deteriorate sharply after the power is restored; (3) Improved: When the user cleans the refrigerator, the odor may improve.

[0005] In these cases, the sensor's equilibrium state may be disrupted, so a method is needed to predict or correct the sensor data to provide more accurate information for user reference. Summary of the invention

[0006] The purpose of the present invention is to provide a method and device for predicting the steady state of an odor sensor after power failure and recovery, so as to solve the problem in the prior art that the measurement data is inaccurate due to power failure of the sensor, which affects the user experience.

[0007] To achieve one of the above-mentioned purposes of the invention, an embodiment of the present invention provides a method for predicting a steady state of an odor sensor after power failure and recovery, the method comprising:

[0008] Collect data information after the odor sensor is powered off and restored for feature extraction;

[0009] The steady-state value information of the odor sensor before power failure is obtained, and a regression model is constructed based on the characteristics of the data information to predict the steady-state value of the sensor, and the refrigerator odor status is output.

[0010] As a further improvement of an embodiment of the present invention, the method further includes: collecting and storing data information of the odor sensor in real time to determine whether the sensor is in a balanced state;

[0011] When the odor sensor is in a balanced state, the steady-state value of the sensor is stored.

[0012] As a further improvement of an embodiment of the present invention, the method further includes: defining a time window win_size=n for data collection, and the data set within the time window is expressed as:

[0013] X=(x1,x2,...,x n}

[0014] Among them, n is the time, which means that the latest n seconds of sensor data are collected;

[0015] The mean of the data set is expressed as:

[0016]

[0017] The variance within the current time window is expressed as:

[0018]

[0019] A difference sequence of a data set X is expressed as:

[0020]

[0021] Among them, the number of 0s in ΔX is recorded as cnt;

[0022] Output the data set and draw a line graph. The slope of the line connecting the head end data and the end end data is expressed as:

[0023]

[0024] As a further improvement of an embodiment of the present invention, the method further includes: when the sensor in the time window is in a balanced state, the variance var changes less than 0.05 compared with the previous time, and |k|≤0.17;

[0025] If the sensor is in a balanced state, the odor of the gas in the refrigerator is predicted based on the sensor data in the current time window, and the sensor steady-state value is stored according to a fixed time interval, and the sensor balance state is no longer determined before the next power failure.

[0026] As a further improvement of an embodiment of the present invention, the method further includes: when the sensor fails to reach a balanced state after power failure, feature extraction is performed on the data information of the sensor in the current time window, including:

[0027] Extract the maximum value X of the data set X within the current time window max , minimum value X min , median, 25% quantile, 75% quantile, and calculate the mean, range, autocorrelation coefficient ρ, variance var, the number of times cnt that 0 appears after splitting two adjacent data, and the slope k of the line connecting the head end data and the end end data to obtain the characteristics of the sensor data information;

[0028] The range is expressed as:

[0029] range=X max -X min

[0030] The autocorrelation coefficient is expressed as:

[0031]

[0032] As a further improvement of an embodiment of the present invention, the method further includes: the “predicting the steady-state value of the sensor by building a regression model in combination with the characteristics of the data information” includes,

[0033] Define that M models are used in a prediction task, then one of the data x i The predicted value of is expressed as:

[0034]

[0035] Among them, b represents the basic model, β represents the weight of each basic model in the total prediction result, and γ represents the parameter of each basic model;

[0036] Input the characteristics of the sensor data information, the steady-state value before power failure, and the power-on time after power failure recovery into the regression model to predict the true measurement value of the air;

[0037] Then the optimization goal of the predicted value is:

[0038]

[0039] Among them, N is the number of training samples and L is the regression model loss function.

[0040] To achieve one of the above-mentioned purposes, an embodiment of the present invention provides a steady-state prediction device for an odor sensor after power failure and recovery, the device comprising a feature extraction module and a steady-state prediction module.

[0041] The feature extraction module is used to collect data information of the odor sensor after power failure and recovery to perform feature extraction;

[0042] The steady-state prediction module is used to obtain the steady-state value information of the odor sensor before power failure, build a regression model based on the characteristics of the data information to predict the steady-state value of the sensor, and output the odor status of the refrigerator.

[0043] As a further improvement of an embodiment of the present invention, the device further includes: a steady-state judgment module and a storage module.

[0044] The steady-state judgment module is used to collect and store data information of the odor sensor in real time to determine whether the sensor is in a balanced state;

[0045] The storage module is used to store the steady-state value of the sensor when the odor sensor is in a balanced state.

[0046] To achieve one of the above-mentioned purposes of the invention, one embodiment of the present invention provides an electronic device, including a memory and a processor, wherein the memory stores a computer program that can be run on the processor, and when the program is executed on the processor, the steps in the steady-state prediction method after the odor sensor is powered off and restored are implemented as described above.

[0047] To achieve one of the above-mentioned purposes of the invention, one embodiment of the present invention provides a storage medium, wherein the storage medium stores a computer program, and is characterized in that when the computer program is executed by a processor, the steps in the steady-state prediction method after the odor sensor is powered off and restored are implemented as described above.

[0048] Compared with the prior art, the method and device for predicting the steady state of the odor sensor after power failure provided by the present invention collects data information and performs feature extraction after the sensor is powered off, and builds a regression model to predict the steady state value of the sensor in combination with the steady state value information before the sensor is powered off, which greatly reduces the time it takes for the odor sensor to recover to a balanced state, and provides a guarantee for accurately predicting the odor in the refrigerator and the user's experience after the sensor is powered on again. The method of the present invention also includes real-time collection and storage of sensor data information, and determining whether the sensor is in a balanced state. When the sensor is in a balanced state, the steady state value will be stored, which helps to timely understand the working state of the sensor and provides a reference for the sensor to maintain a balanced state. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 It is a flow chart of the steady-state prediction method after power failure recovery of the odor sensor of the present invention.

[0050] Figure 2 It is a structural schematic diagram of the steady-state prediction device after the odor sensor of the present invention is powered off and restored. DETAILED DESCRIPTION

[0051] The present invention will be described in detail below in conjunction with the specific embodiments shown in the accompanying drawings. However, these embodiments do not limit the present invention, and any structural, methodological, or functional changes made by a person skilled in the art based on these embodiments are all within the scope of protection of the present invention.

[0052] The embodiments of the present invention are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and cannot be understood as limiting the present invention.

[0053] In the first embodiment of the present invention, the present invention provides a method for predicting the steady state of an odor sensor after power failure recovery, such as Figure 1 As shown, the method includes:

[0054] S1: Collect data information after the odor sensor is powered off and restored for feature extraction;

[0055] S2: Obtain the steady-state value information of the odor sensor before power failure, build a regression model based on the characteristics of the data information to predict the steady-state value of the sensor, and output the refrigerator odor status.

[0056] In a specific embodiment of the present invention, data information of the odor sensor is collected and stored in real time, and it is determined whether the sensor is in a balanced state; when the odor sensor is in a balanced state, the steady-state value of the sensor is stored.

[0057] Specifically, on the one hand, the system obtains the odor data measured by the sensor by connecting to the sensor. This data may include information such as the intensity, composition, and change trend of the odor. The system will collect data regularly or as needed to ensure that the latest sensor data is obtained. By using a storage device, the system will save the collected data to ensure long-term storage and subsequent analysis of the data. The storage method can be to store the data directly on the local device, or to upload the data to a cloud server for storage through a network connection.

[0058] On the other hand, the system will judge the balance state of the sensor. The balance state means that the sensor can stably measure and output accurate data within the normal working range. The system will analyze and compare the collected sensor data according to the preset judgment conditions and algorithms to determine whether the sensor is in a balanced state. These judgment conditions may include data stability, fluctuation range, change trend, etc. When the sensor is judged to be in a balanced state, the system will store the steady-state value of the sensor. The steady-state value refers to the stable value of the data measured by the sensor in a balanced state. The system will save these steady-state values ​​in the storage device for subsequent use. By providing a reference benchmark for comparison and analysis with the subsequently collected data, the working state of the sensor and the change of odor can be judged.

[0059] In a specific embodiment of the present invention, a time window win_size=n is defined for data collection, and the data set within the time window is expressed as:

[0060] X=(x1,x2,...,x n}

[0061] Among them, n is the time, which means that the latest n seconds of sensor data are collected.

[0062] The mean of the data set is expressed as:

[0063]

[0064] It should be noted that a time window refers to a fixed-length time period selected in time series data. It can be a continuous time period or a discrete time point set. In the test environment used in the present invention, a 100-second time window is used.

[0065] Calculating the variance within a time window can provide information about the degree of variation and volatility of the data. Variance is a statistic that measures the degree of dispersion of data, which indicates the degree of difference between a data point and its mean. The variance within the current time window is expressed as:

[0066]

[0067] A difference sequence of a data set X is expressed as:

[0068]

[0069] The number of 0s in ΔX is recorded as cnt; calculating the number of 0s in the difference sequence can provide information about the stability and change of the data. The number of 0s in the difference sequence represents the number of change points in the data within the time window, which can be used to determine the stability of the data and the frequency of change.

[0070] Calculating the slope of the line connecting the head data and the end data can provide information about the trend and rate of change of the data. The slope represents the speed of change of the data within the time window, which can be used to determine the growth or decrease trend of the data and the rate of change. Output the data set and draw a line chart. The slope of the line connecting the head data and the end data is expressed as:

[0071]

[0072] In a specific embodiment of the present invention, when the sensor in the time window is in a balanced state, the variance var changes less than 0.05 compared with the previous time. and |k|≤0.17;

[0073] If the sensor is in a balanced state, the odor of the gas in the refrigerator is predicted based on the sensor data in the current time window, and the sensor steady-state value is stored according to a fixed time interval (such as half an hour). The sensor balance state will no longer be judged before the next power failure, and the odor judgment will be made directly based on the sensor measurement value.

[0074] In a specific embodiment of the present invention, when the sensor fails to reach a balanced state after power failure, feature extraction is performed on the data information of the sensor in the current time window, specifically,

[0075] Extract the maximum value X of the data set X within the current time window max , minimum value X min , median, 25% quantile, 75% quantile, and calculate the mean, range, autocorrelation coefficient ρ, variance var, the number of times cnt when two adjacent data are split, and the slope k of the line connecting the head data and the end data to obtain the characteristics of the sensor data information;

[0076] The degree and range of variation of the measured data are measured by calculating the range. The calculation formula of the range is expressed as:

[0077] range=X max -X min

[0078] The correlation between each measurement data in the time series data is measured by calculating the autocorrelation coefficient. The calculation formula of the autocorrelation coefficient is expressed as:

[0079]

[0080] In a specific embodiment of the present invention, a regression model is constructed by combining the characteristics of data information to predict the steady-state value of the sensor, specifically,

[0081] In the present invention, the computing power problem of the refrigerator system is taken into consideration, so the GBDT algorithm is used to construct the regression model, wherein the GBDT decision tree algorithm is an integrated learning method that improves the prediction performance by iteratively training multiple decision tree models. It is a machine learning algorithm based on gradient boosting, which continuously and iteratively trains weak classifiers (decision trees) and combines them into a strong classifier.

[0082] Define that M models are used in a prediction task, then one of the data x i The predicted value of is expressed as:

[0083]

[0084] Among them, b represents the basic model, β represents the weight of each basic model in the total prediction result, and γ represents the parameter of each basic model;

[0085] The characteristics of the sensor data information, the steady-state value before the power outage, and the power-on time after the power outage are input into the regression model to predict the true measurement value of the air;

[0086] It should be noted that since the sensor is in a steady-state recovery process after power is restored, a reference benchmark is provided for comparison and analysis with subsequently collected data so as to restore to a balanced state as soon as possible.

[0087] The optimization goal of the predicted value is:

[0088]

[0089] Among them, N is the number of training samples and L is the regression model loss function.

[0090] In the second embodiment of the present invention, the present invention provides a steady-state prediction device for an odor sensor after power failure recovery, comprising a feature extraction module and a steady-state prediction module.

[0091] A feature extraction module is used to collect data information of the odor sensor after power failure and recovery for feature extraction;

[0092] The steady-state prediction module is used to obtain the steady-state value information of the odor sensor before power failure, build a regression model based on the characteristics of the data information to predict the steady-state value of the sensor, and output the odor status of the refrigerator.

[0093] In a specific embodiment of the present invention, it also includes a steady-state judgment module and a storage module.

[0094] A steady-state judgment module is used to collect and store data information of the odor sensor in real time and judge whether the sensor is in a balanced state;

[0095] The storage module is used to store the steady-state value of the sensor when the odor sensor is in a balanced state.

[0096] In the third embodiment of the present invention, the present invention provides an electronic device, including a memory and a processor, wherein the memory stores a computer program that can be run on the processor, and when the program is executed on the processor, the steps in the steady-state prediction method after the odor sensor is powered off and restored as described above are implemented.

[0097] In a fourth embodiment of the present invention, the present invention provides a storage medium storing a computer program, and when the computer program is executed by a processor, the steps in the steady-state prediction method after the odor sensor is powered off and restored are implemented as described above.

[0098] In summary, the method and device for predicting the steady state of the odor sensor after power failure and restoration provided by the present invention, after the sensor is powered off, collects data information and performs feature extraction, combines the steady-state value information of the sensor before power failure, and constructs a regression model to predict the steady-state value of the sensor, which greatly reduces the time for the odor sensor to restore the equilibrium state, and provides a guarantee for accurately predicting the odor in the refrigerator and the user's experience after the sensor restores power. The method of the present invention also includes real-time collection and storage of sensor data information, and judging whether the sensor is in a balanced state. When the sensor is in a balanced state, the steady-state value will be stored, which helps to timely understand the working state of the sensor and provides a reference for the sensor to maintain a balanced state.

[0099] It should be understood that although this specification is described according to implementation modes, not every implementation mode contains only one independent technical solution. This description of the specification is only for the sake of clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each implementation mode may also be appropriately combined to form other implementation modes that can be understood by those skilled in the art.

[0100] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the modules described above can refer to the corresponding process in the aforementioned method implementation, and will not be repeated here.

[0101] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical modules, that is, they may be located in one place or distributed on multiple network modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the present implementation scheme.

[0102] In addition, each functional module in each embodiment of the present application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module. The above integrated modules can be implemented in the form of hardware or in the form of hardware plus software functional modules.

[0103] The above-mentioned integrated module implemented in the form of a software function module can be stored in a computer-readable storage medium. The above-mentioned software function module is stored in a storage medium, including a number of instructions for a computer system (which can be a personal computer, a server, or a network system, etc.) or a processor to perform some steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (Read-Only Memory, ROM), random access memory (Random Access Memory, RAM), disk or optical disk and other media that can store program codes.

[0104] Finally, it should be noted that the above implementation modes are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned implementation modes, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned implementation modes, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various implementation modes of the present application.

Claims

1. A method for predicting the steady state of an odor sensor after power failure recovery, characterized in that: include: Collect data information after the odor sensor is powered off and restored for feature extraction; The steady-state value information of the odor sensor before power failure is obtained, and a regression model is constructed based on the characteristics of the data information to predict the steady-state value of the sensor, and the refrigerator odor status is output.

2. The method for predicting the steady state of the odor sensor after power failure recovery according to claim 1 is characterized in that: Also includes, Collect and store data information of the odor sensor in real time to determine whether the sensor is in a balanced state; When the odor sensor is in a balanced state, the steady-state value of the sensor is stored.

3. The method for predicting the steady state of the odor sensor after power failure recovery according to claim 2 is characterized in that: Define the data collection time window win_size = n, then the data set in this time window is expressed as: X={x1,x2,…,x n } Among them, n is the time, which means that the latest n seconds of sensor data are collected; The mean of the data set is expressed as: The variance within the current time window is expressed as: A difference sequence of a data set X is expressed as: Among them, the number of 0s in ΔX is recorded as cnt; Output the data set and draw a line graph. The slope of the line connecting the head end data and the end end data is expressed as:

4. The method for predicting the steady state of the odor sensor after power failure recovery according to claim 3 is characterized in that: When the sensor in the time window is in a balanced state, the variance var changes less than 0.05 compared with the last time. and |k|≤0.17; If the sensor is in a balanced state, the odor of the gas in the refrigerator is predicted based on the sensor data in the current time window, and the sensor steady-state value is stored according to a fixed time interval, and the sensor balance state is no longer determined before the next power failure.

5. The method for predicting the steady state of the odor sensor after power failure recovery according to claim 4 is characterized in that: When the sensor fails to reach a balanced state after power failure, feature extraction is performed on the data information of the sensor in the current time window, including: Extract the maximum value X of the data set X within the current time window max , minimum value X min , median, 25% quantile, 75% quantile, and calculate the mean, range, autocorrelation coefficient ρ, variance var, the number of times cnt that 0 appears after splitting two adjacent data, and the slope k of the line connecting the head end data and the end end data to obtain the characteristics of the sensor data information; The range is expressed as: range=X max -X min The autocorrelation coefficient is expressed as:

6. The method for predicting the steady state of the odor sensor after power failure recovery according to claim 5, characterized in that: The “predicting the steady-state value of the sensor by building a regression model combining the features of the data information” includes: Define that M models are used in a prediction task, then one of the data x i The predicted value of is expressed as: Among them, b represents the basic model, β represents the weight of each basic model in the total prediction result, and γ represents the parameter of each basic model; Input the characteristics of the sensor data information, the steady-state value before power failure, and the power-on time after power failure recovery into the regression model to predict the true measurement value of the air; Then the optimization goal of the predicted value is: Among them, N is the number of training samples and L is the regression model loss function.

7. A steady-state prediction device for an odor sensor after power failure recovery, comprising a feature extraction module and a steady-state prediction module, characterized in that: The feature extraction module is used to collect data information of the odor sensor after power failure and recovery to perform feature extraction; The steady-state prediction module is used to obtain the steady-state value information of the odor sensor before power failure, build a regression model based on the characteristics of the data information to predict the steady-state value of the sensor, and output the odor status of the refrigerator.

8. A steady-state prediction device for an odor sensor after power failure recovery, further comprising a steady-state judgment module and a storage module, characterized in that: The steady-state judgment module is used to collect and store data information of the odor sensor in real time to determine whether the sensor is in a balanced state; The storage module is used to store the steady-state value of the sensor when the odor sensor is in a balanced state.

9. An electronic device, comprising a memory and a processor, characterized in that: The memory stores a computer program that can be run on the processor, and when the program is executed on the processor, the steps in the method for predicting the steady state of the odor sensor after power failure recovery as described in any one of claims 1 to 6 are implemented.

10. A storage medium storing a computer program, characterized in that: When the computer program is executed by the processor, the steps of the method for predicting the steady state of the odor sensor after power failure and recovery as described in any one of claims 1 to 6 are implemented.

Images