Sensor dynamic sampling method based on ambient fluctuation perception and carafe optimization

Abstract

The application discloses a sensor dynamic sampling method based on environmental fluctuation sensing and CARAFE optimization, comprising the following steps: S1, an environmental sensing module calculates the standard deviation sigma of window data from original data collected by a sensor node and judges whether the standard deviation sigma exceeds a dynamic threshold theta, so as to monitor the fluctuation of environmental data; S2, a dynamic sampling controller switches the sampling frequency according to different states of the fluctuation of the environment, and realizes the reduction of the energy consumption of the system while ensuring the data quality through the downsampling processing of content sensing and the intelligent batch transmission strategy. The sensor dynamic sampling method based on environmental fluctuation sensing and CARAFE optimization can adaptively adjust the sampling frequency, and optimize the features of the sampling data through the CARAFE operator of content sensing, so as to improve the accuracy and energy efficiency of data acquisition.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of sensor sampling, and particularly relates to a sensor dynamic sampling method based on environmental fluctuation sensing and CARAFE optimization. BACKGROUND

[0002] In wireless sensor network (WSN) and remote monitoring applications, a large number of unattended wireless sensor nodes are usually deployed in the monitoring area. These nodes realize real-time monitoring, sensing and data collection of various monitoring objects through wireless ad hoc network, and upload information to the base station for storage and processing. Energy efficiency is one of the key indicators of wireless sensor network, and the energy consumption of sensor nodes is mainly concentrated in CPU, sensor module and wireless transceiver, among which the energy consumption of wireless communication is higher, and data transmission accounts for about 80% of the entire network energy consumption.

[0003] Therefore, wireless sensor nodes need to adopt a reasonable sampling strategy to reduce the transmission of redundant data as much as possible to reduce overall energy consumption. The traditional sensor sampling method is usually continuous sampling (periodic sampling), that is, the sensor node collects data at a fixed time interval. Although this method has small sampling error and simple control process, the sampling frequency is fixed, which leads to high power consumption and is difficult to meet the energy demand of wireless sensor nodes. Especially when the environmental parameters are stable, the high sampling frequency will increase unnecessary energy consumption. In addition, the fixed frequency sampling system cannot flexibly respond to rapid changes in the environment and may miss critical data when the environment fluctuates sharply.

[0004] Therefore, the above problems need to be further improved. SUMMARY

[0005] The main purpose of the application is to provide a sensor dynamic sampling method based on environmental fluctuation sensing and CARAFE optimization, which can adaptively adjust the sampling frequency and optimize the features of the sampling data through the content-aware CARAFE operator to improve the accuracy and energy efficiency of data collection. Under the synergistic effect of the environmental sensing module, the dynamic sampling controller, the CARAFE downsampling module and the data transmission and storage strategy, the energy consumption of the wireless sensor node is effectively reduced, which is suitable for application in low-power scenarios such as remote monitoring and wireless sensor network.

[0006] To achieve the above purpose, the application provides a sensor dynamic sampling method based on environmental fluctuation sensing and CARAFE optimization, comprising the following steps:

[0007] Step S1: The environmental sensing module calculates the standard deviation σ of the window data from the original data collected by the sensor node and judges whether the standard deviation σ exceeds the dynamic threshold θ, so as to monitor the fluctuation of the environmental data;

[0008] Step S2: The dynamic sampling controller switches the sampling frequency according to different states of the fluctuation of the environment, and realizes the reduction of the energy consumption of the system while ensuring the data quality through the content-aware down-sampling processing and the intelligent batch transmission strategy.

[0009] As a further preferred technical solution of the above technical solution, step S1 is specifically implemented as:

[0010] Step S1.1: The raw data collected by each sensor node of the environment perception module is allocated to a fixed time window, so as to calculate the standard deviation σ of the window data, the formula being as follows:

[0011]

[0012] wherein x i is the data value of the i-th sampling point, is the average value of the window data, and N is the total number of sampling points in the time window, and the fluctuation of the standard deviation σ directly reflects the change of the environment;

[0013] When σ exceeds the preset dynamic threshold θ, the system judges that the environment is in a fluctuation state; when σ is lower than θ, the system judges that the environment is in a stable state;

[0014] Step S1.2: (in order to improve the accuracy of the environment state judgment) the dynamic threshold θ is adaptively adjusted according to the change of the environment, and the calculation formula is:

[0015] θ = α · σ histoty + β

[0016] wherein σ history represents the average standard deviation of the historical data, and α and β are adjustment parameters for controlling the baseline and sensitivity of the dynamic threshold (such an adaptive dynamic threshold mechanism helps the system to maintain a reasonable sampling frequency in different scenarios and improve the accuracy of the environment state monitoring).

[0017] As a further preferred technical solution of the above technical solution, step S2 is specifically implemented as the following steps:

[0018] Step S2.1: for the stable state, the dynamic sampling controller sets the sampling frequency to a low frequency (such as sampling once per hour) to save energy consumption;

[0019] Step S2.2: for the fluctuation state, the dynamic sampling controller sets the sampling frequency to a high frequency (such as sampling once per minute) to capture the rapidly changing environmental information.

[0020] As a further preferred technical solution of the above technical solution, in step S2, (in order to further optimize the efficiency and real-time performance of the sampling frequency switching) the dynamic sampling controller predicts the environmental fluctuations at future time points through a time series prediction model, so as to switch to the high-frequency mode in advance before the environmental fluctuations occur, and the time series prediction model is as follows:

[0021]

[0022] wherein φ i is the coefficient of the autoregressive term, θ j is the coefficient of the moving average term, ∈ t-j is the white noise term, x t-i represents the past observation data, and the parameters p, q and t are basic elements required for defining and calculating the model; based on the time series prediction model, adaptive sampling frequency adjustment is realized on the basis of predicting the environmental state (reducing the system overhead caused by frequent frequency switching, while ensuring the effectiveness and timeliness of data acquisition).

[0023] As a further preferred technical solution of the above technical solution, for step S2, (in the high-frequency sampling mode, the amount of data collected by the sensor node increases, and direct transmission will significantly increase the communication energy consumption) a kernel prediction module and a feature recombination module of a CARAFE operator are introduced to perform content-aware downsampling processing on the sampling data (the CARAFE operator can identify the change area in the data and perform weighted sampling on the samples to retain the key change information while removing redundant data. Unlike the traditional method using a 1x1 or 2x2 receptive field, the CARAFE operator has a larger receptive field and can consider the entire feature mapping, thereby better retaining the details and edge information of the image and reducing the sawtooth and blur effects in the sampling process. The CARAFE operator supports dynamic generation of adaptive kernels and can perform content-aware processing according to different targets, while the traditional interpolation method cannot dynamically adjust the size of the kernel to match the data content), wherein:

[0024] In the kernel prediction module, the adaptive convolution kernel is generated according to the input features and is related to the position, thereby improving the recognition and retention capability of the change information, and the kernel prediction module generates a position-related adaptive convolution kernel according to the input features, and the formula is as follows:

[0025] K i,j =g(X i,j )

[0026] wherein K i,j is the convolution kernel matrix, X i,j is the feature information of the original data, and the function g is the kernel prediction network, and the adaptive convolution kernel structure allows dynamic weighted optimization according to the characteristics of the environmental data.

[0027] In the weighted sampling step, the feature reorganization module effectively extracts the data change area through resampling, and the CARAFE operator performs content-aware resampling on the data area according to the generated weight W i,j The sampling data is weighted and convoluted, and the data area is content-aware resampled using a convolution kernel K, and the formula is:

[0028]

[0029] Where y i,j is the data after resampling processing, K is the convolution kernel, the size is (2k+1) x (2k+1), x i+m,j+n is the original data point, i and j represent the target position coordinates of the resampled data in the convolution operation, and m and n are the relative position offsets of the convolution kernel, which are used to sample the original data points within the range of the convolution kernel. By weighting the resampling in the significant area, the CARAFE operator can effectively remove unnecessary data redundancy, reduce the data volume, and retain the feature information of the environment (this method can reduce the data transmission volume while maintaining the key change information, effectively improving the energy efficiency of data processing).

[0030] As a further preferred technical solution of the above technical solution, for step S2, the batch data transmission strategy is used to solve the data volume problem caused by batch transmission, wherein:

[0031] In the stable state, the system temporarily stores the sampling data in the local storage unit, and then performs batch transmission when the environment enters the fluctuation state, thereby reducing the transmission frequency and communication power consumption. The system uses the following formula to quantify the total power consumption in the transmission process:

[0032]

[0033] Where T total represents the total power consumption of the system, T sampling represents the power consumption of data acquisition, D represents the data volume of batch transmission, R represents the data transmission rate, and P transmit represents the transmission power consumption. By reasonably controlling the size of the batch data volume D and the transmission frequency, and preferably using a low-power communication protocol (such as LoRa or BLE), the communication power consumption is reduced and the service life of the sensor node is prolonged under the premise of ensuring data integrity.

[0034] The beneficial effects of the present application are:

[0035] The dynamic sampling algorithm of the application realizes the significant reduction of the energy consumption of the system while ensuring the data quality through the steps of environment perception, dynamic sampling adjustment, content-perceived down-sampling and intelligent batch transmission. The adaptive sampling scheme based on environment fluctuation enables the wireless sensor network and remote monitoring application to realize efficient and low-consumption operation in the case of battery power supply, and has wide application prospect and practical value. BRIEF DESCRIPTION OF DRAWINGS

[0036] Fig. 1 is a flowchart of the application.

[0037] Fig. 2 is the overall framework of the CARAFE operator. DETAILED DESCRIPTION

[0038] The following description is used to disclose the application so that those skilled in the art can implement the application. The preferred embodiments in the following description are only used as examples, and other obvious modifications can be thought of by those skilled in the art. The basic principles of the application defined in the following description can be applied to other embodiments, modifications, improvements, equivalents and other technical solutions without departing from the spirit and scope of the application.

[0039] In the preferred embodiments of the application, those skilled in the art should note that the sensors and the like involved in the application can be regarded as prior art.

[0040] Preferred embodiments.

[0041] As shown in Figs. 1-2 The application discloses a sensor dynamic sampling method based on environment fluctuation perception and CARAFE optimization, comprising the following steps:

[0042] Step S1: The environment perception module calculates the standard deviation σ of the window data from the raw data collected by the sensor nodes and judges whether the standard deviation σ exceeds the dynamic threshold θ, so as to monitor the fluctuation of the environment data;

[0043] Step S2: The dynamic sampling controller switches the sampling frequency according to the different states of the environment fluctuation, and realizes the reduction of the energy consumption of the system while ensuring the data quality through the content-perceived down-sampling processing and the intelligent batch transmission strategy.

[0044] Specifically, step S1 is implemented as:

[0045] Step S1.1: The raw data collected by each sensor node of the environment perception module is distributed into a fixed time window, so as to calculate the standard deviation σ of the window data, and the formula is as follows:

[0046]

[0047] where x i is the data value of the i-th sampling point, is the average value of the window data, N is the total number of sampling points within the time window, and the fluctuation of the standard deviation σ intuitively reflects the change of the environment;

[0048] When σ exceeds the preset dynamic threshold θ, the system determines that the environment is in a fluctuating state; when σ is lower than θ, the system determines that the environment is in a stable state;

[0049] Step S1.2: (in order to improve the accuracy of the environmental state judgment) the dynamic threshold θ is adaptively adjusted according to the environmental change, and the calculation formula is:

[0050] θ = α·σ history + β

[0051] where σ history represents the average standard deviation of historical data, and α and β are adjustment parameters for controlling the baseline and sensitivity of the dynamic threshold (such an adaptive dynamic threshold mechanism helps the system to maintain a reasonable sampling frequency in different scenarios and improve the accuracy of environmental state monitoring).

[0052] More specifically, step S2 is implemented as the following steps:

[0053] Step S2.1: for the stable state, the dynamic sampling controller sets the sampling frequency to a low frequency (such as sampling once an hour) to save energy consumption;

[0054] Step S2.2: for the fluctuating state, the dynamic sampling controller sets the sampling frequency to a high frequency (such as sampling once a minute) to capture rapidly changing environmental information.

[0055] Further, in step S2, (in order to further optimize the efficiency and real-time performance of the sampling frequency switching) the dynamic sampling controller predicts the environmental fluctuation at the future time through a time series prediction model, so as to switch to the high-frequency mode in advance before the environmental fluctuation occurs, and the time series prediction model is as follows:

[0056]

[0057] where φ i is the coefficient of the autoregressive term, θ j is the coefficient of the moving average term, ∈ t-j is the white noise term, and x t-iThe past observation data are represented, and the parameters p, q, and t are basic elements required for defining and calculating the model, and affect the accuracy and stability of the prediction; based on the time series prediction model, adaptive sampling frequency adjustment is realized on the basis of predicting the environmental state (reducing the system overhead caused by frequent switching of the frequency, while ensuring the effectiveness and timeliness of data acquisition); in the time series prediction model, the sensor node establishes a model according to the past sampling data to predict the future environmental change trend. When the model predicts that the environment may have a large fluctuation, the system will automatically increase the sampling frequency to capture the rapidly changing data in time; when the environment is expected to remain stable, the system will reduce the sampling frequency, thereby reducing the energy consumption. The adaptive adjustment logic not only improves the accuracy of sampling, but also effectively avoids the system overhead caused by frequent switching of the sampling frequency.

[0058] Furthermore, for step S2, (in the high-frequency sampling mode, the amount of data collected by the sensor node increases, and direct transmission will significantly increase the communication energy consumption) a kernel prediction module and a feature reorganization module of a CARAFE operator are introduced to perform content-aware downsampling processing on the sampling data (the CARAFE operator can identify the change area in the data and perform weighted sampling on the samples to retain key change information while removing redundant data. Unlike the traditional method using a 1x1 or 2x2 receptive field, the CARAFE operator has a larger receptive field and can consider the entire feature mapping, thereby better retaining the details and edge information of the image and reducing the sawtooth and blur effects in the sampling process. The CARAFE operator supports dynamic generation of adaptive kernels and can perform content-aware processing according to different targets, while the traditional interpolation method cannot dynamically adjust the size of the kernel to match the data content), wherein:

[0059] In the kernel prediction module, the adaptive convolution kernel is generated according to the input feature and is related to the position, thereby improving the recognition and retention capability of the change information, the kernel prediction module generates a position-related adaptive convolution kernel according to the input feature, and the formula is:

[0060] K i,j =g(X i,j )

[0061] wherein K i,j is a convolution kernel matrix, X i,j is feature information of the original data, the function g is a kernel prediction network, and the adaptive convolution kernel structure allows dynamic weighted optimization according to the characteristics of the environmental data;

[0062] In the weighted sampling step, the feature reorganization module effectively extracts the data change area through resampling, and the CARAFE operator generates a weight W i,jThe sampling data is weighted and convoluted, and the data region is resampled using a convolution kernel K, and the formula is:

[0063]

[0064] Wherein, y i,j is the data after resampling processing, K is a convolution kernel, and the size is (2k+1) x (2k+1), x i+m,j+n is the original data point, i and j represent the target position coordinates of the resampled data in the convolution operation, and m and n are relative position offsets of the convolution kernel, which are used for sampling the original data point in the range of the convolution kernel, the CARAFE operator can effectively remove unnecessary data redundancy by weighted resampling in the significant area, and the environmental feature information is retained while the data amount is reduced (the method can reduce the data transmission amount in the high-frequency sampling mode, while the key change information is maintained, and the data processing efficiency is effectively improved). The weight W i,j in the CARAFE algorithm is used to weight the convolution kernel of each position to highlight the more important feature area. The actual calculation process of W i,j is not embodied in the formula, it is only mentioned as the result of generating the convolution kernel weight, and the specific application mode is not shown in the formula. This weighting mechanism can more effectively concentrate sampling in the data region with significant changes, and improve the accuracy of data processing.

[0065] Preferably, for step S2, the data amount problem caused by batch transmission is solved by an intelligent batch data transmission strategy, wherein:

[0066] In the stable state, the system temporarily stores the sampling data in the local storage unit, and then performs batch transmission when the environment enters the fluctuation state, so as to reduce the transmission frequency and communication power consumption, and the system uses the following formula to quantify the total power consumption in the transmission process:

[0067]

[0068] Wherein, T total represents the total power consumption of the system, T sampling represents the power consumption of data acquisition, D represents the data amount of batch transmission, R represents the data transmission rate, and P transmit represents the transmission power consumption. By reasonably controlling the size of the batch data amount D and the transmission frequency, and preferably using a low-power communication protocol (such as LoRa or BLE), the communication power consumption is reduced and the service life of the sensor node is prolonged under the premise of ensuring data integrity.

[0069] The advantages of the present application are:

[0070] The dynamic sampling method effectively reduces the energy consumption of the sensor node, and improves the efficiency and accuracy of data collection.

[0071] Through the dual judgment mechanism of standard deviation fluctuation and adaptive threshold, the system effectively distinguishes between the "fluctuation" and "stability" states of the environment, achieving a dynamic balance between sampling frequency and energy consumption. Combined with the time series prediction model, the algorithm can predict the environment fluctuation in advance and switch the sampling frequency in advance, thereby reducing the overhead of frequent switching. At the same time, through the kernel prediction and feature reconstruction module of the CARAFE operator, the system performs weighted sampling on the high-frequency sampled data, significantly compresses the data volume, and improves the energy efficiency of transmission and storage. In summary, the application is suitable for low-power remote monitoring and wireless sensor network scenarios, ensuring the effectiveness of the data without increasing the power consumption, and meeting the application requirements of high accuracy and energy efficiency for environmental state monitoring.

[0072] It is worth mentioning that the sensor and other technical features involved in the present patent application should be considered as prior art, and the specific structure, working principle and possible control method, spatial arrangement method of these technical features can be selected conventionally in the art, and should not be considered as the invention point of the present patent, which will not be further detailed.

[0073] For those skilled in the art, the technical solutions described in the foregoing embodiments can be modified or some technical features can be replaced by equivalents, and any modification, equivalent replacement, improvement, etc. within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A sensor dynamic sampling method based on ambient fluctuation perception and CARAFE optimization, characterized in that, The method comprises the following steps: Step S1: The environment perception module calculates the standard deviation σ of the window data from the raw data collected by the sensor nodes and determines whether the standard deviation σ exceeds the dynamic threshold θ, thereby monitoring the fluctuation of the environmental data; Step S2: The dynamic sampling controller switches the sampling frequency according to different states of the fluctuation of the environment, and realizes the reduction of the energy consumption of the system while ensuring the data quality through the content-aware downsampling processing and the intelligent batch transmission strategy; Step S1 is specifically implemented as: Step S1.1: The raw data collected by each sensor node of the environment perception module is allocated to a fixed time window, thereby calculating the standard deviation σ of the window data, and the formula is as follows: ； wherein, is the data value of the i-th sampling point, is the average value of the window data, N is the total number of sampling points in the time window, and the fluctuation of the standard deviation σ directly reflects the change of the environment. When σ exceeds the preset dynamic threshold θ, the system determines that the environment is in a fluctuation state; when σ is lower than θ, the system determines that the environment is in a stable state; Step S1.2: The dynamic threshold θ is adaptively adjusted according to the environmental changes, and the calculation formula is as follows: ； wherein, the average standard deviation of the historical data, and are tuning parameters for controlling the baseline and sensitivity of the dynamic threshold; Step S2 is specifically implemented as the following steps: Step S2.1: For the stable state, the dynamic sampling controller sets the sampling frequency to a low frequency to save energy consumption; Step S2.2: For the fluctuation state, the dynamic sampling controller sets the sampling frequency to a high frequency to capture the rapidly changing environmental information; In step S2, the dynamic sampling controller predicts the environmental fluctuation at the future time through a time series prediction model, thereby switching to the high-frequency mode in advance before the environmental fluctuation occurs, and the time series prediction model is as follows: ； wherein, is a coefficient of an autoregressive term, is a coefficient of a moving average term, is a white noise term, denotes past observation data, the parameters p, q, t are essential elements required for the definition and calculation of the model; based on the time series prediction model, the adaptive sampling frequency adjustment is realized on the basis of the prediction of the environmental state; For step S2, a kernel prediction module and a feature recombination module of a CARAFE operator are introduced to perform content-aware downsampling processing on the sampling data, wherein: In the kernel prediction module, an adaptive convolution kernel is generated according to the input features, which is related to the position, thereby improving the recognition and retention ability of the changing information, and the kernel prediction module generates a position-related adaptive convolution kernel according to the input features, and the formula is as follows: ； wherein, is a convolution kernel matrix, is the feature information of the original data, the function is the kernel prediction network, and the adaptive convolution kernel structure allows dynamic weighted optimization according to the characteristics of the environmental data; In the weighted sampling step, the feature reorganization module effectively extracts the data change area by resampling, and the CARAFE operator performs content-aware resampling on the data area according to the generated weight The weighted convolution calculation is performed on the sampling data, and the convolution kernel K is used for content-aware resampling on the data area, and the formula is as follows: ； wherein, is the data after resampling, K is the convolution kernel, and the size is , is the original data point, i and j represent the target position coordinates of the resampled data in the convolution operation, and m and n are the relative position offsets of the convolution kernel, which are used to sample the original data points within the range of the convolution kernel. By weighting the resampling in the salient region, the CARAFE operator effectively removes unnecessary data redundancy while reducing the data volume while retaining the feature information of the environment. For step S2, an intelligent batch data transmission strategy is used to solve the data volume problem caused by batch transmission, wherein: In the stable state, the system temporarily stores the sampling data in the local storage unit, and then performs batch transmission when the environment enters the fluctuation state, thereby reducing the transmission times and communication power consumption, and the system uses the following formula to quantify the total power consumption in the transmission process: ； wherein, represents the total power consumption of the system, represents the power consumption of data collection, D represents the data volume of batch transmission, R represents the data transmission rate, represents the transmission power consumption, by reasonably controlling the size of the batch data volume D and the transmission frequency, and using a low-power consumption communication protocol, the communication power consumption is reduced and the service life of the sensor node is prolonged under the premise of ensuring data integrity.

Images