Wireless sensor energy-saving control method based on environmental perception

Abstract

The invention discloses a wireless sensor energy-saving control method based on environmental perception, and belongs to the technical field of wireless sensor network nodes, and the method specifically comprises the following steps: a node collects environmental parameters in an initially set fixed sampling period, and constructs a historical sampling sequence composed of a plurality of continuous sampling period data; according to a historical sampling sequence, a trend characteristic value reflecting environment change is extracted, an environment fluctuation level corresponding to a current sampling period is judged based on the trend characteristic value, and whether a state switching judgment mode is triggered or not is judged based on comparison between the environment fluctuation level and a set activation threshold value and a dormancy threshold value; and judging whether to execute the state switching operation or not through the state switching factor. The method has the beneficial effects that the energy consumption overhead of the node in a stable environment can be effectively reduced, and meanwhile, the response capability of the node in a sudden change scene is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of low-power wireless sensing technology, in particular to a wireless sensor energy-saving control method based on environment perception. BACKGROUND

[0002] With the continuous improvement of the intelligent level in multi-functional space environment, the environment perception system based on wireless sensor network is widely used in indoor climate regulation, energy management, personnel behavior analysis and other fields. Wireless sensor network is usually composed of a large number of sensor nodes distributed in different positions, which are responsible for collecting various environmental parameters such as temperature, humidity, illumination, air flow, gas concentration, human presence, etc., and uploading data to a centralized control platform for comprehensive analysis and decision-making. In actual deployment, the system usually requires continuous operation of each sensor node to ensure real-time reflection and dynamic response of the space state.

[0003] However, existing sensor nodes usually use fixed working cycle or preset strategy, whether there is environmental change in the monitoring area or not, they continue to sample and transmit data according to the set frequency, which causes obvious energy waste in a large number of stable states or low dynamic scenes. In addition, in order to realize the coordination and low-power management of the whole wireless sensor network, the existing system often relies on network-level centralized scheduling mechanism or intermediate control layer to adjust the node state, which not only increases the complexity of system design and operation and maintenance, but also limits the flexible adaptation ability of sensor nodes in heterogeneous deployment or dynamic environment. Especially in the scene of complex actual space structure and variable use mode, if the sensor node itself lacks the ability of environment perception and self-adjustment, the whole system will be difficult to realize fine energy consumption control and low-maintenance operation. Therefore, how to make the sensor node itself have the ability of environment perception and dynamic scheduling of running state is one of the key problems to improve the energy efficiency and intelligent level of response of wireless sensor network. SUMMARY

[0004] In order to solve the problems of high energy consumption, rigid running mechanism and lack of self-adaptive ability to environmental changes of sensor nodes in wireless sensor network in long-term running process, the present application provides a wireless sensor energy-saving control method based on environment perception, which focuses on the optimization design of node-level state scheduling mechanism. The present application realizes the dynamic adjustment of node sampling, activation and sleep state by introducing the judgment mechanism based on environmental fluctuation trend, which helps the sensor node to adjust the running strategy according to the environmental change without relying on centralized control, so as to improve the running efficiency and reduce unnecessary energy consumption, and is suitable for distributed application requirements in various environmental monitoring scenes.

[0005] In order to achieve the above-mentioned purposes, the present application provides a wireless sensor energy-saving scheduling method based on environment perception, which is applied to a wireless sensor node, wherein the node comprises an environment parameter sensor, a communication module and a low-power control processor, and the method specifically comprises the following steps: P1: The node collects environment parameters in an initially set fixed sampling period, and constructs a historical sampling sequence composed of a plurality of continuous sampling period data; P2: According to the historical sampling sequence, a trend characteristic value reflecting the change of the environment is extracted, and the trend characteristic value is mapped to a plurality of membership values of predefined fluctuation levels through a fuzzy membership function, and then the membership values are evaluated according to a preset fuzzy reasoning rule to determine the environment fluctuation level corresponding to the current sampling period; P3: Based on the comparison between the environment fluctuation level and the set activation threshold and sleep threshold, when the node is in a sleep state and the current environment fluctuation level is higher than the activation threshold, or when the node is in an active state and the environment fluctuation level is lower than the sleep threshold in a plurality of continuous sampling periods, a state switching judgment mode is triggered, and in other cases, the current running state is maintained; when the state switching judgment mode is triggered, a state switching factor is used to judge whether to perform a state conversion operation, wherein the state switching judgment action is performed by the node at a fixed time in each sampling period; P4: During the active state of the node, a prediction model is constructed based on the current environment fluctuation level and the historical environment fluctuation level, which is used to estimate the environment fluctuation level of the next sampling period, and the current sampling period, communication interval, transmission power and sleep duration are adjusted according to the prediction result; P5: When the current environment fluctuation level of the node approaches the activation threshold or the sleep threshold, and the state judgment confidence is lower than the set threshold, a neighboring auxiliary judgment mechanism is started, the environment fluctuation level information broadcasted by the neighboring wireless sensor nodes in the current period within the communication range is listened to, and the current environment fluctuation level of the received neighboring nodes is compared with the current environment fluctuation level of the node; when the similarity meets the preset condition, the confidence of the node state judgment is enhanced, and the set values of the activation threshold and the sleep threshold are corrected accordingly.

[0006] In the P1, the historical sampling sequence is formed by obtaining an environment parameter measurement value once in each fixed sampling period and being cached in time sequence, the length of the cache sequence is a preset limited sample number, and when the sequence length reaches the preset value, the first-in-first-out mode is used for updating.

[0007] In the P2, the trend characteristic value reflecting the change of the environment is extracted according to the historical sampling sequence Specifically, ; wherein, represents the environmental parameter collected in the i-th sampling period, represents the mean value calculated based on the previous historical periods in the period, represents the standard deviation of the historical sampling sequence, is a preset non-zero positive constant for avoiding zero denominator, and the trend characteristic value is used as an input basis for judging the environmental fluctuation level of the current period.

[0008] The fuzzy membership function in P2 adopts a symmetrical triangular membership function, and the membership values of the symmetrical triangular membership function are divided into several fluctuation level intervals (which can be five fluctuation level intervals, corresponding to "extremely low, low, medium, high, and extremely high" five levels) according to the numerical size of the trend characteristic value , wherein each fluctuation level interval is equally divided with the statistical mean value of the trend characteristic value as the center, and the fuzzy reasoning rule adopts a Mamdani type structure, and the input of the Mamdani type structure is the fuzzified trend characteristic value, and the output is the environmental fluctuation level of the current sampling period.

[0009] The state switching judgment in P3 adopts a hysteresis control strategy with different trigger conditions, that is, the activation condition and the sleep condition are asymmetric, specifically: When the node is in the sleep state, only under the condition that the environmental fluctuation level of the current period is higher than the activation threshold, the state switching judgment mode is triggered; When the node is in the active state, only under the condition that the environmental fluctuation level is lower than the sleep threshold in the continuous periods, the state switching judgment mode is triggered, is a set positive integer; and in other cases, the current running state is maintained.

[0010] In P3, when the state switching judgment mode is triggered, the node constructs a state switching factor based on the difference between the current environmental fluctuation level and the activation threshold, the fluctuation level change trend intensity in the historical multiple periods, and the fluctuation level change amplitude between adjacent periods, and judges whether to perform the state switching operation from sleep to active or from active to sleep based on the value of the state switching factor. Specifically: the state switching factor , specifically: ; wherein, represents the state switching factor of the current sampling period, is the environmental fluctuation level of the current sampling period, is the activation threshold,​ The exponential enhancement coefficient, The integral gain coefficient represents the fluctuation trend. In history The cumulative first-order change in the level of environmental fluctuations over a given period is used to characterize the strength of the trend. Indicates the first Environmental fluctuation levels for each sampling period; Indicates the first Environmental fluctuation levels for each sampling period.

[0011] The prediction model constructed in P4 specifically includes: the node is based on Environmental fluctuation levels for each historical sampling period Calculate and predict the volatility level The calculation formula is as follows: ; in, This indicates the predicted level of environmental volatility for the next cycle. The fluctuation level for the current cycle. The number of environmental level segments, For level The weighting coefficients, and The first The mean and standard deviation of the level states, Indicates the level in historical sampling Frequency of occurrence The sum of cumulative frequencies for all levels. It is a very small constant.

[0012] In step P5, the current environmental fluctuation level of the received neighboring nodes is compared with the current environmental fluctuation level of the current node. The similarity value is calculated using a formula. : ; in, This indicates the number of neighboring nodes detected. As a smoothing factor, when When this occurs, it is considered that neighboring collaboration is effective, triggering a confidence enhancement action.

[0013] In P5, when the similarity meets a preset condition, the confidence level of the node state judgment is enhanced, and the settings of the activation threshold and the dormancy threshold are adjusted accordingly. Specifically: ; ; in, , These are the original activation threshold and the original dormancy threshold, respectively. This is a dynamic adjustment coefficient. It is a proximity similarity index. , These are the corrected activation threshold and the corrected dormancy threshold.

[0014] The beneficial effects of this invention are: by combining environmental parameter trend analysis with node state adaptive switching mechanism, this invention can effectively reduce the energy consumption of nodes in a stable environment, while improving their response capability in sudden change scenarios. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of Embodiment 1 of the present invention. Detailed Implementation

[0016] To clearly illustrate the technical features of this solution, the following detailed implementation method will be used to explain the solution.

[0017] Example 1: This embodiment of the invention provides a wireless sensor energy-saving control method based on environmental perception, applied to wireless sensor nodes. The method specifically includes: P1: The node collects environmental parameters within the initially set fixed sampling period and constructs a historical sampling sequence consisting of data from multiple consecutive sampling periods. The historical sampling sequence is formed by acquiring environmental parameter measurements once within each fixed sampling period and caching them in chronological order. The length of the cached sequence is a preset finite number of samples. When the sequence length reaches the preset value, it is updated using a first-in-first-out method.

[0018] P2: Based on historical sampling sequences, extract trend feature values ​​reflecting environmental changes, and map these trend feature values ​​to membership values ​​of multiple predefined fluctuation levels using fuzzy membership functions. Then, evaluate the membership values ​​according to preset fuzzy inference rules to determine the environmental fluctuation level corresponding to the current sampling period. The fuzzy membership function uses a symmetrical triangular membership function, and the membership values ​​of the symmetrical triangular membership function are based on the trend feature values. V k The numerical values ​​are divided into several fluctuation level intervals, with each fluctuation level interval being equally divided with the statistical mean of the trend feature value as the center. The fuzzy inference rule adopts the Mamdani type structure. The input of the Mamdani type structure is the fuzzified trend feature value, and the output is the environmental fluctuation level of the current sampling period.

[0019] P3: Based on a comparison of the environmental fluctuation level with the set activation and dormancy thresholds, a state transition judgment mode is triggered when the node is in a dormant state and the current environmental fluctuation level is higher than the activation threshold, or when the node is in an active state and the environmental fluctuation level is lower than the dormancy threshold for multiple consecutive sampling periods. Otherwise, the current operating state is maintained. When the state transition judgment mode is triggered, a state transition factor is used to determine whether to perform a state transition operation. The state transition judgment action is executed periodically by the node within each sampling period. The state transition judgment adopts a hysteresis control strategy with different trigger conditions, specifically: When a node is in a dormant state, the state switching judgment mode is triggered only when the environmental fluctuation level of the current period is higher than the activation threshold. When a node is active, it only occurs in consecutive... If the environmental fluctuation level is below the sleep threshold in each sampling period, the state switching judgment mode is triggered. The specified positive integer; In all other cases, maintain the current operating state.

[0020] When the state switching judgment mode is triggered, the node constructs a state switching factor based on the difference between the current environmental fluctuation level and the activation threshold, the intensity of the fluctuation level change trend in multiple historical periods, and the fluctuation level change amplitude between adjacent periods. Based on the value of the state switching factor, it determines whether to perform a state switching operation from dormant to active or from active to dormant.

[0021] P4: While the node is active, a prediction model is built based on the current environmental fluctuation level and the historical environmental fluctuation level to estimate the environmental fluctuation level of the next sampling period, and the current sampling period, communication interval, transmission power and sleep duration are adjusted according to the prediction results. P5: When the current environmental fluctuation level of a node is close to the activation threshold or dormancy threshold, and the confidence level of the state judgment is lower than the set threshold, the proximity auxiliary judgment mechanism is activated. The node listens to the environmental fluctuation level information broadcast by the neighboring wireless sensor nodes within the communication range in the current period, and compares the current environmental fluctuation level of the neighboring nodes with the current environmental fluctuation level of the node itself. When the similarity meets the preset conditions, the confidence level of the node state judgment is enhanced, and the settings of the activation threshold and dormancy threshold are adjusted accordingly.

[0022] Example 2: See Figure 1 This invention provides an energy-saving control method for wireless sensors based on environmental perception, applied to a wireless sensor node. The node includes an environmental parameter sensor, a communication module, and a low-power control processor. The method specifically includes: P1: The node collects environmental parameters within the initially set fixed sampling period and constructs a historical sampling sequence composed of data from multiple consecutive sampling periods. The historical sampling sequence is formed by acquiring environmental parameter measurements once within each fixed sampling period and caching them in chronological order. The length of the cached sequence is a preset finite number of samples. When the sequence length reaches the preset value, it is updated using a first-in-first-out method.

[0023] P2: Based on historical sampling sequences, extract trend feature values ​​reflecting environmental changes, and map these trend feature values ​​to membership values ​​of multiple predefined fluctuation levels using fuzzy membership functions. Then, evaluate the membership values ​​according to preset fuzzy inference rules to determine the environmental fluctuation level corresponding to the current sampling period; extract trend feature values ​​reflecting environmental changes based on historical sampling sequences. Specifically: ; in, Indicates the first Environmental parameters collected within each sampling period Indicates that within this period, based on the previous The average calculated over a historical period. This represents the standard deviation of the historical sampling sequence. This is a preset non-zero positive constant used to avoid zero denominators and trend characteristic values. This serves as the input for judging the current period's environmental fluctuation level. In this embodiment, the fuzzy membership function uses a symmetrical triangular membership function, and the membership degree values ​​of the symmetrical triangular membership function are based on trend characteristic values. The numerical values ​​are divided into five fluctuation level intervals, corresponding to the five levels of "extremely low, relatively low, medium, relatively high, and extremely high". Each fluctuation level interval is divided equally around the statistical mean of the trend feature value. The fuzzy inference rule adopts the Mamdani type structure. The input of the Mamdani type structure is the fuzzified trend feature value, and the output is the environmental fluctuation level of the current sampling period.

[0024] P3: Based on a comparison of the environmental fluctuation level with the set activation and dormancy thresholds, when a node is in a dormant state and the current environmental fluctuation level is higher than the activation threshold, or when a node is in an active state and the environmental fluctuation level is lower than the dormancy threshold for multiple consecutive sampling periods, a state switching judgment mode is triggered. Otherwise, the current operating state is maintained. When the state switching judgment mode is triggered, a state switching factor is used to determine whether to perform a state transition operation. The state switching judgment action is executed periodically by the node within each sampling period. This embodiment employs a hysteresis control strategy with different trigger conditions for state switching judgment. The activation and dormancy conditions are asymmetrical, specifically: When a node is in a dormant state, the state switching judgment mode is triggered only when the environmental fluctuation level of the current period is higher than the activation threshold. When a node is active, it only occurs in consecutive... If the environmental fluctuation level is below the sleep threshold in each sampling period, the state switching judgment mode is triggered. For each positive integer, the environmental fluctuation level is determined as follows when the node is active: ; in, Indicates the first Environmental fluctuation level for each sampling period Indicates the sleep threshold. This is an indicator function that takes the value 1 when the condition is true and 0 otherwise. For the size of the statistical period window, This is the current sampling period number. As a periodic low-volatility coverage factor; when When a node transitions from an active state to a dormant state, a switch is triggered. To determine whether to switch, a sliding window method is used to statistically analyze environmental fluctuation levels over consecutive sampling periods. If, within a set window length, the fluctuation level for all periods is below a dormant threshold, a dormant state switch is triggered. The hysteresis control strategy also includes a state hold period mechanism. During a set hold period after a state switch, the node does not respond to new fluctuation level judgments to avoid frequent state switches. The activation and dormant thresholds are periodically and adaptively adjusted based on current environmental trend characteristics and the confidence level of state judgments. In all other cases, maintain the current operating state.

[0025] When the state transition judgment mode is triggered, the node constructs a state transition factor based on the difference between the current environmental fluctuation level and the activation threshold, the intensity of fluctuation level changes over multiple historical periods, and the magnitude of fluctuation level changes between adjacent periods. Based on the value of this state transition factor, the node determines whether to perform a state transition operation from dormant to active or from active to dormant. Specifically: State transition factor. Specifically: ; in, This indicates the state switching factor for the current sampling period. The environmental fluctuation level for the current sampling period. As the activation threshold, The exponential enhancement coefficient, The integral gain coefficient represents the fluctuation trend. In history The cumulative first-order change in the level of environmental fluctuations over a given period is used to characterize the strength of the trend. Indicates the first Environmental fluctuation levels for each sampling period; Indicates the first Environmental fluctuation levels for each sampling period.

[0026] P4: While the node is active, a prediction model is built based on the current and historical environmental fluctuation levels to estimate the environmental fluctuation level for the next sampling period. The current sampling period, communication interval, transmission power, and sleep duration are adjusted based on the prediction results. The aforementioned prediction model specifically includes: node-based Environmental fluctuation levels for each historical sampling period Calculate and predict the volatility level The calculation formula is as follows: ; in, This indicates the predicted level of environmental volatility for the next cycle. The fluctuation level for the current cycle. The number of environmental level segments, For level The weighting coefficients, and The first The mean and standard deviation of the level states, Indicates the level in historical sampling Frequency of occurrence The sum of cumulative frequencies for all levels. As a minimal constant, this calculation formula comprehensively considers the degree of fuzzy overlap between the current state and each historical level, as well as statistical weights, to output a continuous predicted value. .

[0027] The node is based on the predicted level of environmental fluctuations in the next cycle. Adjust the following control parameters: sampling time interval of the current sampling period, reporting interval of the communication module, wireless transmission power, and preset sleep duration. The values ​​of each parameter are based on... The numerical range it falls within is mapped to the corresponding parameter control strategy table.

[0028] P5: When the current environmental fluctuation level of a node is close to the activation threshold or dormancy threshold, and the confidence level of the state judgment is lower than the set threshold, the proximity auxiliary judgment mechanism is activated. The node listens to the environmental fluctuation level information broadcast by the neighboring wireless sensor nodes within the communication range in the current period, and compares the current environmental fluctuation level of the neighboring nodes with the current environmental fluctuation level of the node itself. When the similarity meets the preset conditions, the confidence level of the node state judgment is enhanced, and the settings of the activation threshold and dormancy threshold are adjusted accordingly.

[0029] Before a node executes the proximity auxiliary judgment mechanism, the following triggering conditions must be met: the current environmental fluctuation level is within the fuzzy transition range between the activation threshold and the dormancy threshold, and the confidence level of the state judgment is lower than the preset threshold for two consecutive sampling periods. Only after the above conditions are met will the node start the proximity auxiliary judgment mechanism and listen to the fluctuation level information broadcast by neighboring nodes within the communication radius in the current period.

[0030] The similarity comparison between the current environmental fluctuation level of neighboring nodes and the current environmental fluctuation level of the current node is performed by calculating the similarity value using a formula. : ; in, This indicates the number of neighboring nodes detected. As a smoothing factor, when When this occurs, it is considered that neighboring collaboration is effective, triggering a confidence enhancement action.

[0031] When the similarity meets the preset conditions, the confidence level of the node state judgment is enhanced, and the settings of the activation threshold and dormancy threshold are adjusted accordingly, specifically: ; ; in, , These are the original activation threshold and the original dormancy threshold, respectively. This is a dynamic adjustment coefficient. It is a proximity similarity index. , These are the corrected activation threshold and the corrected dormancy threshold.

[0032] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A wireless sensor energy-saving control method based on environmental perception, applied to wireless sensor nodes, wherein the method specifically comprises: P1: The node collects environmental parameters within an initially set fixed sampling period and constructs a historical sampling sequence composed of data from multiple consecutive sampling periods; P2: Based on the historical sampling sequence, extract the trend feature value reflecting environmental changes, and map the trend feature value to multiple predefined fluctuation level membership values ​​through fuzzy membership function. Then, evaluate the membership value according to the preset fuzzy inference rule to determine the environmental fluctuation level corresponding to the current sampling period. P3: Based on the comparison between the environmental fluctuation level and the set activation threshold and dormancy threshold, when the node is in a dormant state and the current environmental fluctuation level is higher than the activation threshold, or when the node is in an active state and the environmental fluctuation level is lower than the dormancy threshold for multiple consecutive sampling periods, the state switching judgment mode is triggered, and the state switching factor is used to determine whether to perform the state transition operation.

2. The method according to claim 1, characterized in that, In P1, the historical sampling sequence is formed by acquiring environmental parameter measurement values ​​once in each fixed sampling period and caching them in chronological order. The length of the cached sequence is a preset finite number of samples. When the sequence length reaches the preset value, it is updated using a first-in-first-out method.

3. The method according to claim 1, characterized in that, The fuzzy membership function in P2 adopts the symmetrical triangular membership function, and the membership degree value of the symmetrical triangular membership function is based on the trend characteristic value. V k The numerical value is divided into several fluctuation level intervals, and each fluctuation level interval is equally divided with the statistical mean of the trend feature value as the center. The fuzzy inference rule adopts the Mamdani type structure. The input of the Mamdani type structure is the fuzzified trend feature value, and the output is the environmental fluctuation level of the current sampling period.

4. The method according to claim 1, characterized in that, The state transition judgment in P3 adopts a hysteresis control strategy with different triggering conditions, specifically: When a node is in a dormant state, the state switching judgment mode is triggered only when the environmental fluctuation level of the current period is higher than the activation threshold. When a node is active, it only occurs in consecutive... If the environmental fluctuation level is below the sleep threshold in each sampling period, the state switching judgment mode is triggered. The specified positive integer; In all other cases, maintain the current operating state.

5. The method according to claim 4, characterized in that, In P3, when the state switching judgment mode is triggered, the node constructs a state switching factor based on the difference between the current environmental fluctuation level and the activation threshold, the intensity of the fluctuation level change trend in multiple historical cycles, and the fluctuation level change amplitude between adjacent cycles, and determines whether to perform a state switching operation from dormant to active or from active to dormant based on the value of the state switching factor.

6. The method according to claim 1, characterized in that, In P3, the state switching judgment action is executed by the node at regular intervals within each sampling period.

7. The method according to any one of claims 1-6, characterized in that, While the node is active, it constructs a prediction model based on the current and historical environmental fluctuation levels to estimate the environmental fluctuation level for the next sampling period, and adjusts the current sampling period, communication interval, transmission power, and sleep duration according to the prediction results.

8. The method according to any one of claims 1-6, characterized in that, When the current environmental fluctuation level of the node is close to the activation threshold or dormancy threshold, and the confidence level of the state judgment is lower than the set threshold, the proximity auxiliary judgment mechanism is activated.

Images