Iot terminal power consumption optimization method based on ultra-lightweight technology

Abstract

This invention discloses a power consumption optimization method for IoT terminals based on ultra-lightweight technology, belonging to the field of IoT terminal power consumption optimization technology. The method includes the following steps: collecting interaction data from the IoT terminal; normalizing the interaction data to obtain organized data, and packaging the organized data into feature vectors; extracting behavioral features of different usage behaviors based on the feature vectors corresponding to those behaviors, and distinguishing usage behaviors based on these behavioral features; classifying usage behaviors according to their characteristics to obtain response levels; learning user behavior on the IoT terminal and summarizing usage habits, and adaptively allocating autonomous wake-up intervals for the IoT terminal at different time periods. This invention addresses the problem that existing IoT terminal power consumption optimization technologies in portable Wi-Fi applications still have issues with unreasonable power consumption optimization for portable Wi-Fi, leading to the inability to ensure users can receive network information in a timely manner.

Description

Technical Field

[0001] This invention relates to the field of power consumption optimization technology for IoT terminals, specifically a power consumption optimization method for IoT terminals based on ultra-lightweight technology. Background Technology

[0002] IoT terminal power consumption optimization technology refers to a method that systematically reduces device energy consumption by co-designing hardware, software, and communication protocols while meeting the functional and performance requirements of the terminal.

[0003] In IoT terminals like portable Wi-Fi devices, power consumption is closely related to user activity. However, portable Wi-Fi devices automatically wake up when actively used by the user, meaning power consumption is difficult to optimize during active use. Optimization is only possible when the user is not using the device. But during this time, the communication device needs to receive messages from others; if the portable Wi-Fi is put into sleep mode, it cannot passively receive information. Existing IoT terminal power optimization technologies for portable Wi-Fi typically analyze user habits and predict user behavior to optimize power consumption. However, since the portable Wi-Fi is woken up when actively used, its primary function during idle time is to receive messages. To receive network messages, which are typically irregular and unpredictable, optimizing the power consumption of portable Wi-Fi via prediction can lead to users not receiving information in a timely manner. For example, patent application CN121547836A discloses a "portable Wi-Fi adaptive power-saving optimization system and method," which uses user behavior prediction to control the sleep and wake-up of portable Wi-Fi. However, since message sending and receiving are unpredictable, behavior prediction cannot guarantee that users will receive network information in a timely manner. Existing IoT terminal power consumption optimization technologies still have problems with unreasonable power consumption optimization in portable Wi-Fi applications, resulting in the inability to ensure that users can receive network information in a timely manner. Summary of the Invention

[0004] This invention aims to at least partially solve one of the technical problems in the prior art. It involves installing an ultra-lightweight processor within an IoT terminal to collect interactive data. This data is then normalized to obtain organized data, which is packaged into feature vectors. Before deployment, feature vectors of different usage behaviors are tested and named test behavior vectors. The behavioral characteristics of these test behavior vectors are analyzed, and usage behaviors are differentiated based on these characteristics. Furthermore, usage behaviors are categorized according to their characteristics, resulting in response levels. Simultaneously, the invention learns user behavior on the IoT terminal during a learning cycle and summarizes user habits. Finally, based on usage habits and response levels, the invention adaptively allocates autonomous wake-up intervals to the IoT terminal at different time periods. This addresses the problem that existing IoT terminal power optimization technologies in portable Wi-Fi applications suffer from unreasonable power optimization, leading to the inability to ensure users can receive network information in a timely manner.

[0005] To achieve the above objectives, this application provides a power consumption optimization method for IoT terminals based on ultra-lightweight technology, comprising the following steps: An ultra-lightweight processor is installed in the IoT terminal to collect the interaction data of the IoT terminal; The interaction data is normalized to obtain sorted data, and the sorted data is packaged into feature vectors. Behavioral features of different usage behaviors are extracted from the feature vectors corresponding to those behaviors, and usage behaviors are distinguished based on these behavioral features. Based on the characteristics of the usage behavior, the usage behavior is classified into different levels, resulting in response levels; Learn user behavior and summarize usage habits when using IoT terminals. Based on usage habits and response levels, adaptively allocate autonomous wake-up intervals for IoT terminals in different time periods.

[0006] Furthermore, the step of installing an ultra-lightweight processor within the IoT terminal and collecting interactive data from the IoT terminal also includes the following sub-steps: Install ultra-lightweight processors in IoT terminals; The interaction data of IoT terminals is monitored through an ultra-lightweight processor. The interaction data includes instantaneous packet rate, instantaneous throughput, traffic direction, and connection status.

[0007] Furthermore, the step of normalizing the interactive data to obtain organized data, and then packaging the organized data into feature vectors, further includes the following sub-steps: A normalization algorithm is used to calculate each type of interaction data independently, converting all different interaction data into values ​​from 0 to 1, resulting in organized data; The compiled data of instantaneous packet rate, instantaneous throughput, traffic direction, and connection status are named in sequence as normalized rate, normalized throughput, normalized direction, and normalized connection. The feature vector is (normalized rate, normalized throughput, normalized direction, normalized connection).

[0008] Furthermore, the step of extracting behavioral features of usage behaviors based on feature vectors corresponding to different usage behaviors, and distinguishing usage behaviors based on behavioral features, also includes the following sub-steps: Before being put into use, feature vectors of different usage behaviors are tested and named test behavior vectors. The behavioral characteristics of usage behaviors are analyzed based on the test behavior vectors of different usage behaviors, and usage behaviors are distinguished based on these behavioral characteristics.

[0009] Furthermore, the step of testing the feature vectors of different usage behaviors before putting them into use, and naming them test behavior vectors, also includes the following sub-steps: The usage behavior is manually categorized, including receiving behavior, sending behavior, and other behaviors. Before being put into use, the usage behavior is tested and learned, and the feature vectors of the usage behavior during the test are extracted and named test behavior vectors. The test behavior vectors include receiving vectors, sending vectors and other vectors.

[0010] Furthermore, the step of analyzing the behavioral characteristics of usage behaviors based on test behavior vectors of different usage behaviors and distinguishing usage behaviors based on these behavioral characteristics also includes the following sub-steps: The normalized rate, normalized throughput, normalized direction, and normalized connection in the test behavior vector are sorted in ascending order, using the symbol TC. n It indicates that the TC n In the expression TC, n is the index of TC and n is a non-zero natural number. Calculate TC. n / TC1, name the calculation result as the eigenratio and use the symbol TB n express; Using the sequence number n as the X-axis, and simultaneously using the characteristic ratio TB n Construct a feature extraction coordinate system for the Y-axis, and set TB n Enter the feature extraction coordinate system according to the serial number n, perform linear regression on the feature extraction coordinate system, and name the slope of the regression line as the behavior feature. Analyze and statistically analyze the range of behavioral characteristics of receiving behavior, and name it the receiving characteristic range; analyze and statistically analyze the range of behavioral characteristics of sending behavior, and name it the sending characteristic range; analyze and statistically analyze the range of behavioral characteristics of other behaviors, and name it the other characteristic range. Obtain the median values ​​of the receive feature range, transmit feature range, and other feature ranges, and name them as the receive feature median, transmit feature median, and other feature medians, respectively. Collectively, these are referred to as feature medians. Sort the usage behaviors according to the feature median values ​​in ascending order, using the symbol BV. h It indicates that the BV h In this context, h is the index of BV and h is a positive integer. BV... h Assign the value h; BV (Best Practices) h A feature partitioning coordinate system is constructed with the horizontal axis representing the feature and the vertical axis representing the behavioral features. The behavioral features are then partitioned according to their corresponding Body Values ​​(BV). h Enter the feature division coordinate system and divide BV h The coordinate point corresponding to this location is named point h; For each BV h Analyze the data and draw two auxiliary lines parallel to the horizontal axis, naming them the downward dividing line and the upward dividing line respectively. Then, divide the BV... h The corresponding downlink and uplink dividing lines are marked as DL. h and UL h , and DL h =UL h-1 UL h =DL h+1 ; When analyzing BV h With BV h+1 When crossing the boundary line between them, move DL h This makes DL h There is no point h-1 above, and UL is moved simultaneously. h+1 This makes UL h+1 There is no h+2 point below, move UL h Real-time statistics are in DL h With UL h The ratio of the number of h points to the number of h+1 points is named the h-th error ratio. Simultaneously, the DL is calculated. h+1 With UL h+1 The ratio of the number of h+1 points to the number of h points is named the h+1th error ratio. Calculate the sum of the h-th error ratio and the (h+1)-th error ratio, and name it the precision exponent. Take the UL when the precision exponent is maximized. h As BV h With BV h+1 The dividing line between them, similarly analyzed for all dividing lines, BV hThe area between the two boundary lines is named the feature area. The range of behavioral characteristics in the feature area is statistically analyzed and named the feature calibration range. The feature calibration range of each usage behavior is analyzed, and the feature calibration range of receiving behavior is named the receiving calibration range, the feature calibration range of sending behavior is named the sending calibration range, and the feature calibration range of other behaviors is named the other calibration range. If the analysis shows that the behavior characteristics of the IoT terminal are within the receiving calibration range, the current usage behavior is classified as receiving behavior; if the analysis shows that the behavior characteristics of the IoT terminal are within the transmitting calibration range, the current usage behavior is classified as transmitting behavior; if the analysis shows that the behavior characteristics of the IoT terminal are within other calibration ranges, the current usage behavior is classified as other behavior.

[0011] Furthermore, the step of classifying usage behavior according to its characteristics to obtain a response classification also includes the following sub-steps: When a user's behavioral characteristics are less than the minimum of the receiving characteristic range and the sending characteristic range, the response level at this time is set to passive level. When a user actively requests to connect to an IoT terminal, the response level at this time is set to active level.

[0012] Furthermore, the step of learning user behavior and summarizing usage habits for IoT terminals, and adaptively allocating autonomous wake-up intervals for IoT terminals in different time periods based on usage habits and response levels, also includes the following sub-steps: During the learning cycle, learn about users' usage behavior of IoT terminals and summarize users' usage habits; Based on usage habits and response levels, the autonomous wake-up interval is adaptively allocated to IoT terminals at different time periods.

[0013] Furthermore, the step of learning user behavior on IoT terminals and summarizing user habits during the learning cycle also includes the following sub-steps: When a user uses an IoT terminal for the first time, a learning period is preset. During the learning period, the IoT terminal remains powered on and simultaneously acquires the user's receiving, sending, and other behaviors in real time. The time when the user's behavior occurs is recorded and named the usage time. The time used for receiving and sending actions are named receiving time and sending time, respectively. Find the first sending time after the receiving time, calculate the time difference between the two, and name it the response time difference; The average response time difference is calculated and named the average wake-up frequency, which is the user's usage habit.

[0014] Furthermore, the step of adaptively allocating autonomous wake-up intervals for IoT terminals in different time periods based on usage habits and response levels also includes the following sub-steps: When the response level is set to passive, the IoT terminal enters a sleep state. Set the autonomous wake-up interval of the IoT terminal to the average wake-up frequency, and wake up the IoT terminal every average wake-up frequency.

[0015] The beneficial effects of this invention are as follows: This invention installs an ultra-lightweight processor within an IoT terminal to collect interaction data from the IoT terminal. The interaction data is then normalized to obtain organized data, which is packaged into feature vectors. Before deployment, feature vectors representing different usage behaviors are tested and named test behavior vectors. The behavioral characteristics of these test behavior vectors are then analyzed to differentiate usage behaviors. The advantage lies in the fact that the user's specific usage process and the software used are considered user privacy, and user behavior is usually not directly obtainable. However, user behavior can be roughly identified through the interaction data between the user and the portable Wi-Fi itself, without obtaining user privacy. By analyzing the behavioral characteristics through test behavior vectors, the behavioral characteristics reveal the characteristics of the interaction data for different usage behaviors, thereby indirectly judging user behavior and analyzing user habits in sending and receiving information. This improves the effectiveness and rationality of optimizing IoT terminal power consumption in portable Wi-Fi applications. This invention categorizes usage behavior into response levels based on its characteristics. Simultaneously, it learns user behavior regarding IoT terminals during a learning cycle and summarizes user habits. Finally, based on these habits and response levels, it adaptively allocates autonomous wake-up intervals for IoT terminals at different time periods. The advantage lies in the fact that message sending and receiving requires minimal data interaction, resulting in smaller behavioral characteristics. Even smaller behavioral characteristics indicate that the user is not using the smart device and does not require frequent network connections, placing the IoT terminal in a passive tier. In this case, the IoT terminal is put into sleep mode. Simultaneously, based on the previously analyzed user's message sending and receiving habits, a user-specific autonomous wake-up interval is assessed. The IoT terminal is actively woken up at each autonomous wake-up interval to allow the user to receive network messages, avoiding missing important information. This further improves the effectiveness and rationality of IoT terminal power consumption optimization in portable Wi-Fi applications.

[0016] Advantages of additional aspects of the invention will be set forth in part in the detailed description of the invention below, and in part will be obvious from the description or may be learned by practice of the invention. Attached Figure Description

[0017] Other features, objects, and advantages of the invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings: Figure 1 This is a flowchart illustrating the steps of the method of the present invention; Figure 2 This is a schematic diagram of the feature extraction coordinate system of the present invention; Figure 3 This is a schematic diagram of the feature division coordinate system of the present invention; Figure 4 This is a schematic diagram of the downlink and uplink dividing lines of the present invention; Figure 5 This is a schematic diagram of DL1 and UL2 after being moved according to the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Example 1, please refer to Figure 1 As shown, this application provides a power consumption optimization method for IoT terminals based on ultra-lightweight technology, including the following steps: Step S100: Install an ultra-lightweight processor in the IoT terminal to collect the interaction data of the IoT terminal; Step S100 also includes the following steps: Step S101: Install an ultra-lightweight processor in the IoT terminal; Step S102: Monitor the interaction data of the IoT terminal through the ultra-lightweight processor. The interaction data includes instantaneous packet rate, instantaneous throughput, traffic direction and connection status. In this specific implementation, the IoT terminal in this embodiment specifically refers to a portable Wi-Fi device. An ultra-lightweight processor is installed inside the portable Wi-Fi device to monitor the instantaneous packet rate, instantaneous throughput, traffic direction, and connection status generated by the interaction between the portable Wi-Fi device and the user. Here, instantaneous packet rate and instantaneous throughput are proper nouns. The connection status refers to the number of clients connected to the Wi-Fi device. The traffic direction is specifically the ratio of uplink data volume to downlink data volume. If the traffic direction is greater than 1, it means that the main data being uploaded at this time is such as uploading files. If the traffic direction is less than 1, it means that the main data being downloaded at this time is such as watching short videos. Under normal circumstances, the traffic direction is difficult to equal 1, that is, it is difficult to achieve a balance between sending and receiving. Therefore, no special explanation is given in this embodiment.

[0020] Step S200 involves normalizing the interactive data to obtain organized data, and then packaging the organized data into feature vectors. Step S200 also includes the following steps: Step S201: Use a normalization algorithm to calculate each type of interaction data independently, converting all different interaction data into values ​​from 0 to 1 to obtain the sorted data; Step S202: The data of instantaneous packet rate, instantaneous throughput, traffic direction and connection status are named in sequence as normalized rate, normalized throughput, normalized direction and normalized connection. Step S203, the feature vector is (normalized rate, normalized throughput, normalized direction, normalized connection); In practice, for example, if a user's interaction data is recorded for a month, the instantaneous packet rate in this month's interaction data ranges from a minimum of 0 to a maximum of 1200. If the instantaneous packet rate in a certain interaction data is 200, the normalized rate is calculated using a normalization algorithm as (200-0) / (1200-0) = 1 / 6. Similarly, the normalized rate, normalized throughput, normalized direction, and normalized connection in each interaction data are calculated. For example, if the normalized rate, normalized throughput, normalized direction, and normalized connection in a certain interaction data are 1 / 6, 0.26, 0.38, and 0.5 respectively, then the feature vector of this interaction data is (1 / 6, 0.26, 0.38, 0.5).

[0021] Step S300 involves extracting behavioral features of different usage behaviors based on their corresponding feature vectors, and then distinguishing usage behaviors based on these behavioral features. Step S300 also includes the following steps: Step S301: Before putting it into use, test the feature vectors of different usage behaviors and name them as test behavior vectors; Step S301 also includes the following steps: Step S301.1, the usage behavior is manually divided, including receiving behavior, sending behavior and other behaviors; Step S301.2: Before putting it into use, test and learn the usage behavior, extract the feature vector of the usage behavior during the test, and name it the test behavior vector. The test behavior vector includes the receiving vector, the sending vector, and other vectors. In practice, before being put into use, testers will simulate various usage behaviors and record the feature vectors corresponding to these usage behaviors to obtain test behavior vectors. The receiving vector is actually the feature vector generated when the user receives network information through the portable Wi-Fi, the sending vector is actually the feature vector generated when the user sends network information through the portable Wi-Fi, and the other vectors are actually the feature vectors generated when the user engages in other network activities such as audio-visual entertainment.

[0022] Step S302: Analyze the behavioral characteristics of the usage behavior based on the test behavior vectors of different usage behaviors and distinguish the usage behaviors based on the behavioral characteristics; In practice, when a user receives a network message while the phone screen is off, the network requests from various applications on the phone are relatively few, resulting in a lower overall reception vector. Similarly, when a user receives a network message while the phone screen is off and then turns on the phone to reply (i.e., sends a message), the user is only focused on the communication software, not other multimedia entertainment software, so the network requests are also relatively few, resulting in a lower overall transmission vector. However, when the user is using multimedia entertainment software, the network requests are more numerous, resulting in a higher overall transmission vector for other behaviors. If the user is using a communication device, there is no need to determine the user's usage behavior, as the portable Wi-Fi needs to remain awake. Therefore, this embodiment only considers the user's information reception needs while the screen is off.

[0023] Step S302 also includes the following steps: Step S302.1: Sort the normalized rate, normalized throughput, normalized direction, and normalized connection in the test behavior vector in ascending order, using the symbol TC. n TC n In the expression TC, n is the index of TC and n is a non-zero natural number. Calculate TC. n / TC1, name the calculation result as the eigenratio and use the symbol TB n express; Please see Figure 2 As shown, in step S302.2, the sequence number n is used as the X-axis, and the feature ratio TB is used as the X-axis. n Construct a feature extraction coordinate system for the Y-axis, and set TB n Enter the feature extraction coordinate system according to the serial number n, perform linear regression on the feature extraction coordinate system, and name the slope of the regression line as the behavior feature. Step S302.3: Analyze and statistically analyze the range of behavioral characteristics of receiving behavior, and name it the receiving characteristic range; analyze and statistically analyze the range of behavioral characteristics of sending behavior, and name it the sending characteristic range; analyze and statistically analyze the range of behavioral characteristics of other behaviors, and name it the other characteristic range. Step S302.4: Obtain the median of the receiving feature range, the transmitting feature range, and other feature ranges, and name them as the receiving feature median, transmitting feature median, and other feature medians, respectively, collectively referred to as the feature median. Sort the usage behaviors according to the feature median of the usage behaviors in ascending order, and use the symbol BV. h BV indicates h In this context, h is the index of BV and h is a positive integer. BV... h Assign the value h; In specific implementation, for example, if a test behavior vector is (1 / 6, 0.26, 0.38, 0.5), after sorting, TC1 to TC4 are obtained as 1 / 6, 0.26, 0.38, and 0.5 respectively. TB1 to TB4 are calculated as 1, 1.56, 2.28, and 3 respectively. The feature extraction coordinate system is then constructed as follows: Figure 2 As shown, further regression yields a behavior feature of 0.672 for this test behavior vector. Similarly, analyzing the behavior features of all test behavior vectors, we obtain the range of receiving features as [0.268, 0.642], the range of sending features as [0.598, 0.995], and the range of other features as [0.846, 1.463]. The corresponding median values ​​of receiving features, sending features, and other features are 0.455, 0.7965, and 1.1545, respectively. Sorting, we obtain BV1 as receiving behavior, BV2 as sending behavior, and BV3 as other behavior. At this point, we assign BV1 a value of 1, BV2 a value of 2, and BV3 a value of 3. That is, in the feature partitioning coordinate system, a vertical axis value of 1 represents receiving behavior, a vertical axis value of 2 represents sending behavior, and a vertical axis value of 3 represents other behavior.

[0024] Please see Figure 3 As shown, in step S302.5, the behavior BV is used. h A feature partitioning coordinate system is constructed with the horizontal axis representing the feature and the vertical axis representing the behavioral features. The behavioral features are then partitioned according to their corresponding Body Values ​​(BV). h Enter the feature division coordinate system and divide BV h The coordinate point corresponding to this location is named point h; Please see Figure 4 As shown, in step S302.6, for each BV h Analyze the data and draw two auxiliary lines parallel to the horizontal axis, naming them the downward dividing line and the upward dividing line respectively. Then, divide the BV... h The corresponding downlink and uplink dividing lines are marked as DL. h and UL h , and DL h =UL h-1 UL h =DL h+1 ; In specific implementation, the feature partitioning coordinate system is constructed as follows: Figure 3 As shown, the entered coordinates yield points 1, 2, and 3, corresponding to receiving, sending, and other actions, respectively. For ease of understanding, in this embodiment, point 1 is referred to as the receiving point, point 2 as the sending point, and point 3 as the other points. Downlink and uplink dividing lines are drawn for BV1, BV2, and BV3 as shown below. Figure 4 As shown.

[0025] Please see Figure 5As shown, in step S302.7, when analyzing BV... h With BV h+1 When crossing the boundary line between them, move DL h This makes DL h There is no point h-1 above, and UL is moved simultaneously. h+1 This makes UL h+1 There is no h+2 point below, move UL h Real-time statistics are in DL h With UL h The ratio of the number of h points to the number of h+1 points is named the h-th error ratio. Simultaneously, the DL is calculated. h+1 With UL h+1 The ratio of the number of h+1 points to the number of h points is named the h+1th error ratio. Step S302.8: Calculate the sum of the h-th error ratio and the (h+1)-th error ratio, name it the accuracy index, and take the UL when the accuracy index is maximum. h As BV h With BV h+1 The dividing line between them, similarly analyzed for all dividing lines, BV h The area between the two boundary lines is named the feature area. The range of behavioral characteristics in the feature area is statistically analyzed and named the feature calibration range. The feature calibration range of each usage behavior is analyzed, and the feature calibration range of receiving behavior is named the receiving calibration range, the feature calibration range of sending behavior is named the sending calibration range, and the feature calibration range of other behaviors is named the other calibration range. Step S302.9: If the analysis shows that the behavior characteristics of the IoT terminal are within the receiving calibration range, then the current usage behavior is classified as receiving behavior; if the analysis shows that the behavior characteristics of the IoT terminal are within the transmitting calibration range, then the current usage behavior is classified as transmitting behavior; if the analysis shows that the behavior characteristics of the IoT terminal are within other calibration ranges, then the current usage behavior is classified as other behavior. In practical implementation, taking BV1 and BV2 as examples, when DL1 is at the bottom, the upper side includes all receiving points, and there are no other coordinate points at the bottom. At this time, there is no need to worry about the specific position of DL1, because there is no point with h=0 below BV1. It is only necessary to ensure that DL1 is below all receiving points. Then, UL2 is moved vertically so that all other points are above UL2, that is, there are no other points below UL2. The moved DL1 and UL2 are as follows: Figure 5As shown, UL1 is then moved vertically up and down, the range of which represents the intersection of the receiving and transmitting feature ranges. During this movement, the ratio of receiving points to transmitting points is continuously calculated among the coordinates above DL1 and below UL1. Simultaneously, the ratio of transmitting points to receiving points above DL2 and below UL2 is also calculated. The final calculated error ratios are 8 for the first error and 7.5 for the second error. Further calculation yields an accuracy index of 15.5. This accuracy index is the highest among all accuracy indexes. The maximum value in the accuracy index is used, so UL1 with an accuracy index of 15.5 is taken as the boundary between BV1 and BV2. Similarly, the boundary between BV2 and BV3 is analyzed. Finally, the receiving calibration range is [0.268, 0.622), the transmitting calibration range is [0.622, 0.938), and the other calibration ranges are [0.938, 1.463]. Then, the user's current usage behavior can be determined based on the range of the behavioral characteristics. This has been explained in detail in step S302.9 and will not be explained again here.

[0026] Step S400: Classify the usage behavior according to its characteristics to obtain a response classification; Step S400 also includes the following steps: Step S401: When the user's behavioral characteristics are less than the minimum of the receiving characteristic range and the sending characteristic range, the response level at this time is set to passive level. Step S402: When a user actively requests to connect to an IoT terminal, the response level at this time is set to active level. In practice, when the user does not operate the smart device, the smart device only makes network requests such as receiving information. The frequency and intensity of interaction between the smart device and the portable Wi-Fi are reduced. If no network information is received, the behavioral characteristics will be further reduced. Therefore, when the behavioral characteristics are less than the minimum value of the receiving characteristics range, it is determined that the user does not have a need to use the smart device. The response level at this time is set to passive level, which means that the user does not have a need to actively use the portable Wi-Fi. When the user turns on the smart device, the smart device will automatically request network from the portable Wi-Fi, that is, actively wake up. The response level at this time is active level, which means that the user actively operates. When the user actively operates, there is no need for power consumption optimization in this embodiment.

[0027] Step S500 involves learning user behavior and summarizing usage habits for IoT terminals, and adaptively allocating autonomous wake-up intervals to IoT terminals at different time periods based on usage habits and response levels. Step S500 also includes the following steps: Step S501: During the learning cycle, learn the user's usage behavior of the IoT terminal and summarize the user's usage habits; Step S501 also includes the following steps: Step S501.1: When a user uses the IoT terminal for the first time, a learning period is preset. During the learning period, the IoT terminal remains in the startup state and simultaneously acquires the user's receiving behavior, sending behavior and other behaviors in real time. The time when the user's usage behavior occurs is recorded and named as the usage time. Step S501.2: Name the time used for receiving and sending actions as receiving time and sending time, respectively. Step S501.3: Find the first sending time after the receiving time, calculate the time difference between the two, and name it the reply time difference; Step S501.4: Calculate the average value of the response time difference and name it the average wake-up frequency. The average wake-up frequency is the user's usage habit. In practice, the learning period is set by the manufacturer. In this embodiment, the learning period is set to 14 days. During the first 14 days of use, the portable Wi-Fi operates according to its existing working mode, without relying on power consumption optimization as described in this embodiment. During this period, the user's usage behavior and time are recorded. For example, if a user receives data at 13:55:36 on a certain day and sends data at 13:59:48 on the same day, with no other sending data in between, the response time difference is 4 minutes and 12 seconds. It should be noted that if multiple consecutive receiving and sending data occur... If the sending time corresponds to the receiving time, the longest response time difference is used for calculation. For example, if a user receives data once at 13:55:36, 13:56:55, and 13:56:59 on the same day, and the first sending time after each of these three receiving times is 13:59:48, then only the response time difference of 4 minutes and 12 seconds is recorded, ignoring the response time differences corresponding to 13:56:55 and 13:56:59. The average response time difference within the user's learning cycle is calculated to obtain the user's average wake-up frequency. The average wake-up frequency of each user is independent.

[0028] Step S502: Based on usage habits and response levels, adaptively allocate autonomous wake-up intervals for IoT terminals at different time periods; Step S502 also includes the following steps: Step S502.1: When the response level is passive, the IoT terminal is put into a sleep state. Step S502.2: Set the autonomous wake-up interval of the IoT terminal to the average wake-up frequency, and wake up the IoT terminal every average wake-up frequency. In practice, for example, if the average wake-up frequency of a user is 3 minutes, when the response analysis is passively graded, the portable Wi-Fi will enter a sleep state and wake up once every 3 minutes, with each wake-up lasting 5 seconds, so that the user's smart device can receive network messages.

[0029] Example 2: This application provides an electronic device, which may include a processor, a communication interface, a memory, and a communication bus. The processor, communication interface, and memory communicate with each other via the communication bus. The memory stores computer-readable instructions, and the processor can call these instructions. When the processor executes a computer-readable instruction, it performs steps as described in the IoT terminal power consumption optimization method based on ultra-lightweight technology to achieve the following functions: collecting interaction data from the IoT terminal; normalizing the interaction data to obtain organized data, and packaging the organized data into feature vectors; extracting behavioral features of different usage behaviors based on the feature vectors corresponding to those behaviors, and distinguishing usage behaviors based on these behavioral features; classifying usage behaviors according to their characteristics to obtain response levels; learning user behavior towards the IoT terminal and summarizing usage habits, and adaptively allocating autonomous wake-up intervals for the IoT terminal at different time periods.

[0030] Furthermore, the logical instructions in the aforementioned memory can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application.

[0031] Example 3: This application also provides a computer program product, which includes a computer program stored on a computer-readable storage medium. The computer program includes program instructions. When the program instructions are executed by a computer, the computer can execute the IoT terminal power consumption optimization method based on ultra-lightweight technology provided by the above methods. The method includes: collecting interaction data of IoT terminals; normalizing the interaction data to obtain sorted data, and packaging the sorted data into feature vectors; extracting behavioral features of usage behaviors based on the feature vectors corresponding to different usage behaviors, and distinguishing usage behaviors based on behavioral features; classifying usage behaviors according to their characteristics to obtain response levels; learning user usage behaviors of IoT terminals and summarizing usage habits, and adaptively allocating autonomous wake-up intervals for IoT terminals in different time periods.

[0032] Example 4: This application may also provide a computer-readable storage medium. When the instructions in the computer-readable storage medium are executed by the processor of a server, the server is able to perform and implement the following functions: collect interactive data from IoT terminals; normalize the interactive data to obtain organized data, and package the organized data into feature vectors; extract behavioral features of usage behaviors based on the feature vectors corresponding to different usage behaviors, and distinguish usage behaviors based on behavioral features; classify usage behaviors according to their characteristics to obtain response levels; learn user usage behaviors of IoT terminals and summarize usage habits, and adaptively allocate autonomous wake-up intervals for IoT terminals in different time periods.

[0033] Furthermore, the computer-readable storage medium provided in this application may be: read-only memory (ROM), random access memory (RAM), random access programmable read-only memory (PROM), electrically erasable programmable read-only memory (EEPROM), dynamic random access memory (DRAM), static random access memory (SRAM), flash memory, non-volatile memory, CD-ROM, CD-R, CD+R, CD-RW, CD+RW, DVD-ROM, DVD-R, DVD+R, DVD-RW, DVD+RW, DVD-RAM, BD-ROM, BD-R, BD-RL The computer program in the aforementioned computer-readable storage medium can run in an environment deployed in computer devices such as clients, hosts, agents, and servers. Furthermore, in one example, the computer program and any associated data, data files, and data structures are distributed across a networked computer system, such that the computer program and any associated data, data files, and data structures are stored, accessed, and executed in a distributed manner through one or more processors or computers.

[0034] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media containing computer-usable program code. These computer program instructions can also be stored in a computer-readable storage medium capable of directing a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means implemented in a process. Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0035] The above-described embodiments are merely specific implementations of the present invention, used to illustrate the technical solutions of the present invention, and are not intended to limit it. The scope of protection of the present invention is not limited thereto. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that any person skilled in the art can still modify or easily conceive of changes to the technical solutions described in the foregoing embodiments within the scope of the technology disclosed in the present invention, or make equivalent substitutions for some of the technical features; and these modifications, changes, or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of protection of the claims.

Claims

1. A power consumption optimization method for IoT terminals based on ultra-lightweight technology, characterized in that, The steps include the following: An ultra-lightweight processor is installed in the IoT terminal to collect the interaction data of the IoT terminal; The interaction data is normalized to obtain sorted data, and the sorted data is packaged into feature vectors. Behavioral features of different usage behaviors are extracted from the feature vectors corresponding to those behaviors, and usage behaviors are distinguished based on these behavioral features. Based on the characteristics of the usage behavior, the usage behavior is classified into different levels, resulting in response levels; Learn user behavior and summarize usage habits when using IoT terminals. Based on usage habits and response levels, adaptively allocate autonomous wake-up intervals for IoT terminals in different time periods.

2. The method for optimizing the power consumption of IoT terminals based on ultra-lightweight technology according to claim 1, characterized in that, The step of installing an ultra-lightweight processor in the IoT terminal and collecting interactive data from the IoT terminal further includes the following sub-steps: Install ultra-lightweight processors in IoT terminals; The interaction data of IoT terminals is monitored through an ultra-lightweight processor. The interaction data includes instantaneous packet rate, instantaneous throughput, traffic direction, and connection status.

3. The method for optimizing the power consumption of IoT terminals based on ultra-lightweight technology according to claim 2, characterized in that, The step of normalizing the interactive data to obtain processed data and packaging the processed data into feature vectors further includes the following sub-steps: A normalization algorithm is used to calculate each type of interaction data independently, converting all different interaction data into values ​​from 0 to 1, resulting in organized data; The compiled data of instantaneous packet rate, instantaneous throughput, traffic direction, and connection status are named in sequence as normalized rate, normalized throughput, normalized direction, and normalized connection. The feature vector is (normalized rate, normalized throughput, normalized direction, normalized connection).

4. The method for optimizing the power consumption of IoT terminals based on ultra-lightweight technology according to claim 3, characterized in that, The step of extracting behavioral features of usage behaviors based on feature vectors corresponding to different usage behaviors, and distinguishing usage behaviors based on behavioral features, further includes the following sub-steps: Before being put into use, feature vectors of different usage behaviors are tested and named test behavior vectors. The behavioral characteristics of usage behaviors are analyzed based on the test behavior vectors of different usage behaviors, and usage behaviors are distinguished based on these behavioral characteristics.

5. The method for optimizing the power consumption of IoT terminals based on ultra-lightweight technology according to claim 4, characterized in that, The step of testing the feature vectors of different usage behaviors before putting them into use, and naming them test behavior vectors, also includes the following sub-steps: The usage behavior is manually categorized, including receiving behavior, sending behavior, and other behaviors. Before being put into use, the usage behavior is tested and learned, and the feature vectors of the usage behavior during the test are extracted and named test behavior vectors. The test behavior vectors include receiving vectors, sending vectors and other vectors.

6. The method for optimizing the power consumption of an IoT terminal based on ultra-lightweight technology according to claim 5, characterized in that, The step of analyzing the behavioral characteristics of usage behaviors based on test behavior vectors of different usage behaviors and distinguishing usage behaviors based on these behavioral characteristics further includes the following sub-steps: The normalized rate, normalized throughput, normalized direction, and normalized connection in the test behavior vector are sorted in ascending order, using the symbol TC. n It indicates that the TC n In the expression TC, n is the index of TC and n is a non-zero natural number. Calculate TC. n / TC1, name the calculation result as the eigenratio and use the symbol TB n express; Using the sequence number n as the X-axis, and simultaneously using the characteristic ratio TB n Construct a feature extraction coordinate system for the Y-axis, and set TB n Enter the feature extraction coordinate system according to the serial number n, perform linear regression on the feature extraction coordinate system, and name the slope of the regression line as the behavior feature. The range of behavioral characteristics of receiving behavior is analyzed and statistically analyzed, and named as the receiving characteristic range; the range of behavioral characteristics of sending behavior is analyzed and statistically analyzed, and named as the sending characteristic range; the range of behavioral characteristics of other behaviors is analyzed and statistically analyzed, and named as the other characteristic range. Obtain the median values ​​of the receive feature range, transmit feature range, and other feature ranges, and name them as the receive feature median, transmit feature median, and other feature medians, respectively. Collectively, these are referred to as feature medians. Sort the usage behaviors according to the feature median values ​​in ascending order, using the symbol BV. h It indicates that the BV h In this context, h is the index of BV and h is a positive integer. BV... h Assign the value h; BV (Best Practices) h A feature partitioning coordinate system is constructed with the horizontal axis representing the feature and the vertical axis representing the behavioral features. The behavioral features are then partitioned according to their corresponding Body Values ​​(BV). h Enter the feature division coordinate system and divide BV h The coordinate point corresponding to this location is named point h; For each BV h Analyze the data and draw two auxiliary lines parallel to the horizontal axis, naming them the downward dividing line and the upward dividing line respectively. Then, divide the BV... h The corresponding downlink and uplink dividing lines are marked as DL. h and UL h , and DL h =UL h-1 UL h =DL h+1 ; When analyzing BV h With BV h+1 When crossing the boundary line between them, move DL h This makes DL h There is no point h-1 above, and UL is moved simultaneously. h+1 This makes UL h+1 There is no h+2 point below, move UL h Real-time statistics are in DL h With UL h The ratio of the number of h points to the number of h+1 points is named the h-th error ratio. Simultaneously, the DL (details of the error ratio) is calculated. h+1 With UL h+1 The ratio of the number of h+1 points to the number of h points is named the h+1th error ratio. Calculate the sum of the h-th error ratio and the (h+1)-th error ratio, and name it the precision exponent. Take the UL when the precision exponent is maximized. h As BV h With BV h+1 The dividing line between them, similarly analyzed for all dividing lines, BV h The area between the two boundary lines is named the feature area. The range of behavioral characteristics in the feature area is statistically analyzed and named the feature calibration range. The feature calibration range of each usage behavior is analyzed, and the feature calibration range of receiving behavior is named the receiving calibration range, the feature calibration range of sending behavior is named the sending calibration range, and the feature calibration range of other behaviors is named the other calibration range. If the analysis shows that the behavior characteristics of the IoT terminal are within the receiving calibration range, then the current usage behavior is classified as receiving behavior; if the analysis shows that the behavior characteristics of the IoT terminal are within the transmitting calibration range, then the current usage behavior is classified as transmitting behavior. If the analysis shows that the behavior characteristics of the IoT terminal are within other calibration ranges, then the current usage behavior will be categorized as other behaviors.

7. The method for optimizing the power consumption of an IoT terminal based on ultra-lightweight technology according to claim 6, characterized in that, The step of classifying usage behavior according to its characteristics to obtain a response classification further includes the following sub-steps: When a user's behavioral characteristics are less than the minimum of the receiving characteristic range and the sending characteristic range, the response level at this time is set to passive level. When a user actively requests to connect to an IoT terminal, the response level at this time is set to active level.

8. The method for optimizing the power consumption of an IoT terminal based on ultra-lightweight technology according to claim 7, characterized in that, The step of learning user behavior and summarizing usage habits for IoT terminals, and adaptively allocating autonomous wake-up intervals for IoT terminals in different time periods based on usage habits and response levels, also includes the following sub-steps: During the learning cycle, learn about users' usage behavior of IoT terminals and summarize users' usage habits; Based on usage habits and response levels, the autonomous wake-up interval is adaptively allocated to IoT terminals at different time periods.

9. The method for optimizing the power consumption of an IoT terminal based on ultra-lightweight technology according to claim 8, characterized in that, The steps of learning user behavior on IoT terminals and summarizing user habits during the learning cycle also include the following sub-steps: When a user uses an IoT terminal for the first time, a learning period is preset. During the learning period, the IoT terminal remains powered on and simultaneously acquires the user's receiving, sending, and other behaviors in real time. The time when the user's behavior occurs is recorded and named the usage time. The time used for receiving and sending actions are named receiving time and sending time, respectively. Find the first sending time after the receiving time, calculate the time difference between the two, and name it the response time difference; The average response time difference is calculated and named the average wake-up frequency, which is the user's usage habit.

10. The method for optimizing the power consumption of an IoT terminal based on ultra-lightweight technology according to claim 9, characterized in that, The step of adaptively allocating autonomous wake-up intervals for IoT terminals in different time periods based on usage habits and response levels also includes the following sub-steps: When the response level is set to passive, the IoT terminal enters a sleep state. Set the autonomous wake-up interval of the IoT terminal to the average wake-up frequency, and wake up the IoT terminal every average wake-up frequency.

Images