Method, device and medium for low-power temperature anomaly alarm based on wireless communication

By collecting temperature data at multiple time points in the temperature alarm device and generating Bluetooth Low Energy broadcast packets, the problems of high power consumption and strong dependence of existing devices are solved, and low power consumption and high reliability temperature anomaly alarm are realized.

CN123111490APending Publication Date: 2026-11-17SHENZHEN INKBIRD TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611477799.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Priority Date
2026-01-09
Filing Date
2026-09-21
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing temperature alarm devices have high power consumption in standby mode and are highly dependent on the terminal. They cannot send alarm information in a timely manner when the mobile phone is offline or the connection is lost. Furthermore, they lack multi-dimensional analysis of temperature trends and change rates, which leads to false alarms or missed alarms.

Method used

Temperature data is collected at multiple time points by multiple temperature probes, target temperature data is extracted, target features are generated using a preset feature extraction method, and Bluetooth low power broadcast packets are generated to trigger an alarm when the alarm threshold is reached. The system adopts an event-triggered working mode, and the main control unit remains in deep sleep when not under alarm conditions.

Benefits of technology

It reduces standby power consumption, improves the accuracy and reliability of alarms, and ensures that it can still issue timely temperature abnormality warnings when the mobile phone is not present or the connection is lost. It is suitable for long-term battery-powered and high-security application scenarios.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123111490A_ABST
    Figure CN123111490A_ABST
Patent Text Reader

Abstract

This application relates to the field of low-power temperature anomaly alarm technology based on wireless communication, and discloses a method, device, and medium for low-power temperature anomaly alarm based on wireless communication. The method includes: collecting temperature data at multiple time points and generating target features based on a preset feature extraction method, achieving a comprehensive judgment of key characteristics such as temperature trends and rates of change. Compared with the traditional single-point threshold method, this significantly improves alarm accuracy. Furthermore, when the target features reach an alarm threshold, a broadcast packet is generated and Bluetooth Low Energy broadcast is executed. The beneficial effects of this invention are: alarms can be completed without maintaining a Bluetooth connection, thus completely solving the high power consumption problem caused by continuous connection in traditional solutions, and enabling the device to quickly issue temperature anomaly warnings even when the mobile phone is not present or the connection is lost.
Need to check novelty before this filing date? Find Prior Art

Description

[0001] This application claims priority to Chinese Patent Application No. 202610031159.2, filed on January 9, 2026, entitled "Method, Device and Medium for Low Power Temperature Anomaly Alarm Based on Wireless Communication", the entire contents of which are incorporated herein by reference. Technical Field

[0002] This invention relates to the field of low-power temperature anomaly alarm technology based on wireless communication, and particularly to a method, device and medium for low-power temperature anomaly alarm based on wireless communication. Background Technology

[0003] Temperature anomaly monitoring systems are widely used in battery management, warehousing and logistics, and home security. Existing temperature alarm devices typically rely on Bluetooth or Wi-Fi networks for real-time communication with mobile terminals. In traditional solutions, maintaining a continuous connection with the mobile terminal requires keeping the Bluetooth link active, resulting in high current consumption even in standby mode. Typical standby power consumption is often greater than 1 mA, which is insufficient for prolonged battery-powered operation. Furthermore, existing solutions are heavily reliant on the terminal; if the phone is offline, too far away, or not connected, alarm information cannot be reported promptly, significantly reducing alarm reliability. In addition, traditional solutions mostly rely on simple threshold judgments based on single-point temperature data, lacking the ability to analyze multi-dimensional characteristics such as temperature trends and rates of change, potentially leading to false alarms or missed alarms. Summary of the Invention

[0004] Therefore, it is necessary to address the existing problem of low-power temperature anomaly alarm based on wireless communication and propose a method, device, and medium for low-power temperature anomaly alarm based on wireless communication.

[0005] A method for low-power temperature anomaly alarm based on wireless communication, the method comprising: Temperature data was continuously collected at multiple time points using multiple temperature probes. Extract target temperature data from multiple preset time points closest to the current time point from the acquired temperature data; The target temperature data are arranged in chronological order to obtain the first target data sequence; The target features of the first target sequence are extracted using a preset feature extraction method; Determine whether the target feature has reached the alarm threshold; If the target feature reaches the alarm threshold, a broadcast packet is generated based on the target feature and an alarm is broadcast.

[0006] Further, the step of extracting the target features of the first target sequence using a preset feature extraction method includes: The first target sequence is transformed using a preset weight matrix to obtain the second target sequence; The second target sequence is encoded by a preset first encoder to obtain a first data feature, and the second target sequence is encoded by a preset second encoder to obtain a second data feature; wherein, the first encoder is used to extract the first data feature of the column data of the second target sequence, and the second encoder is used to extract the second data feature of the horizontal data of the second target sequence. According to a preset data fusion method, the first data feature and the second data feature are fused to obtain the target feature.

[0007] Further, the step of fusing the first data feature and the second data feature according to a preset data fusion method to obtain the target feature includes: The first data feature and the second data feature are concatenated according to a preset feature dimension to form a fusion vector; The fused vector is linearly mapped using a preset linear transformation matrix to obtain the mapping features; The mapped features are processed nonlinearly according to a preset nonlinear activation function to obtain the target features.

[0008] Further, the step of transforming the first target sequence using a preset weight matrix to obtain the second target sequence includes: Using the first target sequence as the input vector, the input vector is multiplied by a preset weight matrix to obtain the matrix operation result. The matrix operation results are numerically summed according to a preset bias vector to form an intermediate sequence; The intermediate sequence is processed element by element according to a preset linear or nonlinear activation rule to obtain the second target sequence.

[0009] Furthermore, the step of generating a broadcast packet and broadcasting an alarm based on the target feature if the target feature reaches the alarm threshold includes: Based on the target features, the temperature numerical parameters and temperature status code are determined according to a preset mapping relationship; According to the preset broadcast protocol format, a broadcast packet is generated based on the temperature numerical parameters and the temperature status code; The broadcast packets are continuously transmitted via the Bluetooth Low Energy Broadcast channel at preset broadcast intervals.

[0010] Furthermore, before the step of extracting target temperature data from multiple preset time points closest to the current time point from the acquired temperature data, the method further includes: Determine whether any of the target temperature values ​​is greater than a preset temperature value; If any of the target temperature values ​​is greater than a preset temperature value, the main control unit connected to the temperature probe is activated to execute the step of extracting the target temperature data of multiple preset time points closest to the current time point from the acquired temperature data.

[0011] Furthermore, after the step of generating a broadcast packet and broadcasting an alarm based on the target feature if the target feature reaches the alarm threshold, the method further includes: Continuously collect gradient temperature data at multiple time points; Detect whether all the gradient temperature data are lower than the preset temperature value and continuously meet the preset judgment period; If all the gradient temperature data are lower than the preset temperature value and the preset judgment period is met continuously, the broadcast alarm will be stopped and the main control unit will re-enter deep sleep mode.

[0012] A low-power temperature anomaly alarm device based on wireless communication, the device comprising: The temperature data acquisition module is used to continuously acquire temperature data at multiple time points through multiple temperature probes. The target temperature data extraction module is used to extract target temperature data from multiple preset time points closest to the current time point from the acquired temperature data. The target temperature data arranging module is used to arrange the target temperature data in chronological order to obtain a first target data sequence; The target feature extraction module is used to extract the target features of the first target sequence using a preset feature extraction method; The judgment module is used to determine whether the target feature has reached the alarm threshold; The broadcast packet generation module is used to generate a broadcast packet based on the target feature and broadcast an alarm if the target feature reaches the alarm threshold.

[0013] An electronic device includes a memory and a processor, the memory storing a computer program that, when executed by the processor, causes the processor to perform the following steps: Temperature data was continuously collected at multiple time points using multiple temperature probes. Extract target temperature data from multiple preset time points closest to the current time point from the acquired temperature data; The target temperature data are arranged in chronological order to obtain the first target data sequence; The target features of the first target sequence are extracted using a preset feature extraction method; Determine whether the target feature has reached the alarm threshold; If the target feature reaches the alarm threshold, a broadcast packet is generated based on the target feature and an alarm is broadcast.

[0014] A computer-readable storage medium storing a computer program, which, when executed by a processor, causes the processor to perform the following steps: Temperature data was continuously collected at multiple time points using multiple temperature probes. Extract target temperature data from multiple preset time points closest to the current time point from the acquired temperature data; The target temperature data are arranged in chronological order to obtain the first target data sequence; The target features of the first target sequence are extracted using a preset feature extraction method; Determine whether the target feature has reached the alarm threshold; If the target feature reaches the alarm threshold, a broadcast packet is generated based on the target feature and an alarm is broadcast.

[0015] The beneficial effects of this invention are as follows: By collecting temperature data at multiple time points and generating target features based on a preset feature extraction method, a comprehensive judgment of key characteristics such as temperature trends and rate of change is achieved. Compared with the traditional single-point threshold method, the accuracy of alarms can be significantly improved. In addition, by generating a broadcast packet and executing Bluetooth Low Energy broadcast when the target feature reaches the alarm threshold, the alarm can be completed without maintaining a Bluetooth connection, thereby completely solving the high power consumption problem caused by continuous connection in the traditional solution, and enabling the device to quickly issue temperature abnormality warnings even when the mobile phone is not present or the connection is disconnected.

[0016] Furthermore, this invention adopts an event-triggered working mode, keeping the main control unit in deep sleep when not under alarm conditions, and only waking it up to process when the temperature changes abnormally. The overall solution significantly reduces standby power consumption, improves the battery life and alarm reliability of the temperature monitoring device, and is suitable for long-term battery-powered and high-safety application scenarios. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] in: Figure 1 This is an application environment diagram of a low-power temperature anomaly alarm method based on wireless communication in one embodiment. Figure 2 This is a flowchart of a low-power temperature anomaly alarm method based on wireless communication in one embodiment. Figure 3 This is a structural block diagram of a low-power temperature anomaly alarm device based on wireless communication in one embodiment. Figure 4 This is a structural block diagram of an electronic device in one embodiment. Detailed Implementation

[0019] 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.

[0020] Figure 1 This is a diagram illustrating an application environment for a low-power temperature anomaly alarm based on wireless communication in one embodiment. (Refer to...) Figure 1 This method for low-power temperature anomaly alarm based on wireless communication is applied to a low-power temperature anomaly alarm system based on wireless communication. The system includes a terminal 110 and a server 120. The terminal 110 and server 120 are connected via a network. The terminal 110 can be a desktop terminal or a mobile terminal; a mobile terminal can be at least one of a mobile phone, tablet, or laptop. The server 120 can be a standalone server or a server cluster consisting of multiple servers. The terminal 110 is used to collect temperature data, and the server 120 is used to generate broadcast packets.

[0021] like Figure 2 As shown, in one embodiment, a method for implementing low-power temperature anomaly alarm based on wireless communication is provided. This method can be applied to both terminals and servers; this embodiment uses a terminal application as an example. The method for implementing low-power temperature anomaly alarm based on wireless communication specifically includes the following steps: S1: Continuously collect temperature data at multiple time points using multiple temperature probes; S2: Extract target temperature data from the acquired temperature data for multiple preset time points closest to the current time point; S3: Arrange the target temperature data in chronological order to obtain the first target data sequence; S4: Extract the target features of the first target sequence using a preset feature extraction method; S5: Determine whether the target feature has reached the alarm threshold; S6: If the target feature reaches the alarm threshold, a broadcast packet is generated based on the target feature and an alarm is broadcast.

[0022] As described in step S1 above, temperature data at multiple time points is continuously collected using multiple temperature probes. Real-time temperature monitoring is performed at specified time intervals using multiple temperature probes (e.g., NTC thermistors or digital temperature sensors). These probes can be distributed in different environmental locations to facilitate multi-point data collection, thereby ensuring the comprehensiveness and accuracy of statistical data. The collected temperature data is stored in memory, forming a time series. The continuous collection time can be customized, such as once per second or once per minute, to meet the needs of different application scenarios.

[0023] As described in step S2 above, target temperature data from multiple preset time points closest to the current time point is extracted from the acquired temperature data. This involves filtering the previously collected temperature data to select data from multiple preset time points closest to the current time point. Specifically, this can rely on timestamps or time markers. By comparing the current real-time time with historical data timestamps, the required target temperature data can be quickly located. The extracted target temperature data can be temperature values ​​from several time points. For different application scenarios, different numbers of historical data points may be required. For example, temperature data for each minute within the last 10 minutes or the last 5 consecutive data points can be selected. By extracting the most recent target data, the system can effectively reflect the short-term temperature change trend, providing the necessary information foundation for subsequent data analysis and feature extraction.

[0024] As described in step S3 above, the target temperature data is arranged in chronological order to obtain the first target data series. The data is typically arranged from the earliest time point to the latest time point. Time series analysis allows the system to capture rising or falling temperature trends, as well as potential abnormal fluctuations. The construction of this data series allows for subsequent deep learning using methods such as statistical analysis and machine learning to extract valuable operational features.

[0025] As described in step S4 above, target features of the first target data series are extracted using a preset feature extraction method. This involves using the preset feature extraction method to perform in-depth analysis of the already arranged first target data series to extract key target features. These features may include statistical features such as temperature change rate, mean, and standard deviation, or other derived indicators generated by time series algorithms (e.g., recent temperature fluctuations). The feature extraction method used can be based on traditional statistical calculations or utilize machine learning algorithms, such as regression analysis and neural networks. After feature extraction, the system can obtain a set of concise numerical indicators. These indicators will help with subsequent alarm decisions. By efficiently extracting these features, the computational complexity of subsequent calculations can be effectively reduced, while improving the accuracy and interpretability of the analysis results.

[0026] As described in step S5 above, it is determined whether the target feature has reached the alarm threshold. The extracted target feature is evaluated to determine whether it exceeds the set alarm threshold. The alarm threshold is usually set based on previous data analysis, considering safety standards or specifications in specific application scenarios such as HVAC and cold chain logistics. If the value of the target feature is greater than or equal to the set alarm threshold, the system will determine that the temperature is abnormal and trigger the subsequent alarm mechanism. This judgment process not only requires simple boundary checks, but can also use various threshold detection algorithms (such as dynamic thresholds or statistical analysis-based methods) to improve the accuracy and feasibility of the judgment. In addition, in some advanced applications, trend analysis based on historical data and machine learning models can also be used to dynamically adjust the alarm threshold, thereby improving the sensitivity and intelligence of the system in more complex environments. If the target feature does not reach the alarm threshold, the system will continue to monitor the current data and maintain a low-power operating state.

[0027] As described in step S6 above, if the target feature reaches the alarm threshold, a broadcast packet is generated based on the target feature and an alarm is broadcast. That is, when it is confirmed that the target feature exceeds the preset alarm threshold, the broadcast alarm stage begins. During this process, a custom broadcast packet is generated based on the target feature data. The broadcast packet contains necessary information, such as device address, temperature status, real-time temperature value, and necessary verification information. Specifically, the broadcast packet must be constructed according to a specific protocol format to ensure that it can be received by devices supporting BLE or other wireless communication protocols. After generation, the system periodically sends this broadcast packet through the Bluetooth Low Energy (BLE) interface. This method greatly improves the reliability of the alarm system because even when the mobile phone is not present or Bluetooth is disconnected, the temperature probe can independently complete the alarm task, thereby reducing the system's dependence on network connectivity. This method ensures that when an anomaly occurs in the monitored environment, an alarm is quickly sent to surrounding support devices or alarm systems, thereby improving security capabilities.

[0028] In one embodiment, step S4, which involves extracting the target features of the first target sequence using a preset feature extraction method, includes: S401: Transform the first target sequence using a preset weight matrix to obtain the second target sequence; S402: The second target sequence is encoded by a preset first encoder to obtain a first data feature, and the second target sequence is encoded by a preset second encoder to obtain a second data feature; wherein, the first encoder is used to extract the first data feature of the column data of the second target sequence, and the second encoder is used to extract the second data feature of the horizontal data of the second target sequence. S403: According to a preset data fusion method, the first data feature and the second data feature are fused to obtain the target feature.

[0029] As described in step S401 above, the first target data sequence is transformed using a preset weight matrix to obtain the second target data sequence. The first target data sequence is used as input and subjected to matrix operations with the preset weight matrix, specifically a linear transformation. This aims to emphasize certain features in the dataset through weighting, while suppressing unimportant noise or redundant information. The preset weight matrix is ​​set based on previous data analysis and modeling results, mapping the characteristics and importance of temperature data. By reducing the weight of certain features and increasing the weight of others, the key factors of temperature changes can be better grasped, enhancing the sensitivity to the main dynamics of the target data sequence and providing more accurate and effective basic data for subsequent feature extraction and information analysis.

[0030] As described in step S402 above, the second target data sequence is encoded using a preset first encoder to obtain a first data feature, and the second target data sequence is encoded using a preset second encoder to obtain a second data feature. The second target data sequence is processed by the two preset encoders to extract different types of data features. First, the first encoder focuses on extracting column data features from the second target data sequence, analyzing changes in temperature data at specific locations. The second encoder focuses on extracting horizontal data features from the second target data sequence, primarily targeting time-varying sequences. By processing the same batch of temperature data using different encoders, the system can capture richer information from multiple dimensions, thus laying a solid foundation for subsequent feature fusion. This ensures the diversity of extracted features and increases the complexity of subsequent analysis methods, facilitating the full exploration of potential indicators of temperature anomalies. The first and second data features will provide multiple effective feature dimensions for subsequent fusion processing.

[0031] As described in step S403 above, the first data feature and the second data feature are fused according to a preset data fusion method to obtain the target feature. That is, the first data feature and the second data feature are integrated using a certain fusion method to generate the target feature representing the integrated features. The data fusion method can take various forms, such as weighted average, principal component analysis, or feature fusion techniques in deep learning. Through this step, the system can effectively combine features extracted from different dimensions, thereby avoiding misjudgments or omissions caused by the inadequacy of a single feature. The fused target feature not only better reflects the overall trend of temperature changes numerically, but also improves the accuracy of the system in identifying temperature anomalies during the subsequent training and inference stages of the model. Finally, the target feature obtained after fusion processing will provide the most reliable basis for the system's alarm mechanism, ensuring an accurate and rapid response in abnormal temperature conditions.

[0032] In one embodiment, step S403, which involves fusing the first data feature and the second data feature according to a preset data fusion method to obtain the target feature, includes: S4031: The first data feature and the second data feature are concatenated according to a preset feature dimension to form a fusion vector; S4032: The fusion vector is linearly mapped using a preset linear transformation matrix to obtain the mapping features; S4033: The mapped features are processed nonlinearly according to a preset nonlinear activation function to obtain the target features.

[0033] As described in step S4031 above, the first data feature and the second data feature are concatenated according to a preset feature dimension to form a fusion vector. The concatenation operation is performed in a specific order to preserve the potential relationships between features. The concatenated fusion vector will contain all information components from the first and second data features, resulting in a richer data structure. This ensures that subsequent steps can extract useful information from various features. Furthermore, to achieve the concatenation, the system must carefully consider the feature types and ranges of the second and first data features during the implementation process to ensure that the concatenated result can fully express the actual temperature changes. After this step, the obtained fusion vector will lay a solid foundation for subsequent linear transformations.

[0034] As described in step S4032 above, the fused vector is linearly mapped using a preset linear transformation matrix to obtain mapped features. Matrix operations are used to map the fused vector to a new feature space, aiming to optimize feature representation and reduce data dimensionality through linear transformation. Using a linear transformation matrix helps reduce redundant information in the concatenated data and highlights important features through weighting. Because the mapped features generated after linear transformation are more suitable for processing by backend models or algorithms, the information contained in the mapped features can effectively convey the basic properties of temperature data, thereby improving the understanding of temperature change patterns. When implementing linear mapping, the system must ensure that the accuracy and characteristics of matrix multiplication remain consistent to avoid introducing errors or deviations.

[0035] As described in step S4033 above, the mapped features are nonlinearly processed using a preset nonlinear activation function to obtain the target features. Nonlinear processing of the mapped features using a preset nonlinear activation function (such as ReLU, Sigmoid, or Tanh) helps introduce nonlinear elements, enabling the final generated target features to better reflect the complexity and changing trends of temperature data. The role of nonlinear activation is to introduce nonlinear transformation capabilities into the features, allowing subsequent models to capture more complex data patterns and interactions between features. For example, the ReLU function can effectively suppress negative values, while the Sigmoid function can limit the output to a specified range. After this activation process, the final target features will have higher expressive power, thus better serving the subsequent alarm decision-making process. This processing step, to a certain extent, ensures the robustness and accuracy of the system, helping it to make more efficient and accurate responses when dealing with abnormal temperature conditions.

[0036] In one embodiment, step S401, which transforms the first target sequence using a preset weight matrix to obtain the second target sequence, includes: S4011: Using the first target sequence as the input vector, perform matrix multiplication on the input vector and the preset weight matrix to obtain the matrix operation result; S4012: The matrix operation results are numerically summed according to a preset bias vector to form an intermediate sequence; S4013: Process the intermediate sequence element by element according to a preset linear or nonlinear activation rule to obtain the second target sequence.

[0037] As described in step S4011 above, the first target sequence is used as the input vector, and the input vector is multiplied by a preset weight matrix to obtain the matrix operation result. In this sub-step, the first target sequence is first used as the input vector, denoted as X, and multiplied by a preset weight matrix W. Mathematically, this can be represented as Y = WX, where Y is the calculation result, representing the benchmark for the next step of processing. The weight matrix is ​​usually set during the system's training phase to assign appropriate weights to input data of different dimensions, thereby emphasizing important features and suppressing irrelevant features. Through matrix operations, the representation of the first target sequence in the new feature space can be extracted and calculated, forming a new data structure. The generated matrix operation result will provide the necessary foundation for subsequent bias addition and activation processing.

[0038] As described in step S4012 above, the matrix operation result is numerically summed according to a preset bias vector to form an intermediate sequence. A bias operation is then performed on the matrix operation result. Specifically, a preset bias vector b is added to the calculated matrix operation result Y to form a new intermediate sequence Z, mathematically represented as Z = Y + b. The introduction of the bias vector aims to provide an additional adjustment factor for each feature, enabling a greater capture of the potential variation patterns of temperature data in subsequent processing. By adding the bias, it can better adapt to temperature fluctuations and changes, avoiding reliance solely on the results obtained from weight transformations. This helps stabilize and optimize the network's learning process and enhances the model's flexibility. The resulting intermediate sequence Z lays the foundation for subsequent nonlinear activation processing, making the target feature generation process more adaptive and effective.

[0039] As described in step S4013 above, the intermediate sequence is processed element by element according to a preset linear or nonlinear activation rule to obtain the second target sequence. A preset activation function is applied to the formed intermediate sequence Z, and each element is processed to generate the required second target sequence. The activation function can be linear, such as a simple identity function; or it can be nonlinear, such as the common ReLU (Rectified Linear Unit), Sigmoid, or Tanh function. The choice of a suitable activation function depends on the characteristics of the data and the specific application requirements. Nonlinear activation can introduce more nonlinear decision boundaries, thereby enhancing the model's representational ability and allowing the system to capture more complex feature relationships. After the intermediate sequence undergoes activation processing, the final second target sequence will be the result of the model after transformation and processing, possessing a more accurate expression of temperature features. This step is a crucial stage in feature extraction, providing important data for the system's subsequent alarm decisions and temperature anomaly monitoring.

[0040] In one embodiment, step S6, which involves generating a broadcast packet and broadcasting an alarm based on the target feature if the target feature reaches an alarm threshold, includes: S601: Determine the temperature numerical parameters and temperature status code according to the target features and a preset mapping relationship; S602: Generate a broadcast packet based on the temperature numerical parameters and the temperature status code according to the preset broadcast protocol format; S603: The broadcast packet is continuously transmitted via the Bluetooth Low Energy Broadcast Channel at a preset broadcast interval.

[0041] As described in step S601 above, the temperature numerical parameters and temperature status codes are determined according to the target features based on a preset mapping relationship. Based on the previously extracted target feature data, the current temperature numerical parameters and corresponding temperature status codes are established through a preset mapping relationship. This mapping relationship typically relies on prior experience, data modeling, or historical data analysis, enabling the temperature features to be effectively transformed into usable parameters. For example, the system may determine the actual temperature value based on the specific numerical value of the target feature (e.g., taking a certain multiple of the real-time temperature). Simultaneously, status codes are defined according to the temperature range to indicate whether the temperature is in a "normal," "over-temperature," or "low-temperature fault" state. This concretizes the abstract temperature features and gives them practical meaning, providing necessary data support for subsequent broadcast packet generation. Accurate parameter and status code determination is crucial for the accuracy of the alarm, ensuring that the receiving device can reasonably interpret the transmitted information.

[0042] As described in step S602 above, a broadcast packet is generated based on the temperature numerical parameters and the temperature status code according to a preset broadcast protocol format. The obtained temperature numerical parameters and temperature status code are integrated into a broadcast packet according to the pre-defined broadcast protocol format. Here, the broadcast protocol format refers to a set of structured rules that define the order of fields, data types, and the number of bytes occupied by each field in the broadcast packet. The generated broadcast packet will include multiple fields, such as a preamble, device address, data length, flag bits, temperature status, and temperature value. Through this explicit structure, the broadcast packet can be correctly parsed by any receiving device conforming to the protocol. Once the generated broadcast packet contains complete and reasonable information, the system can transmit the packet to the receiving device via Bluetooth or other wireless communication protocols.

[0043] As described in step S603 above, the broadcast packets are continuously transmitted via the Bluetooth Low Energy (BLE) broadcast channel at preset broadcast intervals. The generated broadcast packets are continuously transmitted via the Bluetooth Low Energy (BLE) broadcast channel. Bluetooth Low Energy is a high-efficiency communication protocol specifically designed for IoT devices and sensors, characterized by low power consumption and long battery life. During this process, the system continuously transmits broadcast packets according to the set broadcast interval (e.g., once every 100 milliseconds) to ensure that surrounding receiving devices can promptly obtain alarm information about abnormal temperatures. This periodic transmission method provides redundancy, effectively improving the reliability of message transmission and ensuring successful capture of broadcast information even in complex signal environments.

[0044] In one embodiment, before step S2 of extracting target temperature data from multiple preset time points closest to the current time point from the acquired temperature data, the method further includes: S101: Determine whether any of the target temperature values ​​is greater than a preset temperature value; S102: If any of the target temperature values ​​is greater than a preset temperature value, the main control unit connected to the temperature probe is activated to execute the step of extracting the target temperature data of multiple preset time points closest to the current time point from the acquired temperature data.

[0045] As described in step S101 above, it is determined whether any of the target temperature values ​​exceeds a preset temperature value. The previously collected target temperature values ​​are evaluated to determine if they exceed a preset temperature threshold. This preset temperature threshold is typically set based on safety standards for a specific environment, equipment operating requirements, or system design goals. When the system acquires multiple sets of target temperature data, it iterates through these data one by one to find if any value exceeds the preset value. This allows for timely detection of potential risks when abnormal temperature changes occur, preparing for subsequent event response. Abnormal temperature fluctuations can be quickly identified, enabling the system to react rapidly. Temperature anomalies may lead to equipment malfunctions or safety hazards; therefore, early detection is crucial for ensuring system stability and safety.

[0046] As described in step S102 above, if any of the target temperature values ​​exceeds a preset temperature value, the main control unit connected to the temperature probe is activated. If the temperature data analysis confirms that a value exceeds the preset temperature value, a wake-up operation is immediately performed, restoring the main control unit from a low-power or sleep state to a working state. Since the system's functionality may be limited when the main control unit is in a low-power state, making it unable to effectively analyze temperature data or trigger alarms, the wake-up process typically relies on sensors or hardware triggers to ensure rapid state switching when necessary. After activating, the main control unit can immediately begin data processing tasks, such as extracting the target temperature data closest to the current time. This design allows the system to conserve power while maintaining normal operation, but it can react quickly once a temperature anomaly is detected. In this way, the system not only improves the timeliness and accuracy of alarms but also extends the equipment's lifespan and minimizes maintenance costs.

[0047] In one embodiment, after step S6, which involves generating a broadcast packet and broadcasting an alarm based on the target feature if the target feature reaches an alarm threshold, the method further includes: S701: Continuously collects gradient temperature data at multiple time points; S702: Detect whether all the gradient temperature data are lower than the preset temperature value and continuously meet the preset judgment period; S703: If all the gradient temperature data are lower than the preset temperature value and the preset judgment period is met continuously, then stop broadcasting the alarm and put the main control unit back into deep sleep state.

[0048] As described in step S701 above, gradient temperature data at multiple time points is continuously collected. This involves continuously collecting temperature gradient data using multiple temperature probes to monitor temperature changes in the current environment. Unlike the target temperature data, gradient temperature data emphasizes the rate of temperature change and fluctuation trends, reflecting the dynamic characteristics of environmental temperature changes. This acquisition method typically involves continuous temperature monitoring within specific time intervals (e.g., per second or per minute) to obtain a continuous sequence of temperature change data. The key to this step is ensuring the continuity and real-time nature of the data, enabling the system to respond quickly to changes in the external environment. Continuous acquisition of gradient temperature data provides a foundation for subsequent anomaly detection and decision-making. Unless there are significant temperature fluctuations, the system can effectively maintain a low-power state. This step can also help achieve early warning of potential risks by identifying rising or falling temperature trends, laying a data foundation for subsequent alarm and response measures.

[0049] As described in step S702 above, it is detected whether all the gradient temperature data are below the preset temperature value and continuously meet the preset judgment period. The continuously collected gradient temperature data is detected to verify whether all data are below the preset temperature threshold and remain in this state for a period of time. This detection process is crucial because it determines whether the device state has returned to normal, preventing continued operation in overheated or unsafe conditions. In implementation, the system first checks each collected gradient temperature data to ensure that all data are below the set safe temperature value. This logic can be implemented through a set loop or conditional branch. If all gradient temperature data are indeed below the set threshold, the system also records the duration of this state to confirm whether the preset judgment period is met. For example, the system may be set to require that temperature data within 60 seconds be below the threshold to confirm temperature recovery. Through this judgment, the system can confirm temperature state changes within an effective time window, avoiding false alarms or missed alarms. After reaching the normal time, subsequent state adjustments can be made to ensure that the system can effectively return to low-power mode when the temperature is safe, improving the overall efficiency of the device.

[0050] As described in step S703 above, if all the gradient temperature data are lower than the preset temperature value and continuously meet the preset judgment period, the broadcast alarm will stop, and the main control unit will re-enter deep sleep mode. If all the gradient temperature data are lower than the preset temperature value, and this state has continuously met the set time period, the system will execute a stop alarm command to ensure that the alarm function can only be stopped under a truly safe condition, thereby avoiding unnecessary interference or waste of resources. Once the condition is confirmed to be met, the system will send a stop signal to the previously generated broadcast mechanism to terminate the alarm. This process is not only crucial for the normal operation of the device, but also helps to improve the user's operating experience and avoid continuous alarms affecting normal life or work environment. At the same time, the system needs to return the main control unit to deep sleep mode again to reduce energy consumption. Once the safety standard is reached, the system no longer needs to continuously perform high-frequency data acquisition and analysis, thereby saving power and extending the service life of the device.

[0051] In one embodiment, after step S5 of determining whether the target feature has reached the alarm threshold, the method further includes: S611: If the target feature reaches the alarm threshold, then the lighting control parameters are determined according to the target feature based on a preset mapping relationship. The lighting control parameters include the lighting effect mode and / or the light color value. S612: Based on the lighting control parameters, drive the preset lighting device to emit the corresponding lighting effect and / or color.

[0052] As described in step S611 above, in existing temperature anomaly alarm schemes, even if the main control unit has determined that the target characteristic has reached the alarm threshold, its output is often limited to Bluetooth broadcast packets or remote notifications. For on-site personnel, if their mobile phone is not nearby, the application is not open, the Bluetooth connection is disconnected, or they are in a noisy environment where it is inconvenient to view the terminal, it is still difficult to intuitively perceive the temperature anomaly in a timely manner. At the same time, traditional light prompts usually only provide simple "on / off" or fixed color alarms, and cannot differentiate based on the target characteristics such as the degree of temperature anomaly, temperature change trend, and rate of change, resulting in insufficient light feedback information. Furthermore, in multi-temperature probe scenarios, if the status of each probe cannot be mapped to distinguishable light parameters, on-site personnel cannot quickly determine which probe or which area is experiencing an anomaly.

[0053] The target feature can be obtained after weight matrix transformation, dual encoder encoding, and fusion processing of the first target sequence, or it can be a derived value such as temperature value, temperature change rate, temperature trend, alarm level, or abnormal duration corresponding to the target feature. Then, the lighting control parameters are determined based on a preset mapping relationship. The preset mapping relationship can be a lookup table, piecewise function, linear mapping, nonlinear mapping, fuzzy rule, empirical formula, or machine learning mapping model. For example, multiple temperature abnormality intervals can be preset: when the target feature is in the first abnormality interval, it is mapped to a solid yellow light; when the target feature is in the second abnormality interval, it is mapped to a slow orange flash; when the target feature is in the third abnormality interval, it is mapped to a fast red flash; when the target feature indicates a rapid temperature rise, it is mapped to a red breathing or gradient effect. The lighting control parameters include the lighting effect mode and / or the light color value. The lighting effect mode can include solid light, flashing, breathing, gradient, double flash, fast flash, slow flash, SOS rhythm, etc.; the light color value can include RGB value, HSV value, color temperature value, brightness value, saturation value, etc. For multi-probe systems, separate lighting control parameters can be determined for each temperature probe, allowing abnormal probes to have different light colors or effects than normal probes. In the example above, the first abnormal zone can be understood as the "warning zone," indicating that the temperature has deviated from the normal range but has not yet reached a serious level; the second abnormal zone can be understood as the "over-temperature zone," indicating that the temperature abnormality is already quite obvious; and the third abnormal zone can be understood as the "high-risk zone," indicating that the temperature abnormality is severe or the temperature rise is extremely rapid, requiring immediate action. Alarm thresholds can include a first threshold, a second threshold, and a third threshold. For example, the first threshold can be set to 45℃ or a normalized target characteristic of 0.60, the second threshold can be set to 55℃ or a normalized target characteristic of 0.75, and the third threshold can be set to 65℃ or a normalized target characteristic of 0.90.

[0054] In one embodiment, the preset mapping relationship includes a first abnormal interval, a second abnormal interval, and a third abnormal interval. The first abnormal interval is, for example, a temperature value of 45°C to 55°C and a temperature rise rate of 0.5°C / min to 1.0°C / min, or a normalized target feature of 0.60 to 0.75, corresponding to a solid yellow light or slow flashing. The second abnormal interval is, for example, a temperature value of 55°C to 65°C and a temperature rise rate of 1.0°C / min to 2.0°C / min, or a normalized target feature of 0.75 to 0.90, corresponding to an orange breathing or slow flashing. The third abnormal interval is, for example, a temperature value greater than or equal to 65°C and a temperature rise rate greater than 2.0°C / min, or a normalized target feature greater than or equal to 0.90, corresponding to a red fast flashing, breathing, or gradual change. The above values ​​are merely examples, and this application is not limited to them; specific thresholds can be calibrated according to the application scenario.

[0055] Based on the aforementioned lighting control parameters, preset lighting devices are driven to emit corresponding lighting effects and / or colors. This scheme converts the lighting control parameters into actual light output through a main control unit or a dedicated lighting driver circuit. The lighting devices can be one or more LEDs, RGB LEDs, light strips, indicator lights, display backlight modules, etc. Driving methods can include GPIO switch control, PWM dimming, constant current driving, SPI / I2C driving, dedicated LED driver chip control, etc. For lighting effect modes, constant light, blinking, breathing, and gradient effects can be achieved by adjusting the PWM duty cycle, frequency, and timing mode; for color values, different colors can be achieved by adjusting the duty cycle of the RGB three channels or writing to a color register. In multi-probe scenarios, each lighting device can be synchronously refreshed according to its corresponding lighting control parameters, thereby displaying the temperature status of each temperature probe. If the lighting device is a monochrome LED, its on / off, blinking frequency, and duty cycle can be controlled according to the lighting effect mode; if the lighting device is an RGB LED, the brightness of the red, green, and blue channels can be controlled separately according to the color value, thereby synthesizing different colors such as red, yellow, green, orange, blue, and purple. For breathing and gradual effects, the main control unit can gradually change the PWM duty cycle according to a preset time sequence to smoothly increase or decrease the brightness. For flashing effects, it can alternately turn the lights on and off according to a preset frequency and duty cycle. For synchronous lighting display of multiple probe statuses, the main control unit can refresh multiple lighting devices simultaneously or according to a unified clock, so that the light corresponding to each probe displays the status corresponding to its target characteristic. For example, an abnormal probe corresponds to a fast red flash, a normal probe corresponds to a solid green light, and a warning probe corresponds to a slow yellow flash. During the driving process, intermittent driving, low duty cycle driving, and zone driving can also be used to reduce power consumption. When the alarm is cleared, the main control unit can determine the temperature recovery based on subsequent gradient temperature data and stop the light feedback, causing the lighting device to turn off or return to the normal indication state, while simultaneously putting the main control unit back into deep sleep mode.

[0056] In one embodiment, the light effect mode includes one or more of constant light, flashing, breathing, and gradient, and / or the light color includes one or more of red, yellow, and green.

[0057] In this embodiment, the lighting control parameters are further defined: the lighting effect mode includes one or more of constant light, flashing, breathing, and gradual change, and / or the light color includes one or more of red, yellow, and green. This definition does not replace broadcast alarms with lighting feedback, but rather provides configurable, combinable, and distinguishable lighting effect and color expression methods for the local lighting feedback dimension, in addition to broadcast alarms. "And / or" indicates that only lighting effect modes, only light colors, or a combination of lighting effect modes and light colors can be used. Constant light means that the lighting device continuously emits light at a basically constant brightness during the alarm period, without periodic extinguishing or significant brightness fluctuations. Constant light can be used as the default lighting effect for the first or second abnormal interval, or it can be combined with different light colors, for example, constant yellow for warning and constant red for high risk. The constant light mode can be used for single-probe alarms, as well as for normal status indication or continuous abnormal status indication in multi-probe status synchronous lighting displays. It can adapt to different hardware configurations and cost requirements, improving implementation flexibility. Since only lighting effect or only light color can be used, the system's dependence on the lighting device is reduced. Because light effects and colors can be combined, richer alarm semantics can be expressed. Since each sensor can be configured separately in multi-sensor scenarios, synchronized lighting display of multiple sensor statuses can be achieved.

[0058] Reference Figure 3 The present invention also provides a low-power temperature anomaly alarm device based on wireless communication, the device comprising: Temperature data acquisition module 902 is used to continuously acquire temperature data at multiple time points through multiple temperature probes; The target temperature data extraction module 904 is used to extract target temperature data from multiple preset time points closest to the current time point from the acquired temperature data; The target temperature data arrangement module 906 is used to arrange the target temperature data in chronological order to obtain a first target data sequence; The target feature extraction module 908 is used to extract the target features of the first target sequence using a preset feature extraction method. The judgment module 910 is used to determine whether the target feature has reached the alarm threshold; The broadcast packet generation module 912 is used to generate a broadcast packet based on the target feature and broadcast an alarm if the target feature reaches the alarm threshold.

[0059] In one embodiment, the target feature extraction module 908 includes: The second target sequence submodule is used to transform the first target sequence using a preset weight matrix to obtain the second target sequence; The encoding submodule is used to encode the second target sequence using a preset first encoder to obtain a first data feature, and to encode the second target sequence using a preset second encoder to obtain a second data feature; wherein, the first encoder is used to extract the first data feature of the column data of the second target sequence, and the second encoder is used to extract the second data feature of the horizontal data of the second target sequence. The fusion submodule is used to fuse the first data feature and the second data feature according to a preset data fusion method to obtain the target feature.

[0060] In one embodiment, the fusion submodule includes: A fusion vector forming unit is used to concatenate the first data feature and the second data feature according to a preset feature dimension to form a fusion vector; The mapping feature acquisition unit is used to perform linear mapping on the fusion vector through a preset linear transformation matrix to obtain mapping features; The target feature acquisition unit is used to perform nonlinear processing on the mapped features according to a preset nonlinear activation function to obtain the target features.

[0061] In one embodiment, the second target sequence submodule includes: The matrix operation result acquisition unit is used to take the first target sequence as the input vector, perform matrix multiplication on the input vector and a preset weight matrix to obtain the matrix operation result; The intermediate sequence forming unit is used to perform numerical summation on the matrix operation results according to a preset bias vector to form an intermediate sequence; The second target sequence acquisition unit is used to process the intermediate sequence element by element according to a preset linear or nonlinear activation rule to obtain the second target sequence.

[0062] In one embodiment, the broadcast packet generation module 912 includes: The determination submodule is used to determine the temperature numerical parameters and temperature status code according to the target features and a preset mapping relationship. The broadcast packet generation submodule is used to generate broadcast packets based on the temperature numerical parameters and the temperature status code according to a preset broadcast protocol format. The transmitting submodule is used to continuously transmit the broadcast packets through the Bluetooth Low Energy Broadcast channel at preset broadcast intervals.

[0063] In one embodiment, the low-power temperature anomaly alarm device based on wireless communication further includes: The temperature value determination module is used to determine whether there is a temperature value among the target temperature values ​​that is greater than a preset temperature value. The wake-up module is used to wake up the main control unit connected to the temperature probe if there is a temperature value greater than a preset temperature value among the target temperature values, so as to execute the step of extracting the target temperature data of multiple preset time points closest to the current time point from the acquired temperature data.

[0064] In one embodiment, the low-power temperature anomaly alarm device based on wireless communication further includes: The gradient temperature data acquisition module is used to continuously acquire gradient temperature data at multiple time points; The detection module is used to detect whether all the gradient temperature data are lower than the preset temperature value and continuously meet the preset judgment period. The broadcast alarm stop module is used to stop broadcasting alarms and cause the main control unit to re-enter deep sleep state if all the gradient temperature data are lower than the preset temperature value and the preset judgment period is met continuously.

[0065] In one embodiment, the low-power temperature anomaly alarm device based on wireless communication further includes: A lighting control parameter determination module is used to determine lighting control parameters according to a preset mapping relationship based on the target feature if the target feature reaches an alarm threshold. The lighting control parameters include a light effect mode and / or a light color value. The lighting device driving module is used to drive a preset lighting device to emit corresponding light effects and / or light colors according to the lighting control parameters.

[0066] In one embodiment, the light effect mode includes one or more of constant light, flashing, breathing, and gradient, and / or the light color includes one or more of red, yellow, and green.

[0067] Figure 4 An internal structural diagram of an electronic device in one embodiment is shown. This electronic device can specifically be a terminal or a server, and more specifically, a computer device. Figure 4As shown, the electronic device includes a processor, a memory, and a network interface connected via a system bus. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system and may also store a computer program. When executed by the processor, this computer program enables the processor to implement a low-power temperature anomaly alarm method based on wireless communication. The internal memory may also store a computer program, which, when executed by the processor, enables the processor to implement a low-power temperature anomaly alarm method based on wireless communication. Those skilled in the art will understand that… Figure 4 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the electronic device to which the present application is applied. The specific electronic device may include more or fewer components than shown in the figure, or combine certain components, or have different component arrangements.

[0068] In one embodiment, an electronic device is provided, including a memory and a processor, the memory storing a computer program that, when executed by the processor, causes the processor to perform the following steps: Temperature data was continuously collected at multiple time points using multiple temperature probes. Extract target temperature data from multiple preset time points closest to the current time point from the acquired temperature data; The target temperature data are arranged in chronological order to obtain the first target data sequence; The target features of the first target sequence are extracted using a preset feature extraction method; Determine whether the target feature has reached the alarm threshold; If the target feature reaches the alarm threshold, a broadcast packet is generated based on the target feature and an alarm is broadcast.

[0069] By collecting temperature data at multiple time points and generating target features based on a preset feature extraction method, a comprehensive judgment of key characteristics such as temperature trends and rate of change is achieved. Compared with the traditional single-point threshold method, the accuracy of alarms can be significantly improved. In addition, by generating a broadcast packet and executing Bluetooth Low Energy broadcast when the target feature reaches the alarm threshold, the alarm can be completed without maintaining a Bluetooth connection, thus completely solving the high power consumption problem caused by continuous connection in the traditional solution. It also enables the device to quickly issue temperature abnormality warnings even when the mobile phone is not present or the connection is lost.

[0070] In one embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, causes the processor to perform the following steps: Temperature data was continuously collected at multiple time points using multiple temperature probes. Extract target temperature data from multiple preset time points closest to the current time point from the acquired temperature data; The target temperature data are arranged in chronological order to obtain the first target data sequence; The target features of the first target sequence are extracted using a preset feature extraction method; Determine whether the target feature has reached the alarm threshold; If the target feature reaches the alarm threshold, a broadcast packet is generated based on the target feature and an alarm is broadcast.

[0071] By collecting temperature data at multiple time points and generating target features based on a preset feature extraction method, a comprehensive judgment of key characteristics such as temperature trends and rate of change is achieved. Compared with the traditional single-point threshold method, the accuracy of alarms can be significantly improved. In addition, by generating a broadcast packet and executing Bluetooth Low Energy broadcast when the target feature reaches the alarm threshold, the alarm can be completed without maintaining a Bluetooth connection, thus completely solving the high power consumption problem caused by continuous connection in the traditional solution. It also enables the device to quickly issue temperature abnormality warnings even when the mobile phone is not present or the connection is lost.

[0072] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments described above. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and RAMbus dynamic RAM (RDRAM), etc.

[0073] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0074] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A low-power-consumption temperature abnormality alarm method based on wireless communication, characterized in that, The method includes: Temperature data was continuously collected at multiple time points using multiple temperature probes. Extract target temperature data from multiple preset time points closest to the current time point from the acquired temperature data; The target temperature data are arranged in chronological order to obtain the first target data sequence; The target features of the first target sequence are extracted using a preset feature extraction method; Determine whether the target feature has reached the alarm threshold; If the target feature reaches the alarm threshold, a broadcast packet is generated based on the target feature and an alarm is broadcast.

2. The method for realizing low-power-consumption temperature abnormality alarm based on wireless communication according to claim 1, characterized in that, The step of extracting the target features of the first target sequence using a preset feature extraction method includes: The first target sequence is transformed using a preset weight matrix to obtain the second target sequence; The second target sequence is encoded by a preset first encoder to obtain a first data feature, and the second target sequence is encoded by a preset second encoder to obtain a second data feature; wherein, the first encoder is used to extract the first data feature of the column data of the second target sequence, and the second encoder is used to extract the second data feature of the horizontal data of the second target sequence. According to a preset data fusion method, the first data feature and the second data feature are fused to obtain the target feature.

3. The method of claim 2, wherein the wireless communication is Bluetooth Low Energy (BLE). The step of fusing the first data feature and the second data feature according to a preset data fusion method to obtain the target feature includes: The first data feature and the second data feature are concatenated according to a preset feature dimension to form a fusion vector; The fused vector is linearly mapped using a preset linear transformation matrix to obtain the mapping features; The mapped features are processed nonlinearly according to a preset nonlinear activation function to obtain the target features.

4. The method for realizing low-power-consumption temperature abnormality alarm based on wireless communication according to claim 2, characterized in that, The step of transforming the first target sequence using a preset weight matrix to obtain the second target sequence includes: Using the first target sequence as the input vector, the input vector is multiplied by a preset weight matrix to obtain the matrix operation result. The matrix operation results are numerically summed according to a preset bias vector to form an intermediate sequence; The intermediate sequence is processed element by element according to a preset linear or nonlinear activation rule to obtain the second target sequence.

5. The method for realizing low-power-consumption temperature abnormality alarm based on wireless communication according to claim 1, characterized in that, The step of generating a broadcast packet and broadcasting an alarm based on the target feature if the target feature reaches the alarm threshold includes: Based on the target features, the temperature numerical parameters and temperature status code are determined according to a preset mapping relationship; According to the preset broadcast protocol format, a broadcast packet is generated based on the temperature numerical parameters and the temperature status code; The broadcast packets are continuously transmitted via the Bluetooth Low Energy Broadcast channel at preset broadcast intervals.

6. The method for realizing low-power-consumption temperature abnormality alarm based on wireless communication according to claim 1, characterized in that, Before the step of extracting target temperature data from multiple preset time points closest to the current time point from the acquired temperature data, the method further includes: Determine whether any of the target temperature values ​​is greater than a preset temperature value; If any of the target temperature values ​​is greater than a preset temperature value, the main control unit connected to the temperature probe is activated to execute the step of extracting the target temperature data of multiple preset time points closest to the current time point from the acquired temperature data.

7. The method of claim 6, wherein the wireless communication is Bluetooth Low Energy (BLE). After the step of generating a broadcast packet and broadcasting an alarm based on the target feature if the target feature reaches the alarm threshold, the method further includes: Continuously collect gradient temperature data at multiple time points; Detect whether all the gradient temperature data are lower than the preset temperature value and continuously meet the preset judgment period; If all the gradient temperature data are lower than the preset temperature value and the preset judgment period is met continuously, the broadcast alarm will be stopped and the main control unit will re-enter deep sleep mode.

8. The method for realizing low-power-consumption temperature abnormality alarm based on wireless communication according to claim 1, characterized in that, After the step of determining whether the target feature has reached the alarm threshold, the method further includes: If the target feature reaches the alarm threshold, then the lighting control parameters are determined according to the target feature based on a preset mapping relationship. The lighting control parameters include the lighting effect mode and / or the light color value. Based on the lighting control parameters, the preset lighting device is driven to emit corresponding lighting effects and / or colors.

9. The method of claim 8, wherein the wireless communication is Bluetooth Low Energy (BLE). The light effect mode includes one or more of constant light, flashing, breathing, and gradient, and / or the light color includes one or more of red, yellow, and green.

10. A low power consumption temperature abnormality alarm device based on wireless communication, characterized in that, The device includes: The temperature data acquisition module is used to continuously acquire temperature data at multiple time points through multiple temperature probes. The target temperature data extraction module is used to extract target temperature data from multiple preset time points closest to the current time point from the acquired temperature data. The target temperature data arranging module is used to arrange the target temperature data in chronological order to obtain a first target data sequence; The target feature extraction module is used to extract the target features of the first target sequence using a preset feature extraction method; The judgment module is used to determine whether the target feature has reached the alarm threshold; The broadcast packet generation module is used to generate a broadcast packet based on the target feature and broadcast an alarm if the target feature reaches the alarm threshold.

11. A computer readable storage medium, characterized in that, The device stores a computer program, which, when executed by a processor, causes the processor to perform the steps of the method for implementing a low-power temperature anomaly alarm based on wireless communication as described in any one of claims 1 to 9.

12. An electronic device, comprising: The device includes a memory and a processor. The memory stores a computer program, which, when executed by the processor, causes the processor to perform the steps of the method for implementing a low-power temperature anomaly alarm based on wireless communication as described in any one of claims 1 to 9.