Context-based Internet of Things data pushing method

Abstract

The invention discloses a context-based Internet of Things data pushing method, which relates to the technical field of data pushing, and comprises the following steps: collecting original data, inputting preprocessed data into a context fusion engine, fusing the preprocessed data into context data, identifying agricultural context events based on a context data application rule engine and a machine learning reasoning model, and outputting pushing event data. According to the method, the soil humidity data, the air temperature and humidity data, the illumination intensity data and the meteorological platform data are fused through the unified agricultural situation model, unified processing and situation reasoning of the multi-source data are achieved, and the accuracy and practicability of the pushed information are improved. The agricultural situation event is recognized by combining the rule engine and the machine learning reasoning model, and the pushing frequency and content are adaptively adjusted according to the pushing event data, so that the communication and calculation burden is reduced, the user experience is optimized, and a good application effect is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of data push technology, and in particular to a context-based Internet of Things data push method. Background Art

[0002] With the rapid development of IoT technology, a large number of soil moisture sensors, air temperature and humidity sensors, light intensity sensors, and terminal devices that access meteorological platform data have been deployed in agricultural scenarios. Existing data push methods typically use threshold judgments based on single sensor data to trigger push notifications, either timed or triggered by events. These methods suffer from the following deficiencies: First, data from different sources lacks unified fusion processing, resulting in incomplete environmental perception; second, push decisions lack contextual understanding, making it prone to false positives or missed positives due to individual sensor anomalies or errors; and third, push content and frequency are not dynamically adjusted based on the actual context, which can easily lead to user information overload or delayed responses. To address these issues, a new method based on multi-source data fusion, context-aware analysis, and intelligent push notification is urgently needed to improve the accuracy, timeliness, and system stability of IoT data push notifications.

[0003] At present, the Chinese patent application number CN202110559811.5 discloses a big data push method and system based on the Internet of Things. It includes the following steps: a mobile terminal establishes a communication connection with a first base station; the mobile terminal monitors the system information and synchronization signal sent by the second base station; in response to monitoring the synchronization signal sent by the second base station, the mobile terminal determines the link quality of the communication link between the mobile terminal and the second base station; if the link quality of the communication link between the mobile terminal and the second base station is higher than the link quality threshold, the mobile terminal sends a measurement report to the first base station; in response to receiving the measurement report, the first base station sends a handover request message to the second base station; in response to receiving the handover request message and in response to the second base station using the first random access mode, the second base station allocates a plurality of random access preamble codes specifically for the mobile terminal to the mobile terminal.

[0004] In related technologies, IoT data push relies on the judgment of a single sensor threshold, which makes it difficult to perform multi-source data fusion and situational reasoning mechanisms, resulting in low accuracy of pushed information, high false alarm rate, and poor user experience. Summary of the Invention

[0005] The technical problem solved by the present invention is that in the existing technology, IoT data push relies on a single sensor threshold judgment, which makes it difficult to perform multi-source data fusion and situational reasoning mechanism, resulting in low accuracy of pushed information, high false alarm rate and poor user experience.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] A context-based IoT data push method includes the following steps:

[0008] Step S1, collecting raw data;

[0009] Step S2: standardize, remove outliers, and perform preliminary aggregation on the raw data at the edge node, and output it as preprocessed data;

[0010] Step S3: input the pre-processed data into a context fusion engine, perform time stamp alignment and spatial location association based on a unified agricultural context model, and fuse them into context data;

[0011] Step S4: applying a rule engine and a machine learning inference model based on the contextual data to identify agricultural contextual events and output push event data;

[0012] Step S5: Generate push content based on the push event data, adjust the push frequency and content level according to the urgency and user preference, and send it to the user terminal.

[0013] Preferably, the step S1 includes the following sub-steps:

[0014] Step S101, obtaining current soil moisture raw data from the soil moisture sensor;

[0015] Step S102, obtaining current air temperature raw data and air humidity raw data from the air temperature and humidity sensor;

[0016] Step S103, obtaining current ambient light raw data from the light intensity sensor;

[0017] Step S104: obtaining current raw meteorological data and meteorological forecast data for the next 48 hours by calling the meteorological platform interface.

[0018] Preferably, step S2 includes the following sub-steps:

[0019] Step S201, performing format standardization processing on the soil moisture raw data, the air temperature raw data, the air humidity raw data and the ambient light intensity raw data respectively;

[0020] Step S202: detecting and removing abnormal values ​​in the soil moisture raw data, the air temperature raw data, the air humidity raw data, and the ambient light intensity raw data, wherein the abnormal values ​​include missing values, out-of-range values, and abnormal sensor fluctuation values;

[0021] Step S203 , locally aggregate the standardized soil moisture raw data, air temperature raw data, air humidity raw data, ambient light intensity raw data, meteorological raw data, and meteorological forecast data at preset time intervals to generate preprocessed data.

[0022] Preferably, step S3 includes the following sub-steps:

[0023] Step S301: input the pre-processed data into a context fusion engine. The context fusion engine includes a semantic annotation module, a time-space synchronization module, and a situational reasoning module. The logic of the semantic annotation module is:

[0024] The pre-processed data are semantically annotated based on a unified agricultural scenario model, and the raw data of soil moisture, air temperature, air humidity, ambient light intensity, meteorological data, and meteorological forecast data are respectively associated with the corresponding agricultural environmental factors;

[0025] Step S302: The logic of the time and space synchronization module is:

[0026] Based on timestamp synchronization rules, the soil moisture raw data, air temperature raw data, air humidity raw data, ambient light intensity raw data, meteorological raw data, and weather forecast data are aligned on the same timeline and spatially associated based on geographic location information to generate consistent contextual data.

[0027] Step S303: The logic of the contextual reasoning module is:

[0028] Preset fusion inference rules are applied to the consistent scenario data to convert the multi-source environmental factor data into unified agricultural scenario indicators, which include water sufficiency, climate suitability and light suitability, and output standardized scenario data.

[0029] Preferably, step S4 includes the following sub-steps:

[0030] Step S401, preliminarily screening the standardized context data based on a preset rule engine;

[0031] Step S402: Analyze the filtered standardized context data using a machine learning inference model trained based on historical context data, identify potential risks and action suggestion events, and output them as push event data.

[0032] Step S403: classify the pushed event data according to the preset event importance and mark them as high priority events, medium priority events or low priority events.

[0033] Preferably, step S5 includes the following sub-steps:

[0034] Step S501: Determine the level of detail and frequency of push content based on the importance of the push event data. For high-priority events, push detailed content and send it immediately. For medium- and low-priority events, push summary content and send it at a fixed time each day.

[0035] Step S502: Adjust the push strategy based on the user's preset receiving preferences, including filtering the data type of interest, limiting the push time period, and selecting an emergency notification method, and generate push content;

[0036] Step S503: Send the push content to the user terminal via a mobile application, SMS, or email, and attach a situation description and suggested measures derived from the analysis of the standardized situation data to the push content.

[0037] Preferably, the outlier removal in step S2 includes: comparing the mean value in a continuous time window with the single point data, identifying and deleting data points that exceed a set deviation range.

[0038] Preferably, the machine learning inference model in step S4 is based on a classification algorithm, the input features include soil moisture change trends, future meteorological changes, light intensity changes and historical crop growth response patterns, and the output is a situational event category.

[0039] Preferably, when the push content is generated in step S5, an environmental change chart and operation suggestions are attached based on the results of the situational reasoning analysis.

[0040] Preferably, the data processing logic package of the agricultural scenario model in step S3 is:

[0041] Define the corresponding relationship between soil moisture raw data, air temperature raw data, air humidity raw data, ambient light intensity raw data, meteorological raw data and meteorological forecast data and agricultural environmental factors, and map various pre-processed data to situational indicators, including water sufficiency, climate suitability and light suitability indicators;

[0042] Based on the corresponding relationship, the pre-processed data is aligned according to the unified timestamp synchronization rules, and spatial registration is performed based on the geographic location information to generate a context data set;

[0043] Based on the contextual data set, the fusion inference rules are applied to infer the comprehensive environmental status of the current farmland and output standardized contextual data.

[0044] The present invention has the following beneficial effects: This invention integrates soil moisture data, air temperature and humidity data, light intensity data, and meteorological platform data through a unified agricultural context model, enabling unified multi-source data processing and contextual reasoning, thereby improving the accuracy and practicality of pushed information. It also combines a rule engine with a machine learning inference model to identify agricultural contextual events and adaptively adjust push frequency and content based on push event data, reducing communication and computing burdens, optimizing the user experience, and demonstrating excellent application results. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 A flowchart of the steps of a context-based IoT data push method provided by one embodiment of the present invention. DETAILED DESCRIPTION

[0046] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings. It is obvious that the described embodiments are only part of the embodiments of the present invention, but not all of the embodiments.

[0047] Example, see Figure 1 , provides a context-based IoT data push method, including the following steps:

[0048] Step S1, collecting raw data;

[0049] Step S2: standardize, remove outliers, and perform preliminary aggregation on the raw data at the edge node, and output it as preprocessed data;

[0050] Step S3: input the pre-processed data into a context fusion engine, perform time stamp alignment and spatial location association based on a unified agricultural context model, and fuse them into context data;

[0051] Step S4: applying a rule engine and a machine learning inference model based on the contextual data to identify agricultural contextual events and output push event data;

[0052] Step S5: Generate push content based on the push event data, adjust the push frequency and content level according to the urgency and user preference, and send it to the user terminal.

[0053] Step S1 includes the following sub-steps:

[0054] Step S101: obtaining current soil moisture raw data from a soil moisture sensor.

[0055] Step S101 ensures real-time monitoring of soil moisture status and provides basic data for water adequacy assessment.

[0056] Step S102: Acquire current air temperature raw data and air humidity raw data from the air temperature and humidity sensor.

[0057] Step S102 provides important parameters of farmland climate conditions to support climate suitability calculation.

[0058] Step S103: obtaining current ambient light raw data from the light intensity sensor.

[0059] Step S103 captures changes in light intensity required for crop growth to support light suitability analysis.

[0060] Step S104: obtaining current raw meteorological data and meteorological forecast data for the next 48 hours by calling the meteorological platform interface.

[0061] Step S104 introduces external weather change information to improve environmental prediction capabilities and enhance the foresight of reasoning.

[0062] Step S1 collects rich soil, climate, and light raw data through multiple types of sensors and meteorological platforms, providing a reliable basis for subsequent multi-source data fusion and environmental state reasoning, and avoiding misjudgments caused by a single data source.

[0063] Step S2 includes the following sub-steps:

[0064] In step S201 , format standardization processing is performed on the soil moisture raw data, the air temperature raw data, the air humidity raw data, and the ambient light intensity raw data.

[0065] Step S201 unifies the data formats output by different devices to eliminate data incompatibility issues.

[0066] Step S202 , detecting and removing abnormal values ​​in the soil moisture raw data, the air temperature raw data, the air humidity raw data and the ambient light intensity raw data, wherein the abnormal values ​​include missing values, out-of-range values ​​and abnormal sensor fluctuation values.

[0067] Step S202 eliminates missing values, out-of-range values, and abnormal fluctuation values ​​to ensure data authenticity and stability.

[0068] Step S203 , locally aggregate the standardized soil moisture raw data, air temperature raw data, air humidity raw data, ambient light intensity raw data, meteorological raw data, and meteorological forecast data at preset time intervals to generate preprocessed data.

[0069] Step S203 reduces the amount of data, extracts valid information, and improves data utilization efficiency and processing speed.

[0070] The outlier removal in step S2 includes: comparing the mean value in a continuous time window with the single point data, identifying and deleting data points that exceed the set deviation range.

[0071] Step S2 improves data quality, reduces noise interference, reduces back-end processing burden, improves the accuracy of reasoning analysis and system efficiency through standardization, cleaning and aggregation processing, further improves data quality, and ensures that the data used for reasoning is representative and reliable.

[0072] Step S3 includes the following sub-steps:

[0073] Step S301: input the pre-processed data into the context fusion engine. The context fusion engine includes a semantic annotation module, a time-space synchronization module, and a situational reasoning module. The logic of the semantic annotation module is:

[0074] The preprocessed data are semantically annotated based on a unified agricultural scenario model, and the soil moisture raw data, air temperature raw data, air humidity raw data, ambient light intensity raw data, meteorological raw data and meteorological forecast data are respectively associated with the corresponding agricultural environmental factors.

[0075] Step S301 associates each data with a specific agricultural environmental factor to establish a clear semantic basis for contextual reasoning.

[0076] Step S302: The logic of the time and space synchronization module is:

[0077] Based on the timestamp synchronization rules, the soil moisture raw data, air temperature raw data, air humidity raw data, ambient light intensity raw data, meteorological raw data and meteorological forecast data are aligned on the same time axis, and spatially associated based on geographic location information to generate consistent situational data.

[0078] Step S302 ensures the temporal and spatial uniformity of different data sources, thereby improving the consistency and accuracy of the situational analysis.

[0079] Step S303: The logic of the contextual reasoning module is:

[0080] Preset fusion inference rules are applied to the consistent scenario data to convert the multi-source environmental factor data into unified agricultural scenario indicators, which include water sufficiency, climate suitability and light suitability, and output standardized scenario data.

[0081] Step S303 extracts core characteristic indicators of the environment to support situation classification and intelligent decision-making push.

[0082] The data processing logic package of the agricultural scenario model in step S3 is:

[0083] Define the correspondence between soil moisture raw data, air temperature raw data, air humidity raw data, ambient light intensity raw data, meteorological raw data and meteorological forecast data and agricultural environmental factors, and map various preprocessed data to situational indicators, which include water sufficiency, climate suitability and light suitability indicators.

[0084] Based on the corresponding relationship, the pre-processed data is aligned according to the unified timestamp synchronization rules, and spatial registration is performed according to the geographic location information to generate a context data set.

[0085] Based on the contextual data set, the fusion inference rules are applied to infer the comprehensive environmental status of the current farmland and output standardized contextual data.

[0086] Step S3 uses a unified agricultural context model to integrate multi-source data, align time and space, and perform contextual derivation. This accurately depicts the farmland environment and supports subsequent precision delivery. Agricultural context model data processing enhances the system's multi-source perception capabilities and reasoning reliability.

[0087] Step S4 includes the following sub-steps:

[0088] Step S401 : Preliminary screening of standardized context data is performed based on a preset rule engine.

[0089] Step S401 quickly eliminates irrelevant or normal situations to improve subsequent reasoning efficiency.

[0090] In step S402 , the machine learning inference model trained based on historical context data analyzes the filtered standardized context data, identifies potential risks and task suggestion events, and outputs them as push event data.

[0091] The machine learning inference model is based on a classification algorithm. The input features include soil moisture change trends, future meteorological changes, light intensity changes, and historical crop growth response patterns, and the output is the situational event category.

[0092] Step S402 intelligently identifies anomalies and risks to improve the accuracy of push decisions.

[0093] Step S403: classify the pushed event data according to the preset event importance and mark them as high priority events, medium priority events or low priority events.

[0094] Step S403 is to set the basis for the push strategy, supporting adaptive adjustment of the push frequency and content level.

[0095] Step S4 intelligently identifies potential risks or operational opportunities in farmland based on the integrated contextual data, significantly improving the accuracy and timeliness of event push.

[0096] Step S5 includes the following sub-steps:

[0097] Step S501: Determine the detail level and frequency of the pushed content based on the importance of the pushed event data. For high-priority events, push detailed content and send it immediately. For medium- and low-priority events, push summary content and send it at a fixed time period each day.

[0098] Step S501 avoids information redundancy and delay, and improves push efficiency.

[0099] Step S502: Adjust the push strategy according to the user's preset receiving preference. The push strategy includes filtering the data type of interest, limiting the push time period, and selecting an emergency notification method to generate push content.

[0100] When push content is generated, it is accompanied by environmental change charts and task suggestions based on the results of situational reasoning analysis to enhance the push effect.

[0101] Step S502 improves the relevance and acceptability of push notifications to meet the needs of different users.

[0102] Step S503: Send the push content to the user terminal via a mobile application, SMS, or email, and attach a situation description and suggested measures derived from the analysis of the standardized situation data to the push content.

[0103] Step S503 improves the intuitiveness and guidance of the pushed information and enhances the decision support capability.

[0104] Step S5 dynamically adjusts the push frequency and content level based on the situational analysis results to achieve accurate, timely, and personalized information push, optimize user experience, and reduce system load.

[0105] The present invention introduces a unified agricultural context model to fuse soil moisture data, air temperature and humidity data, light intensity data, and meteorological platform data, thereby achieving temporal and spatial alignment of different data sources, thereby deriving an accurate comprehensive state of the farmland environment. Compared with the traditional method of relying on a single sensor data for push, the present invention can effectively reduce false alarms through multi-source cross-validation and improve the accuracy and practicality of pushed information. Agricultural context events are identified through a rule engine and a machine learning inference model, and the push frequency and content level are adaptively adjusted based on the push event data, making the push more in line with actual needs, significantly reducing the data transmission and storage burden, improving user experience and system stability, and having high application value and promotion prospects.

[0106] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a fully hardware embodiment, a fully software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media containing computer-usable program code. The storage medium may be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0107] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A context-based IoT data push method, characterized in that: The steps include: Step S1, collecting raw data; Step S2: standardize, remove outliers, and perform preliminary aggregation on the raw data at the edge node, and output it as preprocessed data; Step S3: input the pre-processed data into a context fusion engine, perform time stamp alignment and spatial location association based on a unified agricultural context model, and fuse them into context data; Step S4: applying a rule engine and a machine learning inference model based on the contextual data to identify agricultural contextual events and output push event data; Step S5: Generate push content based on the push event data, adjust the push frequency and content level according to the urgency and user preference, and send it to the user terminal.

2. The context-based IoT data push method according to claim 1, wherein: The step S1 includes the following sub-steps: Step S101, obtaining current soil moisture raw data from the soil moisture sensor; Step S102, obtaining current air temperature raw data and air humidity raw data from the air temperature and humidity sensor; Step S103, obtaining current ambient light raw data from the light intensity sensor; Step S104: obtaining current raw meteorological data and meteorological forecast data for the next 48 hours by calling the meteorological platform interface.

3. The context-based IoT data push method according to claim 2, wherein: The step S2 includes the following sub-steps: Step S201, performing format standardization processing on the soil moisture raw data, the air temperature raw data, the air humidity raw data and the ambient light intensity raw data respectively; Step S202: detecting and removing abnormal values ​​in the soil moisture raw data, the air temperature raw data, the air humidity raw data, and the ambient light intensity raw data, wherein the abnormal values ​​include missing values, out-of-range values, and abnormal sensor fluctuation values; Step S203 , locally aggregate the standardized soil moisture raw data, air temperature raw data, air humidity raw data, ambient light intensity raw data, meteorological raw data, and meteorological forecast data at preset time intervals to generate preprocessed data.

4. The context-based IoT data push method according to claim 3, wherein: The step S3 includes the following sub-steps: Step S301: input the pre-processed data into a context fusion engine. The context fusion engine includes a semantic annotation module, a time-space synchronization module, and a situational reasoning module. The logic of the semantic annotation module is: The pre-processed data are semantically annotated based on a unified agricultural scenario model, and the raw data of soil moisture, air temperature, air humidity, ambient light intensity, meteorological data, and meteorological forecast data are respectively associated with the corresponding agricultural environmental factors; Step S302: The logic of the time and space synchronization module is: Based on timestamp synchronization rules, the soil moisture raw data, air temperature raw data, air humidity raw data, ambient light intensity raw data, meteorological raw data, and weather forecast data are aligned on the same timeline and spatially associated based on geographic location information to generate consistent contextual data. Step S303: The logic of the contextual reasoning module is: Preset fusion inference rules are applied to the consistent scenario data to convert the multi-source environmental factor data into unified agricultural scenario indicators, which include water sufficiency, climate suitability and light suitability, and output standardized scenario data.

5. The context-based IoT data push method according to claim 4, wherein: The step S4 includes the following sub-steps: Step S401, preliminarily screening the standardized context data based on a preset rule engine; Step S402: Analyze the filtered standardized context data using a machine learning inference model trained based on historical context data, identify potential risks and action suggestion events, and output them as push event data. Step S403: classify the pushed event data according to the preset event importance and mark them as high priority events, medium priority events or low priority events.

6. The context-based IoT data push method according to claim 5, wherein: The step S5 includes the following sub-steps: Step S501: Determine the level of detail and frequency of push content based on the importance of the push event data. For high-priority events, push detailed content and send it immediately. For medium- and low-priority events, push summary content and send it at a fixed time each day. Step S502: Adjust the push strategy based on the user's preset receiving preferences, including filtering the data type of interest, limiting the push time period, and selecting an emergency notification method, and generate push content; Step S503: Send the push content to the user terminal via a mobile application, SMS, or email, and attach a situation description and suggested measures derived from the analysis of the standardized situation data to the push content.

7. The context-based IoT data push method according to claim 6, wherein: The outlier elimination in step S2 includes: comparing the mean value in a continuous time window with the single point data, identifying and deleting data points that exceed a set deviation range.

8. The context-based IoT data push method according to claim 7, wherein: The machine learning inference model in step S4 is based on a classification algorithm, and the input features include soil moisture change trends, future meteorological changes, light intensity changes and historical crop growth response patterns, and the output is a situational event category.

9. The context-based IoT data push method according to claim 8, wherein: When the push content is generated in step S5, it is accompanied by an environmental change chart and operation suggestions based on the results of the situational reasoning analysis.

10. The context-based IoT data push method according to claim 8, wherein: The data processing logic package of the agricultural scenario model in step S3 is: Define the corresponding relationship between soil moisture raw data, air temperature raw data, air humidity raw data, ambient light intensity raw data, meteorological raw data and meteorological forecast data and agricultural environmental factors, and map various pre-processed data to situational indicators, including water sufficiency, climate suitability and light suitability indicators; Based on the corresponding relationship, the pre-processed data is aligned according to the unified timestamp synchronization rules, and spatial registration is performed based on the geographic location information to generate a context data set; Based on the contextual data set, the fusion inference rules are applied to infer the comprehensive environmental status of the current farmland and output standardized contextual data.

Images