An iot shared device data acquisition and abnormality analysis method

Abstract

This invention relates to the field of shared equipment monitoring and analysis, specifically disclosing a method for data acquisition and anomaly analysis of IoT shared equipment. The method includes: acquiring equipment locations and clustering them into zones; correcting and determining the equipment lifecycle based on usage duration and frequency; constructing dynamic state envelopes for equipment at different lifecycle stages; calculating a degradation index based on real-time data and the envelopes, identifying anomalies, and statistically analyzing available equipment; integrating the degradation index with historical anomaly events to assess equipment health, and analyzing the zone's acceptable demand based on health; predicting future demand, comparing it with the acceptable demand to calculate a supply-demand balance index, and determining whether cross-regional scheduling is necessary. This invention improves anomaly detection accuracy through lifecycle segmentation modeling; by quantifying equipment health and correlating it with acceptable demand, combined with multi-dimensional prediction of future demand, it achieves forward-looking cross-regional resource optimization allocation.

Description

Technical Field

[0001] This invention relates to the field of shared equipment monitoring and analysis, and specifically to a method for data acquisition and anomaly analysis of IoT shared equipment. Background Technology

[0002] With the rapid development of IoT technology, IoT-enabled shared devices (such as shared bicycles, shared power banks, and shared cars) have been widely applied in various aspects of urban life, providing users with great convenience. In the operation and management of shared devices, the ability to efficiently and accurately grasp the operational status of a massive number of devices and respond promptly to regional supply and demand imbalances is crucial in determining operating costs and user experience.

[0003] Currently, the monitoring and scheduling of shared equipment mainly relies on simple threshold alarms and manual experience-based scheduling. Existing technologies typically set uniform operating parameter thresholds for all devices, triggering alarms once data exceeds these limits. However, the performance of shared equipment changes significantly with usage duration and frequency over long-term use. Devices at different lifecycle stages (such as break-in, stabilization, and degradation) have vastly different baseline values ​​for their normal operating parameters. Using uniform, fixed thresholds for anomaly detection is highly prone to false alarms for normal degradation in aging equipment or missed detections for subtle anomalies in earlier equipment, resulting in low accuracy.

[0004] Furthermore, existing scheduling strategies are mostly based on simple statistics of historical order volumes or manual inspections, lacking in-depth analysis of the health status of the equipment itself. The number of available devices in a region does not equate to its actual service capacity; old or "sub-healthy" devices cannot withstand high-intensity continuous use like new devices.

[0005] Therefore, there is an urgent need for a method that can comprehensively consider individual equipment health differences and regional dynamic needs, thereby achieving accurate anomaly warning and intelligent scheduling. Summary of the Invention

[0006] To address the above problems, this invention proposes a method for data collection and anomaly analysis of IoT shared devices. The specific technical solution is as follows: A method for data collection and anomaly analysis of IoT shared devices includes the following steps: obtaining the current location of all shared devices within a target area, clustering them based on spatial distance to obtain several device groups, and recording the area where the device groups are located as partitions.

[0007] The current lifecycle stage of each device is determined based on its usage duration and frequency.

[0008] Based on historical operating data of devices at different lifecycle stages in an interference-free environment, dynamic state envelopes are constructed to characterize the normal operating boundaries of the devices.

[0009] Based on the real-time operating data of each device and the dynamic status envelope corresponding to its life cycle, calculate the status degradation index and determine whether the operation is abnormal. If abnormal, mark and warn; if normal, record as available device and count available devices in each partition.

[0010] Based on the condition deterioration index of available equipment and the cumulative effect of historical abnormal events, the health of the equipment is assessed and the demand heat that each zone can withstand is analyzed accordingly.

[0011] Predict the demand intensity of each region in the future and compare it with the acceptable demand intensity to calculate the supply and demand balance index, and then determine whether cross-regional equipment scheduling is necessary and provide feedback.

[0012] Compared with existing technologies, the IoT sharing device data acquisition and anomaly analysis method described in this invention has the following beneficial effects: 1. By introducing a lifecycle segmentation modeling strategy, this invention constructs dynamic state envelopes for devices at different lifecycle stages. These envelopes can adaptively fit the operating characteristics of devices at different stages, making the anomaly judgment benchmark more scientific and significantly reducing false alarms and false negatives.

[0013] 2. This invention achieves a quantitative assessment of equipment health by integrating a condition deterioration index that reflects the real-time deviation of the equipment with the abnormal cumulative effect of recording historical damage to the equipment, providing a reliable basis for analyzing the remaining lifespan and load-bearing capacity of the equipment.

[0014] 3. To address the issue of delayed scheduling decisions, this invention, on the one hand, correlates equipment health with the level of demand to quantify the upper limit of the actual service capacity of shared equipment within a region (supply side); on the other hand, by integrating multi-dimensional factors such as date, weather, and population flow, it enables the prediction of future demand (demand side). The quantification and prediction of both supply and demand lays a solid foundation for scientific decision-making. Attached Figure Description

[0015] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments 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.

[0016] Figure 1 This is a schematic diagram of the method flow of the present invention.

[0017] Figure 2 This is a structural block diagram of the present invention.

[0018] Figure 3 This is a schematic diagram of the overall process of the present invention. Detailed Implementation

[0019] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a data acquisition and anomaly analysis method for IoT sharing devices proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0020] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0021] The specific scheme of the data acquisition and anomaly analysis method for IoT sharing devices provided by the present invention will be described in detail below with reference to the accompanying drawings.

[0022] In one embodiment of the present invention, considering that the operating status of IoT shared devices changes continuously with usage time and frequency after deployment, accurately identifying anomalies from a massive number of devices, assessing the true health status of devices, and dynamically adjusting resource distribution based on device health status and regional demand are key to improving operational efficiency and user experience. The present invention provides a method for data collection and anomaly analysis of IoT shared devices. This method first performs spatial clustering based on device location and then models lifecycle segments to accurately identify abnormal devices; next, it quantitatively assesses the health of normal devices; based on the correlation between health and the level of acceptable demand, it determines whether supply and demand in each region are balanced, ultimately triggering cross-regional scheduling and feedback, thereby achieving optimized resource allocation.

[0023] Please see Figure 1 , Figure 2 and Figure 3 As shown, the present invention provides a method for data collection and anomaly analysis of IoT sharing devices, including the following steps S1 to S5.

[0024] S1. Locate the devices and cluster them into zones. Specifically, this includes the following steps: obtain the current location information of all shared devices in the target area, perform cluster analysis based on spatial distance, divide them into several device groups, and record the area where each device group is located as a zone.

[0025] In one specific embodiment of the present invention, all shared devices are the same type of device, with identical model and parameters.

[0026] It should be noted that the purpose of dividing the equipment group into zones in step S1 is to conduct subsequent demand heat prediction and resource scheduling on a regional basis. Compared with predicting for each isolated piece of equipment, regional data is more statistically consistent and feasible, and can effectively smooth out random fluctuations in individual equipment. At the same time, clear regional division provides a clear spatial unit basis for subsequent judgments on whether cross-regional scheduling from areas with surplus equipment to areas with shortage equipment is necessary, making scheduling decisions more operational and scientific.

[0027] After completing the spatial division of the target area, in order to accurately assess the overall status of the equipment in each partition, it is first necessary to accurately characterize the current wear and tear of each individual device.

[0028] S2. Abnormal device identification, specifically including the following steps: S2-1. Determine the current lifecycle stage of each shared device based on its cumulative runtime and usage count. The specific process includes: S2-1-1. Obtain the usage time of each device from its first activation to the current moment, and count its cumulative usage count.

[0029] S2-1-2. Calculate the actual usage frequency of the device based on the cumulative number of uses and the duration of use, and compare the actual usage frequency with the preset standard usage frequency range.

[0030] S2-1-3. If the actual usage frequency is within the preset range, the usage duration will be used directly as the runtime for lifecycle determination.

[0031] S2-1-4. If the actual usage frequency exceeds the preset range, calculate the compensation amount for the used time based on the deviation between the actual usage frequency and the preset range, and use the sum of the used time and the compensation amount as the running time. The compensation amount for the used time is calculated using the following mathematical model: .

[0032] in, This indicates the amount of compensation for the duration already used. Indicates the actual frequency of equipment use. These represent the lower and upper limits of the preset standard's operating frequency range, respectively. This represents the preset usage time correction amount corresponding to the unit usage frequency deviation. It should be noted that the usage time correction amount corresponding to the unit usage frequency deviation can be obtained by conducting accelerated life tests on the same batch of equipment or by performing regression analysis based on historical operating data. Specifically, this involves fitting a functional relationship between the usage frequency deviation and the additional wear time; the slope of this relationship is... value.

[0033] It should be noted that the standard usage frequency range is set based on the fact that, under normal operating conditions, the usage frequency of shared equipment typically fluctuates around a typical value. This typical value can be preset based on the equipment's design life, historical operating statistics of similar equipment, or industry experience. When the actual usage frequency deviates from this range, it indicates that the equipment is in atypical operating conditions, which may be due to accelerated aging caused by overuse or slowed aging due to idleness, thus affecting the accurate assessment of its remaining lifespan. Compensation is introduced to correct for the used time: when the actual usage frequency is higher than the upper limit, it indicates that the equipment is under high load and requires positive compensation to reflect the increased actual wear and tear; when the actual frequency is lower than the lower limit, it indicates that the equipment is under low load or idle and requires negative compensation to reflect the reduced wear and tear, thereby achieving a more accurate dynamic match between the equipment's operating status and its lifespan.

[0034] S2-1-5. Based on the preset lifecycle division standard, the runtime is matched with the runtime range corresponding to different lifecycles to determine the current lifecycle of each device.

[0035] In one specific embodiment of the present invention, the life cycle of the device is divided into a break-in period, a stabilization period, and a degradation period.

[0036] It's important to note that relying solely on usage time to determine equipment lifespan is often inaccurate. This is because two typical scenarios exist in actual operation: First, some equipment, despite long deployment times and usage durations far exceeding the average, remains largely idle due to remote locations, resulting in actual wear and tear far below the time-based indicators. Second, some equipment, though deployed for a short time, is overloaded due to its location in high-traffic areas, causing its aging process to far exceed the time-based indicators. By introducing usage frequency to dynamically adjust usage time, these issues can be effectively addressed. This incorporates the slowed aging caused by idleness or the accelerated aging caused by overload, significantly improving the accuracy and rationality of lifespan determination and laying the foundation for subsequent precise modeling.

[0037] After accurately determining the lifecycle stage of each device, it is necessary to establish evaluation criteria for the normal operating status of each stage in order to conduct subsequent anomaly detection.

[0038] S2-2. Based on the historical operating data of shared devices at different lifecycle stages in an interference-free environment, construct dynamic state envelopes to characterize the normal operating boundary of the devices. The specific process includes: S2-2-1. Extract the historical operating data of shared devices at different lifecycle stages in an interference-free environment as the base sample corresponding to each lifecycle.

[0039] S2-2-2. The isolated forest algorithm is used to perform unsupervised learning on the base samples corresponding to each life cycle, and the normal fluctuation range of various operating data of the equipment under each life cycle is identified and obtained, thereby constructing a dynamic state envelope that represents the normal operation boundary of the equipment at different life cycles.

[0040] As an example, the specific steps for constructing a dynamic state envelope using the Isolation Forest algorithm are as follows: Collected multidimensional historical operational data are used as training samples, with each sample being a multidimensional vector. An Isolation Forest model is constructed. For each sample, its path length in all isolated trees is calculated, and the average path length is obtained. An anomaly score is calculated based on the average path length; a higher score indicates a greater likelihood of an anomaly. The 95th percentile of the anomaly scores for all historical training samples is used as the anomaly determination threshold, thereby defining the normal operation boundary of the device. This boundary is the dynamic state envelope.

[0041] In another specific embodiment of the present invention, the normal fluctuation range of various operating data of the shared device in each life cycle is obtained through experimental testing.

[0042] It's important to note that segmenting the model for different lifecycle stages is necessary because the baseline values ​​and normal fluctuation ranges of key operating parameters (such as motor current, battery voltage, and vibration frequency) differ significantly between newly deployed, stable, and declining IoT sharing devices. Using a uniform, fixed threshold for anomaly detection throughout the entire lifecycle can easily lead to numerous false alarms or missed alarms. For example, normal wear and tear on older devices might be flagged as anomalies, while subtle anomalies in new devices might be overlooked due to their normal fluctuation range. Employing a dynamic envelope based on lifecycle segmentation adaptively matches the operational characteristics of devices at different stages, providing a more accurate comparison benchmark for subsequent anomaly detection.

[0043] Once a dynamic state envelope that accurately reflects the normal operating boundaries of equipment at each lifecycle stage is obtained, the real-time operating status of the equipment can be monitored and diagnosed based on this.

[0044] S2-3. Calculate the state degradation index and determine whether the equipment is operating abnormally. The specific process includes: S2-3-1. Calculate the state degradation index based on the real-time operating data of the shared equipment and the dynamic state envelope corresponding to its life cycle stage. The specific method is: S2-3-1-1. Collect the operating data of the equipment in real time within the sliding time window and obtain the dynamic state envelope corresponding to the life cycle of the equipment.

[0045] S2-3-1-2. For each time window, compare each of the equipment's operating data with the normal fluctuation range defined by its dynamic state envelope.

[0046] S2-3-1-3. Calculate the ratio of the number of operational data items that exceed the normal fluctuation range to the total number of monitored items, and use this as the item exceedance ratio.

[0047] S2-3-1-4. Identify the running data item with the largest fluctuation range, calculate the ratio of its amplitude exceeding the interval boundary to the length of the interval, and use it as the amplitude over-limit ratio.

[0048] S2-3-1-5. The normalized ratio of the number of items exceeding the limit and the ratio of the magnitude exceeding the limit are weighted and summed to obtain the equipment condition deterioration index.

[0049] S2-3-1-6. Calculate the equipment condition deterioration index in sequence for each time window to form a time sequence of equipment condition deterioration index.

[0050] It should be noted that when acquiring real-time operating data of a device, if the device is currently in use, its current operating data will be collected directly; if the device is currently idle, the operating data collected during its most recent use will be processed as real-time operating data.

[0051] S2-3-2. Based on the state deterioration index, determine whether the equipment is operating abnormally. The specific process includes: according to the state deterioration index time sequence of the equipment, determine whether the following conditions are met: (1) There is a single value in the time sequence that exceeds the preset state deterioration index threshold, and the number of times the limit is exceeded reaches or exceeds the set number threshold.

[0052] (2) The time series as a whole shows a monotonically increasing trend.

[0053] If any of the above conditions are met, the equipment is determined to be malfunctioning; otherwise, it is determined to be operating normally.

[0054] S2-3-3 If the operation is abnormal, mark the device and issue an early warning. If the operation is normal, record it as an available device and count the number of available devices in each partition.

[0055] It should be noted that the core purpose of detecting operational anomalies in equipment is twofold: firstly, to promptly identify and mark equipment with operational anomalies so that operators can perform precise maintenance, prevent minor faults from escalating into major problems, and reduce equipment scrap rates; secondly, and more importantly, to screen out currently functioning equipment from all available devices, preparing for a subsequent effective assessment of the demand capacity of the entire zone. Only normally functioning equipment can reliably provide services, and including abnormal equipment in the statistics would artificially inflate the service capacity of the zone, leading to scheduling decision errors.

[0056] After identifying all available devices, although they are currently functioning properly, their remaining lifespan and ability to withstand future demands vary. To quantify this difference, the health of each available device needs to be assessed.

[0057] S3. Assess the acceptable demand level, specifically including the following steps: S3-1. Assess the health of the equipment based on the condition deterioration index of available equipment and the cumulative effect of its historical abnormal events. The specific process includes: S3-1-1. Obtain the number of abnormal events that have occurred in the history of available equipment and the warning level corresponding to each abnormal event. Based on the preset impact factors corresponding to different warning levels, determine the impact factors of each abnormal event and accumulate them to obtain the historical abnormal cumulative impact coefficient.

[0058] S3-1-2. Divide the preset constant by the sum of the condition deterioration index of available equipment and the cumulative impact coefficient of historical anomalies to obtain the health assessment value of available equipment.

[0059] In another specific embodiment of the present invention, the health assessment value = 1 / (1 + condition deterioration index + historical abnormal cumulative impact coefficient).

[0060] S3-2. Based on the health status of the equipment and its life cycle stage, determine the demand heat that each zone can withstand in its current state. The method for analyzing the demand heat that a zone can withstand is as follows: S3-2-1. Select equipment at different life cycle stages and belonging to different health status ranges for experimental testing, and establish a correlation model between the equipment's life cycle, health status, and its withstand demand heat. The experimental testing process is as follows: S3-2-2. Use the test equipment continuously multiple times, with each use duration set as the baseline duration. Record the critical number of uses corresponding to the first occurrence of abnormal equipment operation data, and use this as the equipment's withstand demand heat.

[0061] S3-2-3. For each available device within a partition, based on its lifecycle stage and current health status, and combined with the correlation model, determine the acceptable demand intensity for each available device.

[0062] S3-2-4. Sum up the demand heat of all available devices in the partition to obtain the current demand heat that the partition can withstand.

[0063] As an example, the correlation model between a device's lifespan, health, and its available demand is a multiple regression equation fitted to experimental data: ,in To ensure the equipment can withstand the required heat. This refers to the lifecycle stage (which can be represented by numerical encoding). This is a health assessment value. Multiple groups were obtained through experiments. After obtaining the data, the function can be fitted using the least squares method. The specific form.

[0064] It should be noted that the base duration is set based on the average single-use duration designed for the device. For example, for shared bicycles, the base duration can be set to 30 minutes.

[0065] It should be noted that a comprehensive health index can be quantified by combining the fusion state degradation index and historical anomaly events. A higher health index means that the device has a longer remaining lifespan, greater robustness, and can withstand higher continuous usage intensity (i.e., demand intensity).

[0066] The health status of all available devices within a partition is transformed into their respective acceptable demand levels through an experimentally established correlation model. These levels are then summed to obtain the maximum service demand that the partition as a whole can reliably support, providing a key input for supply and demand balance analysis.

[0067] After determining the current upper limit of demand that a region can handle, it is also necessary to predict the total amount of demand that users may generate in the future. Only by making a forward-looking comparison between supply and demand can scientific scheduling decisions be made.

[0068] S4. Predict the demand for each partition in the future period, which includes the following steps: S4-1. Obtain the order data of each partition in the historical period, count the cumulative number of times the device is used in the partition and the duration of each use in the historical period, and convert the number of times the usage duration is converted into the equivalent number of times under the base duration, and use it as the demand for each partition in the historical period.

[0069] S4-2. Collect demand influencing factors for each region during each historical period. These factors include date attributes, regional population flow characteristics, and weather data.

[0070] S4-3. Based on the demand heat of each region and the corresponding demand influencing factors in each historical period, a training dataset is constructed. An LSTM network model with spatiotemporal attention mechanism is used to train the dataset to construct a spatiotemporal prediction model for demand heat.

[0071] S4-4. Obtain the demand influencing factors for the region in the future time period and input them into the prediction model to obtain the predicted demand heat value of the region in the future time period.

[0072] It should be noted that the demand heat spatiotemporal prediction model is updated using an online learning or periodic (e.g., daily, weekly) retraining mechanism to incorporate the latest order data and ensure that the model can reflect the latest changing trends in demand in a timely manner.

[0073] S5. Determine whether cross-regional equipment scheduling is required. Specifically, the following steps are included: S5-1. Calculate the supply-demand balance index based on the predicted demand and the tolerable demand for each region in the future. The specific process includes: calculating the ratio of the tolerable demand for the current region to the predicted demand in the future, and using it as the supply-demand balance index for the region.

[0074] S5-2. Determine whether cross-regional equipment scheduling is required and provide feedback. The specific process includes: S5-2-1. Compare the supply and demand balance index of each region with the preset expected range of supply and demand balance index.

[0075] S5-2-2 If the supply and demand balance index of all partitions is within the expected range, it is determined that there is no need to perform cross-regional equipment scheduling at present; otherwise, it is determined that scheduling is required.

[0076] When the predicted demand exceeds the capacity of a partition (i.e., the supply-demand balance index is too low), the system determines that equipment needs to be transferred from other partitions with surplus equipment; conversely, if the predicted demand is far below the capacity, the system may consider transferring equipment out.

[0077] It should be noted that when the equipment or partition is in the initial operation phase and the amount of historical data is less than a preset threshold, a cold start mode is executed: a common model or factory calibration data of similar equipment is used as the dynamic state envelope; the cumulative impact coefficient of historical anomalies is set to zero or a preset initial value; and the population and geographical characteristics of similar areas are compared to estimate the tolerable demand and future demand. After sufficient data has been accumulated, the system switches to the standard operation mode as described in this invention.

[0078] In one embodiment of the present invention, when it is determined that cross-regional equipment scheduling is required, corresponding scheduling suggestion information can be further output. For example, when there is a partition where supply is insufficient, according to the principle of proximity and the capacity margin of the oversupply partition, shared equipment is selected from the oversupply partition and scheduled to the partition with tight demand.

[0079] In summary, this invention acquires device locations and clusters them into zones; corrects and determines device lifecycles based on usage duration and frequency; constructs dynamic state envelopes for devices at different lifecycle stages; calculates degradation indices based on real-time data and envelopes, identifies anomalies, and counts available devices; integrates degradation indices with historical anomaly events to assess device health, and analyzes the zone's acceptable demand based on health; predicts future demand, compares it with acceptable demand to calculate a supply-demand balance index, and determines whether cross-zone scheduling is necessary.

[0080] This invention improves the accuracy of anomaly detection through lifecycle segmentation modeling; by quantifying equipment health and linking it with the level of demand that can be absorbed, combined with multidimensional prediction of future demand, it achieves forward-looking cross-regional resource optimization and allocation.

[0081] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0082] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0083] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.

[0084] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0085] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0086] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0087] Finally, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for data acquisition and anomaly analysis of IoT shared devices, characterized in that, Includes the following steps: Obtain the current location of all shared devices within the target area, cluster them based on spatial distance to obtain several device groups, and record the area where the device groups are located as partitions; Determine the current lifecycle stage of each device based on its usage duration and frequency. Based on historical operating data of devices at different lifecycle stages in an interference-free environment, dynamic state envelopes are constructed to characterize the normal operating boundaries of the devices. Based on the real-time operating data of each device and the dynamic status envelope corresponding to its life cycle, calculate the status degradation index and determine whether the operation is abnormal. If it is abnormal, mark it and issue a warning. If it is normal, record it as a usable device and count the usable devices in each partition. Based on the condition deterioration index of available equipment and the cumulative effect of historical abnormal events, the health of the equipment is assessed and the demand heat that each zone can withstand is analyzed accordingly. Predict the demand intensity of each region in the future and compare it with the acceptable demand intensity to calculate the supply and demand balance index, and then determine whether cross-regional equipment scheduling is necessary and provide feedback.

2. The method for data acquisition and anomaly analysis of IoT sharing devices according to claim 1, characterized in that: The method for determining the current lifecycle stage of each device is as follows: Obtain the usage time of each device from its first activation to the current time, and calculate its cumulative usage count; The actual usage frequency of the device is calculated based on the cumulative number of uses and the duration of use, and then compared with the preset standard usage frequency range. If the actual usage frequency is within the preset range, the usage duration will be used directly as the runtime for lifecycle determination. If the actual usage frequency exceeds the preset range, the compensation amount for the used time is calculated based on the deviation between the actual usage frequency and the preset range, and the sum of the used time and the compensation amount is taken as the running time. Based on the preset lifecycle segmentation criteria, the runtime is matched with the runtime range corresponding to different lifecycles to determine the current lifecycle of each device.

3. The method for data acquisition and anomaly analysis of IoT sharing devices according to claim 1, characterized in that: The method for constructing the dynamic state envelope is as follows: Historical operational data of shared devices at different lifecycle stages under interference-free conditions are extracted and used as base samples for each lifecycle stage. The isolated forest algorithm is used to perform unsupervised learning on the base samples corresponding to each life cycle, and the normal fluctuation range of various operating data of the device under each life cycle is identified and obtained, thereby constructing a dynamic state envelope that represents the normal operation boundary of the device at different life cycles.

4. The method for data acquisition and anomaly analysis of IoT sharing devices according to claim 1, characterized in that: The method for calculating the state degradation index is as follows: Real-time acquisition of device operating data within a sliding time window, and acquisition of the dynamic state envelope corresponding to the device's lifecycle; For each time window, the various operating data of the equipment are compared one by one with the normal fluctuation range defined by its dynamic status envelope; The ratio of the number of operational data items that exceed the normal fluctuation range to the total number of monitored items is used as the item excess ratio. Identify the data item with the largest fluctuation range, calculate the ratio of its amplitude exceeding the interval boundary to the length of the interval, and use this as the amplitude over-limit ratio; The equipment condition deterioration index is obtained by weighted summing the normalized number of out-of-limit ratio and the amplitude out-of-limit ratio. The equipment condition degradation index is calculated sequentially for each time window to form a time series of equipment condition degradation indexes.

5. The method for data acquisition and anomaly analysis of IoT sharing devices according to claim 4, characterized in that: The method for determining whether the equipment is operating abnormally is as follows: Based on the time series of the equipment's condition deterioration index, determine whether the following conditions are met: (1) There is a single value in the time series that exceeds the preset state degradation index threshold, and the number of times it exceeds the limit reaches or exceeds the set number threshold. (2) The time series as a whole shows a monotonically increasing trend; If any of the above conditions are met, the equipment is determined to be malfunctioning; otherwise, it is determined to be operating normally.

6. The method for data acquisition and anomaly analysis of IoT sharing devices according to claim 1, characterized in that: The method for assessing the health of available equipment is as follows: Obtain the number of abnormal events that have occurred in the history of available devices and the warning level corresponding to each abnormal event. Based on the preset impact factors corresponding to different warning levels, determine the impact factors of each abnormal event and accumulate them to obtain the historical abnormal cumulative impact coefficient. Divide the preset constant by the sum of the condition deterioration index of available equipment and the cumulative impact coefficient of historical anomalies to obtain the health assessment value of available equipment.

7. The method for data acquisition and anomaly analysis of IoT sharing devices according to claim 1, characterized in that: The method for determining the demand intensity of the analysis partition is as follows: Experimental tests were conducted on devices at different stages of their life cycle and within different health ranges to establish a correlation model between the device's life cycle, health level, and the level of demand it can withstand. The experimental testing process is as follows: The test equipment was used continuously multiple times, with each use duration set as the baseline duration. The critical number of uses corresponding to the first occurrence of abnormal equipment operation data was recorded, and this number was used as the equipment's withstandable heat load. For each available device within a partition, based on its lifecycle stage and current health, and combined with the correlation model, determine the demand intensity that each available device can withstand. The current acceptable demand level of a partition is obtained by summing up the demand levels of all available devices within the partition.

8. The method for data acquisition and anomaly analysis of IoT shared devices according to claim 7, characterized in that: The methods for predicting the demand intensity of a region in future periods include: Obtain order data for each partition within a historical time period, calculate the cumulative number of times devices are used and the duration of each use within each partition under each historical time period, and convert the number of times of use with different usage durations into the equivalent number of times of use under the base duration, and use it as the demand heat of each partition under each historical time period. Collect demand influencing factors for each region in different historical periods. These factors include date attributes, regional population flow characteristics, and weather data. A training dataset was constructed based on the demand intensity of each region and the corresponding demand influencing factors in each historical period. An LSTM network model with spatiotemporal attention mechanism was used to train the dataset to build a spatiotemporal prediction model for demand intensity. Obtain the demand influencing factors for each region in the future time period and input them into the prediction model to obtain the predicted demand heat value for each region in the future time period.

9. The method for data acquisition and anomaly analysis of IoT sharing devices according to claim 1, characterized in that: The method for calculating the supply-demand balance index is as follows: The ratio of the current demand that a region can withstand to the projected demand in the future is calculated and used as the supply-demand balance index for that region.

10. The method for data acquisition and anomaly analysis of IoT sharing devices according to claim 1, characterized in that: The method for determining whether cross-regional equipment scheduling is required is as follows: The supply and demand balance index of each region is compared with the preset expected range of supply and demand balance index. If the supply and demand balance index of all zones is within the expected range, it is determined that there is no need to perform cross-zone equipment scheduling at present; otherwise, it is determined that scheduling is required.

Images