Hospital informatization-based end-cloud collaborative information storage method

Abstract

The invention relates to the technical field of end-cloud collaboration, and particularly discloses an end-cloud collaboration information storage method based on hospital informatization, which comprises the following steps of: dividing data generated by terminal equipment into real-time data and non-real-time data according to a real-time requirement, and uploading the real-time data to a cloud platform in real time for storage; determining target data; drawing target curves, and classifying the target curves; for a single classification, determining a standard curve, and determining a target device and other devices; and the other devices delete the stored target data and only store the deviation condition with the target data of the target device, and the target data of the target device and the deviation condition are uploaded to the cloud platform. The system can reduce the overall power consumption cost of a hospital.

Description

Technical Field

[0001] This invention relates to the field of edge-cloud collaboration technology, specifically to an edge-cloud collaborative information storage method based on hospital informatization. Background Technology

[0002] Hospital informatization is a systematic project that takes patient information sharing as its core and uses modern information technologies such as cloud computing, the Internet of Things, and artificial intelligence to digitize, network, and intelligently transform the entire process of hospital medical care, management, and scientific research.

[0003] In an edge-cloud collaborative architecture, physiological parameters, images, or logs collected by terminal devices need to be encoded, encrypted, and retransmitted locally before being uploaded to a remote data center via Wi-Fi / cellular links. Compared to local storage, information of the same bit length has to go through stages such as RF power amplifiers, protocol handshakes, and modulation / demodulation, each step consuming additional power. Increased power consumption means increased electricity costs, and with the surge in high-resolution images and continuous monitoring data, the overall electricity costs for hospitals are rising accordingly. Summary of the Invention

[0004] The purpose of this invention is to provide a cloud-edge collaborative information storage method based on hospital informatization, thereby solving the above-mentioned technical problems.

[0005] The objective of this invention can be achieved through the following technical solutions:

[0006] A cloud-edge collaborative information storage method based on hospital informatization divides the data generated by terminal devices into real-time data and non-real-time data according to real-time requirements, and the real-time data is uploaded to the cloud platform for storage in real time.

[0007] Non-real-time data of the same type are grouped together, and non-real-time data generated by the same terminal device in the same group are marked as target data.

[0008] Plot the curve f(t) of the target data changing over time, denoted as the target curve. Classify the target curves. The similarity between any two target curves in the same category is greater than a preset value. t represents time and t belongs to the monitoring period [t1+Δt, t1']. t1' represents the upload time point of the next day, t1 represents the upload time point of the current day, Δt represents the preset upload duration, and the upload time point is set based on the electricity price of the hospital.

[0009] For a single category, the target curve with the highest overall similarity to the remaining target curves is taken as the standard curve, the terminal device corresponding to the standard curve is taken as the target device, and the remaining terminal devices are marked as the remaining devices.

[0010] The remaining devices delete the stored target data and only save the deviation from the target data of the target device. At upload time t1', the target data and deviation of the target device are uploaded to the cloud platform.

[0011] As a further aspect of the present invention: before plotting the curve of the target data changing over time, the method further includes: encoding the non-numerical target data to obtain numerical target data.

[0012] As a further aspect of the present invention: setting the upload time point includes:

[0013] The system obtains the electricity price of the hospital during the monitoring period, takes the time point with the lowest electricity price as the initial point, and marks the corresponding initial point as the target point if the electricity price at the next time point adjacent to the initial point is not the lowest. The target points are then sorted according to the time axis, and the last target point in the sort is taken as the upload time point.

[0014] As a further aspect of the present invention: before determining the deviation, the following steps are included:

[0015] The target device generates a compressed package of standard curves locally and adds a complete list of timestamps of the standard curves to the compressed package;

[0016] The target device initiates a short-range one-way broadcast within the local area network and sends the compressed packet cyclically only within a fixed frame window.

[0017] The remaining devices listen to the broadcast channel and cache the compressed package within a fixed frame window, and restore the standard curve locally.

[0018] As a further aspect of the present invention: determining the deviation includes:

[0019] The remaining devices read the ordinate value of their own target curve at each sampling moment and simultaneously read the ordinate value of the standard curve at the same moment, recording the difference between the two as a single-point deviation value.

[0020] The single-point deviation values ​​are sorted according to the time axis to form a deviation sequence, and a continuity check is performed on the deviation sequence. Adjacent deviation values ​​with the same sign and consistent trend are merged to obtain several deviation segments.

[0021] The remaining devices are considered as a set of deviation line segments as the deviation situation.

[0022] As a further aspect of the present invention, determining the deviation further includes:

[0023] The remaining devices only store the start time, end time, and corresponding deviation value for each deviation line segment, discarding the deviation values ​​within the line segment.

[0024] Before generating a new deviation segment, the remaining devices compare the similarity between the end deviation value of the previous deviation segment and the new deviation value; if the similarity is higher than a preset condition, the end time of the previous deviation segment is extended; otherwise, the next deviation segment is created.

[0025] As a further aspect of the present invention: data with high real-time requirements is treated as real-time data, and data with low real-time requirements is treated as non-real-time data.

[0026] The beneficial effects of this invention compared to the prior art are as follows:

[0027] The implementation steps of this invention can bring about the following beneficial effects:

[0028] 1) By dividing the data generated by the terminal device into real-time data and non-real-time data according to their real-time nature, only real-time data is uploaded immediately, while only deviation information is uploaded for non-real-time data. This method avoids the high-power operation of RF power amplification, protocol handshake, modulation and demodulation of complete data by all terminal devices, which greatly reduces the usage time and energy consumption of the wireless link on the terminal side, thereby reducing the overall communication energy consumption and extending the device's battery life;

[0029] 2) This invention obtains the hospital's electricity price curve for a day, uses the period with the lowest electricity price as candidate nodes, and automatically determines the final upload time point for each day by combining it with a preset upload duration. By arranging the upload of non-real-time data during the period with the lowest or relatively low electricity price, it can maximize the use of cost differences caused by electricity price fluctuations while ensuring data reliability and integrity, thereby reducing the power consumption of network operation and server-side storage management, and effectively reducing the hospital's overall electricity expenses.

[0030] 3) Through edge-cloud collaboration, this invention achieves layered processing of data transmission and storage, improving the overall reliability and operational efficiency of the system. It enables scientific and accurate management data analysis, enhances hospital operational levels, and provides accurate, scientific, and multi-dimensional data analysis support for further refined hospital management, supporting group-based medical collaboration and assisting in the construction of hospital groups. Furthermore, it fulfills the requirements for holistic hospital management information management, enabling a hospital director's management dashboard, management decision analysis, clinical quality control analysis, workload indicators, and tertiary hospital indicator decision analysis through a medical big data platform, comprehensively supporting the visualization of hospital leaders' decision-making and management, and the refinement of operational management. Attached Figure Description

[0031] The invention will now be further described with reference to the accompanying drawings.

[0032] Figure 1 This is a flowchart illustrating a cloud-edge collaborative information storage method based on hospital informatization according to the present invention. Detailed Implementation

[0033] 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.

[0034] Please see Figure 1 As shown, this invention is a cloud-edge collaborative information storage method based on hospital informatization, comprising the following steps:

[0035] Each piece of data collected by the device is timestamped and classified into two categories according to a pre-set response time limit: real-time data and non-real-time data. Data with high real-time requirements is classified as real-time data, and data with low real-time requirements is classified as non-real-time data. For example, ECG monitoring data is classified as real-time data, while shift records or equipment status logs are classified as non-real-time data.

[0036] Two parallel processing threads are started locally. One thread is responsible for packaging real-time data and pushing it directly to the cloud platform through a secure channel, while the other thread caches non-real-time data in local storage.

[0037] By differentiating the urgency of different data, critical vital sign information can be delivered to the cloud uninterruptedly and with low latency for real-time monitoring and decision support, while logs and status information that do not require immediate processing are uploaded later. This not only balances network bandwidth usage but also reduces the battery consumption of the terminal during continuous wireless transmission. By using local caching and batch uploading, excessive contention for link resources during peak periods is avoided, while historical data can be efficiently transmitted at appropriate times. This achieves intelligent scheduling of data streams, ultimately helping to maintain stable operation and providing reliable support for subsequent data analysis and storage management.

[0038] In local storage, all non-real-time data to be uploaded is scanned and divided according to business type. For example, image files, equipment operation logs and environmental monitoring records are grouped into different groups. For each group of data, all entries from the same device are aggregated according to the acquisition terminal identifier. For example, the temperatures uploaded by the same monitor over a period of time are grouped together as target data.

[0039] In a preferred embodiment of the present invention, non-numerical target data is encoded to obtain numerical target data;

[0040] Each piece of non-numerical information to be encoded is examined, such as device status text, alarm level labels, or diagnostic conclusions in inspection reports. This batch of data is deduplicated, and all possible different text values ​​are extracted, such as "normal," "warning," "fault," or "pending review." Each text value is assigned a unique numerical identifier, for example, "normal" is mapped to 1, "warning" to 2, "fault" to 3, and "pending review" to 4.

[0041] According to the time sequence of the original data records, the text content in each record is replaced with the corresponding numerical identifier to generate a set of numerical sequences; if a new text value is encountered, it is added to the existing mapping table, the next available identifier is assigned, and the historical data is updated accordingly; finally, what is obtained locally is the numerical target data that corresponds one-to-one with the original non-numerical information, which can be used for curve plotting and similarity calculation later.

[0042] For image data, non-real-time data of image type is processed locally by frame segmentation. Each frame of image is scanned line by line, and different brightness and color areas are mapped to corresponding numerical sequences according to preset grayscale or color block levels. For example, in a chest X-ray image, the bright parts of the bone contour and the low-brightness parts of the lung field are mapped to a set of continuous numbers respectively.

[0043] For video data, the entire video is first cut into several key frames. Contour tracking is performed on the motion region of interest in each frame. The turning points on the contour are extracted and converted into numerical points. Then, the contour difference between adjacent frames is calculated, and the difference magnitude is arranged into a numerical curve in chronological order.

[0044] Other types of data will not be elaborated here, as they can be converted using existing technologies.

[0045] After encoding and compression are completed locally, values ​​are read sequentially from the target data at preset time intervals, and each data point is mapped to a coordinate system in chronological order to generate a curve, such as the fluctuation of heart rate over time or the change of device temperature with working time.

[0046] Align the curves of each device with the same start and end times, where the start point is the first sample after adding a preset duration to the upload time of the day, and the end point is the upload time of the next day, so as to ensure that the curve range is consistent.

[0047] Compare the similarity of these aligned curves pairwise, such as the relative time of the peak and the duration of the upward or downward trend, and group curves with similarity exceeding the threshold into one category until all curves are classified.

[0048] After classification, the curves in the same category are highly consistent in shape, which makes it easier to select typical curves as representatives for deviation analysis.

[0049] By aligning and classifying the curves of each terminal within the same time interval, data from devices with similar operating characteristics can be automatically merged, ensuring that only a small number of representative curves need to be processed instead of all the original data. This reduces the amount of data that needs to be transmitted and stored while retaining key trend information. At the same time, determining the upload interval based on hospital electricity price fluctuations can unify all non-real-time data into the same period, facilitating batch processing and cost optimization. Thus, while ensuring data consistency and integrity, it reduces link usage and storage overhead, supporting efficient and stable operation.

[0050] In another preferred embodiment of the present invention, setting the upload time point includes:

[0051] Locally retrieve electricity price data for different time periods within the monitoring period, such as reading in prices for nighttime, early morning, daytime, and evening in sequence;

[0052] Iterate through these price values, find the period with the lowest price, and record this period as the initial low point; then compare the next period after the initial low point. If the electricity price in the next period starts to rise, consider the initial point to have ended and mark it as a target point.

[0053] Continue this process on the price curve throughout the day, and whenever a trough is found followed by a price rise, mark that trough as the new target point.

[0054] Sort all target points by time from earliest to latest, and select the latest one as the unified upload time for non-real-time data;

[0055] It is worth noting that the upload time is a fixed time, and the specific frequency of adjustment can be set by the administrator;

[0056] By using time-slot filtering and tagging processes, the upload of non-urgent data can be scheduled during the cheapest electricity periods, thereby avoiding the consumption of large amounts of electricity during peak electricity price periods. This reduces the energy expenditure required by the equipment for wireless transmission, lowers the overall electricity cost of the hospital at night and during non-core periods, and ensures that the data can be transmitted quickly and securely after the price drops, providing a solid foundation for the energy saving, consumption reduction and cost optimization of the solution.

[0057] Within the same category, the target curves generated by each device are paired up in turn, the similarity value of each pair of curves within the same time interval is calculated, and the sum of the similarity between each curve and all other curves is recorded.

[0058] Iterate through the sum of similarity of each curve, find the one with the highest value, identify it as the standard curve in this category, mark the terminal device corresponding to the standard curve as the target device, and mark the other devices as the other devices, so that in subsequent comparisons, only the deviation analysis of their respective target curves and standard curves needs to be performed.

[0059] After all deviation segments are generated locally, the other devices will clear the previously cached original numerical sequences and only retain representative deviation segments and their start and end times and deviation value records.

[0060] When the scheduled upload time arrives, a secure connection is initiated to the cloud, and the complete numerical sequence of the target device and the deviation segments of all devices are packaged into two data blocks; after the upload is completed, the device will receive a confirmation instruction from the cloud, and then clear the local cache of uploaded deviation information to make room for the next round of data collection and processing.

[0061] To avoid repeatedly sending the entire data segment to each device, only a complete curve plus a few lightweight deviation records are needed to reconstruct the monitoring curves of all devices. This significantly reduces transmission volume and cache usage, alleviating the burden on cloud storage and network. After uploading, the cloud platform will merge and reconstruct the standard curve with the deviation segments of each device to restore the complete data trend of each device.

[0062] After local processing is completed, other terminals will delete their original cached complete patient data, retain only the deviation segments between the data and the complete data uploaded by the target terminal, and record the start and end times and deviation values ​​of these deviation segments locally.

[0063] When the scheduled upload time arrives, the complete patient data of the target terminal and the deviation segments of each terminal are packaged locally, and then these two parts are uploaded to the cloud platform in sequence through the secure channel.

[0064] After the upload is confirmed, each terminal receives a cloud instruction to clear the local cache of uploaded deviation information, in order to prepare for the data collection and deviation calculation of the next cycle.

[0065] To avoid repeatedly uploading large amounts of complete patient data to each terminal, only one complete set of data and multiple sets of minor deviation information need to be uploaded. The complete data curves for all patients can then be reconstructed in the cloud. After uploading, the cloud platform will merge the complete data of the target terminal with the deviation segments of each terminal to restore the true data trend. Based on this, clinical monitoring, abnormal warnings, and data analysis will be performed to support timely decision-making by medical staff. At the same time, it will reduce the burden on network transmission and local storage, ensuring the continuous, efficient, and secure operation of the system.

[0066] Storing data on a cloud platform enables interconnectivity and resource integration among various business systems, reducing redundant construction. It provides interfaces to existing hospital business and information systems, allowing them to interface with the hospital information platform and inherit existing data resources and services. The hospital information platform is an open system, capable of adapting to various policies, technologies, and business developments. Software systems that adhere to information standardization can be connected to the platform, achieving data and application integration. The hospital information platform reduces the high coupling between business systems; figuratively speaking, it's like a standardized socket, with each hospital business system acting as a plug, making connection and replacement easier. By building a hospital information platform, patient information exchange, previously distributed across various business systems, is integrated into the platform, achieving interconnectivity between hospital departments and between hospitals, maximizing convenience for patients, frontline medical staff, and management personnel in analysis and decision-making.

[0067] In another preferred embodiment of the present invention, the process includes determining the deviation before:

[0068] The target device packages the selected standard curve with the corresponding complete timestamp list locally, compressing the curve sampling points and time nodes into a single data packet;

[0069] Initiate a short-range one-way broadcast in the local area network, divide the compressed package into several frames of the same size, and send these frames in a loop within a fixed time slot window. Each frame contains a sequence number so that the listening end can detect loss.

[0070] The remaining devices open the listening channel in the same window, receive and buffer all numbered frames, and after receiving them all, they are assembled into a complete compressed package according to the sequence number order, and decompressed to obtain the standard curve and all its timestamps, thus achieving local restoration;

[0071] By broadcasting the standard curve compressed package once within the local area network, other devices do not need to frequently request the same data from the cloud or target device, thereby significantly reducing the repeated transmission of similar data through the wireless link, reducing the RF power amplifier and handshake overhead between the terminal and the base station, and effectively reducing power consumption by utilizing the low power consumption characteristics of short-range broadcasting, ensuring the rapid distribution and local restoration of the standard curve, and providing an efficient and energy-saving foundation for subsequent deviation calculation and data upload.

[0072] In a preferred embodiment of the present invention, determining the deviation includes:

[0073] During local processing, other devices will first extract the current sampled value of their target curve from the local cache at each preset sampling time, such as the heart rate value recorded at the previous moment. At the same time, they will read the heart rate value corresponding to the same time point from the standard curve that was just restored, subtract the two values ​​to obtain the single-point deviation value, and record it in the local deviation list in chronological order.

[0074] The device scans the deviation list, treats adjacent deviation values ​​with the same sign and the same direction of change as a continuous interval, merges these continuous deviation values ​​into a deviation line segment, records the start and end times of the line segment and the corresponding start and end deviation values, and repeats the above merging operation until the entire list has been traversed.

[0075] The device saves all generated deviation segments locally as deviation information, preparing for subsequent uploading.

[0076] Comparing the point-by-point deviation between its own curve and the standard curve can accurately reflect the differences in data between devices. By checking and merging the deviation values ​​at each point continuously, not only is the number of storage items reduced, but the evolution trend of the deviation over time can also be preserved. This makes it easier to quickly restore the data trend and locate the range of change in the cloud or other analysis stages. This saves local storage space while ensuring data integrity and provides support for reducing the amount of data during subsequent network transmission.

[0077] The approach of combining single-point deviations into deviation segments is based on the characteristic that deviations tend to be stable or continuously increasing or decreasing over a period of time. By detecting the consistency of the sign and trend of adjacent deviation values, the segments are merged. This approach can accurately capture the start and end times of the deviation pattern while avoiding too many small segments due to short-term fluctuations. This makes the deviation information both temporally continuous and compressible, and forming a set of segments as a deviation case is both reasonable and efficient, which is helpful for subsequent difference analysis and cloud storage.

[0078] It should be noted that determining the aforementioned deviation also includes:

[0079] During local processing, the other devices will traverse the merged list of deviation segments. For each new sampling point, the deviation value corresponding to the sampling time will be read. Then, the end time of the previous deviation segment and its deviation value at the end time will be retrieved and compared. For example, the numerical differences or trends will be compared to see if they are consistent.

[0080] If the difference is within the preset fluctuation range, the end time of the previous line segment will be updated to the current time, and the original end deviation value will be replaced with the current deviation value.

[0081] If the difference exceeds this range, the current time is taken as the starting time of the new line segment, the current deviation value is taken as the initial deviation value of the new line segment, and the same method is used to determine whether the line segment should be extended or the next new line segment should be started in subsequent sampling.

[0082] By retaining only the start and end times of each line segment and the corresponding deviation value at the end time, the starting and ending intervals of the deviation trend can be accurately marked, while significantly reducing local storage requirements. At the same time, when determining whether to extend a line segment, the deviation is compared with the end of the previous segment to dynamically capture the boundary of deviation changes, ensuring that each line segment represents a continuous and relatively stable deviation trend. This provides concise yet information-rich key data for subsequent uploading and cloud analysis, helping the overall solution to reduce resource consumption while ensuring the integrity of difference detection.

[0083] Retaining only the start time, end time, and deviation value of the end time of a line segment will not miss important deviation fluctuations. This is because in the judgment logic for generating a new line segment, the creation of a new line segment will be triggered as long as there is a significant deviation jump. Any sudden change that exceeds the preset fluctuation range will be recorded as the starting point of the new line segment, thereby ensuring that all large-scale deviation changes can be accurately captured and saved in the boundary information of their respective line segments.

[0084] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the present invention should still fall within the scope of the present invention.

Claims

1. A method for end-to-cloud collaborative information storage based on hospital informatization, characterized in that, Includes the following steps: The data generated by the terminal device is divided into real-time data and non-real-time data according to the real-time requirements. The real-time data is uploaded to the cloud platform for storage in real time. Non-real-time data of the same type are grouped together, and non-real-time data generated by the same terminal device in the same group are marked as target data. Plot the curve f(t) of the target data changing over time, denoted as the target curve. Classify the target curves. The similarity between any two target curves in the same category is greater than a preset value. t represents time and t belongs to the monitoring period [t1+Δt, t1']. t1' represents the upload time point of the next day, t1 represents the upload time point of the current day, Δt represents the preset upload duration, and the upload time point is set based on the electricity price of the hospital. For a single category, the target curve with the highest overall similarity to the remaining target curves is taken as the standard curve, the terminal device corresponding to the standard curve is taken as the target device, and the remaining terminal devices are marked as the remaining devices. The remaining devices delete the stored target data and only save the deviation from the target data of the target device. At upload time t1', the target data and deviation of the target device are uploaded to the cloud platform.

2. The terminal-cloud collaborative information storage method based on hospital informatization according to claim 1, characterized in that, Before plotting the curve of the target data changing over time, the process also includes encoding the non-numerical target data to obtain numerical target data.

3. The terminal-cloud collaborative information storage method based on hospital informatization according to claim 1, characterized in that, Setting the upload time point includes: The system obtains the electricity price of the hospital during the monitoring period, takes the time point with the lowest electricity price as the initial point, and marks the corresponding initial point as the target point if the electricity price at the next time point adjacent to the initial point is not the lowest. The target points are then sorted according to the time axis, and the last target point in the sort is taken as the upload time point.

4. The terminal-cloud collaborative information storage method based on hospital informatization according to claim 1, characterized in that, Before determining the deviation, the following steps are included: The target device generates a compressed package of standard curves locally and adds a complete list of timestamps of the standard curves to the compressed package; The target device initiates a short-range one-way broadcast within the local area network and sends the compressed packet cyclically only within a fixed frame window. The remaining devices listen to the broadcast channel and cache the compressed package within a fixed frame window, and restore the standard curve locally.

5. The terminal-cloud collaborative information storage method based on hospital informatization according to claim 1, characterized in that, Determining the deviation includes: The remaining devices read the ordinate value of their own target curve at each sampling moment and simultaneously read the ordinate value of the standard curve at the same moment, recording the difference between the two as a single-point deviation value. The single-point deviation values ​​are sorted according to the time axis to form a deviation sequence, and a continuity check is performed on the deviation sequence. Adjacent deviation values ​​with the same sign and consistent trend are merged to obtain several deviation segments. The remaining devices are considered as a set of deviation line segments as the deviation situation.

6. The terminal-cloud collaborative information storage method based on hospital informatization according to claim 1, characterized in that, Determining the deviation also includes: The remaining devices only store the start time, end time, and corresponding deviation value for each deviation line segment, discarding the deviation values ​​within the line segment. Before generating a new deviation segment, the remaining devices compare the similarity between the end deviation value of the previous deviation segment and the new deviation value; if the similarity is higher than a preset condition, the end time of the previous deviation segment is extended; otherwise, the next deviation segment is created.

7. The terminal-cloud collaborative information storage method based on hospital informatization according to claim 1, characterized in that, Data with high real-time requirements is treated as real-time data, while data with low real-time requirements is treated as non-real-time data.

Images