A diabetes collaborative remote monitoring system based on 5G and internet of things

Abstract

The application relates to the field of information technology, and provides a diabetes collaborative remote monitoring system based on 5G and the Internet of Things. The system comprises the following steps: collecting blood glucose, activity, device identification and geographical position information through an initial data packet generation module based on 5G and the Internet of Things technology, and generating an initial data packet containing device type, data priority, patient position and time stamp. A service node matching module matches the service node according to the device type and the data priority. A path instruction generation module generates path instructions for directly connecting a specialized team or a regular platform in combination with a patient condition file. The system adopts 5G network slicing technology to dynamically allocate and adjust slice bandwidth, optimizes the transmission channel and monitors the transmission delay. Finally, a customized health guidance generation module generates a warning probability value based on the initial data packet and a pre-training analysis model, and provides customized health guidance. Remote precise monitoring and efficient management of diabetes patients are realized, and the quality and efficiency of medical services are improved.

Description

TECHNICAL FIELD

[0001] The application relates to the field of information technology, in particular to a diabetes collaborative remote monitoring system based on 5G and Internet of Things. BACKGROUND

[0002] Diabetes remote monitoring is a key direction in the field of smart medical care, which provides real-time and personalized health management for patients by integrating advanced communication technology and Internet of Things devices. With the continuous growth of the global number of diabetes patients, an efficient and accurate remote monitoring system has become a core means to improve the quality of life of patients and reduce medical costs. However, existing solutions have significant limitations in data transmission efficiency, device collaboration and personalized services. Many systems rely on traditional network architecture, which is difficult to cope with large-scale device connection and diversified data processing needs, resulting in high latency and slow response. In addition, existing methods generally lack dynamic adaptation ability for individual differences of patients, and cannot optimize data flow according to the complexity of the disease or geographical location, limiting the actual application value of the system.

[0003] In the diabetes collaborative remote monitoring system based on 5G and Internet of Things, the core challenge is concentrated in the precision of network connection and the differentiated needs of data processing. The high concurrency connection of a large number of monitoring devices puts high requirements on network bandwidth and stability, while the low latency characteristics of 5G network have potential, but how to realize intelligent matching of devices to service nodes is still a problem. This problem further deduces the optimization problem of data path, that is, how to dynamically select the optimal transmission path according to the device type and data characteristics to ensure the rapid processing of blood glucose data or activity data. The realization of path optimization depends on the dynamic perception of patient condition, and patients with complex conditions need to be directly connected to specialist teams, while stable patients' data needs to be efficiently shunted to the regular platform. If a multi-dimensional dynamic targeting mechanism cannot be realized, the system will be difficult to balance real-time and resource utilization.

[0004] Therefore, how to design a multi-dimensional targeting attachment mechanism based on 5G and Internet of Things technology to realize the precise matching of device connection, data path and patient condition has become a key problem to be solved in the diabetes collaborative remote monitoring system. SUMMARY

[0005] The application provides a diabetes collaborative remote monitoring system based on 5G and Internet of Things, mainly comprising:

[0006] An initial data packet generation module generates an initial data packet containing device type, data priority, patient geographical location and timestamp based on blood glucose data, activity data, device identifier and geographical location information collected by the monitoring device;

[0007] a service node matching module, matching a service node according to the device type, a preset device-service node mapping table and the data priority;

[0008] a path instruction generating module, extracting the patient geographic location and the data priority from the service node allocation result, combining the patient condition archive to obtain a condition complexity score, generating a direct specialist team path instruction if the condition complexity score is higher than a first threshold value, and generating a regular platform path instruction if the condition complexity score is lower than the first threshold value;

[0009] network slice allocation and adjustment, allocating a dedicated slice based on the path instruction using 5G network slice technology, dynamically adjusting the slice bandwidth based on the data priority to form an optimized transmission channel, monitoring the real-time state of the optimized transmission channel and calculating the transmission delay in combination with the time stamp, and re-adjusting the slice bandwidth and service node allocation if the transmission delay is higher than a second threshold value;

[0010] a customized health guidance generating module, generating a warning probability value based on the initial data packet using a pre-trained analysis model, and generating customized health guidance based on the warning probability value.

[0011] Preferably, the initial data packet generating module comprises:

[0012] obtaining blood glucose data, activity data, device identification and geographic location information from the monitoring device through the 5G network, and generating an original data set;

[0013] if the original data set is complete, judging an abnormal state according to preset threshold values of the blood glucose data and the activity data, and determining a data priority based on the abnormal state;

[0014] obtaining device type and patient information according to the data priority and the device identification, and generating a temporary data packet containing a time stamp;

[0015] calculating a patient location change trend using the geographic location and the time stamp, and generating location-related data;

[0016] if the location-related data is related to the abnormal state of the blood glucose data or the activity data, embedding the location-related data into the temporary data packet to generate an initial data packet.

[0017] Preferably, the service node matching module comprises:

[0018] obtaining a service node set, the service node set comprising a plurality of service nodes, each service node having corresponding service capability parameters and a current load state;

[0019] calculating a processing priority of each service node according to the service capability parameters and the current load state, and obtaining a node processing sequence;

[0020] According to the node processing sequence, a set of available data transmission channels is obtained from the set of service nodes, the set of data transmission channels including a plurality of data transmission channels, each data transmission channel having a corresponding bandwidth parameter and a delay parameter;

[0021] According to the bandwidth parameter and the delay parameter, a data transmission channel is determined;

[0022] A set of to-be-transmitted data is obtained, the set of to-be-transmitted data including a plurality of data packets, each data packet having a corresponding data priority;

[0023] If the data priority is higher than a preset threshold, a low-delay node in the service node is selected, and if the data priority is lower than the preset threshold, a load-balancing node in the service node is selected.

[0024] Preferably, the path instruction generation module comprises:

[0025] The geographical location of the patient and the data priority are obtained from the service node allocation result, and the complexity of the illness is calculated by a preset illness assessment model combined with the illness profile to obtain a complexity score;

[0026] If the complexity score is higher than a preset first threshold, the nearest specialist team node is obtained by querying a preset specialist team node table according to the geographical location of the patient, and a specialist team path instruction is generated;

[0027] If the complexity score is lower than the preset first threshold, a general platform path instruction is generated by querying a general platform node table according to the geographical location of the patient.

[0028] Preferably, the allocation and adjustment of the network slice comprises:

[0029] The data priority and the target node information are parsed from the path instruction, a dedicated network slice is generated using a preset slice allocation rule, and a dedicated slice configuration is obtained;

[0030] According to the dedicated slice configuration, the bandwidth parameter and the channel state of the current network are obtained, a slice bandwidth allocation ratio is calculated using a dynamic adjustment algorithm, and a bandwidth allocation scheme is obtained;

[0031] According to the bandwidth allocation scheme and the data priority, a priority identifier is generated, the dedicated network slice is embedded, and an optimized slice transmission configuration is obtained;

[0032] The real-time state of the transmission channel is obtained through channel monitoring, state data including a timestamp is extracted, and channel state data is obtained;

[0033] According to the channel state data, the transmission delay is calculated using the timestamp, and a delay value is obtained;

[0034] If the delay value is higher than a preset second threshold value, a dynamic adjustment algorithm is used to calculate the allocation proportion of the slice bandwidth, and a bandwidth adjustment scheme is obtained;

[0035] According to the bandwidth adjustment scheme, the current load state of the service node is queried, and an allocation algorithm is used to re-allocate the service node, and a node allocation scheme is obtained;

[0036] According to the bandwidth adjustment scheme and the node allocation scheme, an optimized configuration containing slice bandwidth and service node parameters is generated, and an optimized transmission configuration is obtained;

[0037] Through the optimized transmission configuration, the bandwidth parameters and node allocation of the transmission channel are updated, a verification algorithm is used to verify the availability of the configuration, and an available transmission channel is obtained.

[0038] Preferably, the customized health guidance generation module comprises:

[0039] Obtain time series data from blood glucose trend, and process the time series data using a data standardization algorithm to obtain standardized trend data;

[0040] According to the standardized trend data, a warning signal analysis method is used to calculate the warning signal probability, and a warning probability value is obtained;

[0041] If the warning probability value is higher than a preset threshold value, an intervention instruction generation algorithm is used to generate a professional intervention instruction, and intervention instruction data is obtained;

[0042] If the warning probability value is lower than a preset threshold value, a suggestion generation algorithm is used to generate a regular health suggestion, and health suggestion data is obtained;

[0043] According to the intervention instruction data or the health suggestion data, a data integration algorithm is used to generate a health management scheme, and health scheme data is obtained;

[0044] From the health scheme data, extract individualized features, and use a feature matching algorithm to generate customized health guidance, and obtain customized guidance data.

[0045] Preferably, the system further comprises an update module, and the update module comprises:

[0046] Obtain patient feedback data and blood glucose data from the patient terminal, and use a data cleaning algorithm to format and denoise the patient feedback data, and obtain standardized feedback data;

[0047] According to the standardized feedback data and blood glucose data, a trend analysis algorithm is used to calculate the blood glucose change curve, and blood glucose trend data is generated;

[0048] The blood glucose trend data and standardized feedback data are integrated into the patient's medical record using the record update rules to obtain the updated patient medical record.

[0049] Based on the updated patient medical records, a mapping adjustment algorithm is used to reallocate the relationship between devices and service nodes, generating an adjustment mapping table.

[0050] If the load of the service nodes in the adjusted mapping table is higher than a preset threshold, a load balancing algorithm is used to redistribute the nodes to obtain an optimized mapping table.

[0051] The technical solutions provided in this application embodiment may include the following beneficial effects:

[0052] This application discloses a collaborative remote monitoring system for diabetes based on 5G and the Internet of Things. It acquires blood glucose and activity data transmitted by monitoring devices, generates an initial data packet containing information such as device type and priority, matches appropriate service nodes according to a device-service node mapping table, and generates path instructions based on the complexity of the patient's condition. It employs 5G network slicing technology to allocate dedicated slices, dynamically adjusts bandwidth to form an optimized transmission channel, and monitors the transmission status in real time and makes timely adjustments. A pre-trained model is used to analyze blood glucose change trends, generate a health management plan, and send it to the patient's terminal. This application achieves intelligent collection, efficient transmission, real-time analysis, and personalized management of blood glucose data, improving the efficiency and quality of health management for diabetic patients and providing new technical support for telemedicine. Attached Figure Description

[0053] Figure 1 This is a schematic diagram of the framework of a 5G and IoT-based collaborative remote monitoring system for diabetes, as described in this application.

[0054] Figure 2 This is a schematic diagram of the path instruction generation module framework in the system of this application.

[0055] Figure 3 This is a schematic diagram of the network slice allocation and adjustment module in the system of this application. Detailed Implementation

[0056] The technical solutions of the embodiments of this application will be clearly and thoroughly described below with reference to the accompanying drawings. The described embodiments are merely some embodiments of this application.

[0057] like Figure 1 As shown in the figure, this embodiment is a 5G and IoT-based collaborative remote monitoring system for diabetes. The system specifically includes:

[0058] An initial data packet generation module generates an initial data packet containing device type, data priority, patient geographic location, and time stamp based on blood glucose data, activity data, device identification, and geographic location information collected by the monitoring device.

[0059] In this embodiment, blood glucose data, activity data, device identification, and geographic location information are obtained from the monitoring device through the 5G network to generate an original data set. If the original data set is complete, the abnormal state is determined according to the preset threshold of blood glucose data and activity data, and the data priority is determined based on the abnormal state. According to the data priority and the device identification, the device type and the patient information are obtained to generate a temporary data packet containing a time stamp. The patient location change trend is calculated using the geographic location and the time stamp to generate location-related data.

[0060] If the location-related data is related to the abnormal state of blood glucose data or activity data, the location-related data is embedded in the temporary data packet to generate an initial data packet.

[0061] In one possible implementation, blood glucose data, activity data, device identification, and geographic location information are obtained from the monitoring device through the 5G network to generate an original data set. The high bandwidth and low latency characteristics of the 5G network ensure real-time data transmission.

[0062] For example, the monitoring device is a wearable blood glucose monitor that collects blood glucose values (e.g., 6.5 mmol / L), activity data (e.g., 1000 steps), device identification (e.g., ID: CGM123), and geographic location (e.g., latitude and longitude coordinates) every 5 minutes. These data are quickly uploaded to the cloud through the 5G network to generate an original data set containing a time stamp. Real-time ensures the timeliness of the data, facilitating subsequent analysis.

[0063] Specifically, if the original data set is complete, the abnormal state is determined according to the preset threshold of blood glucose data and activity data, and the data priority is determined.

[0064] Preferably, the blood glucose threshold is set to hypoglycemia <3.9 mmol / L or hyperglycemia >10.0 mmol / L, and the activity data threshold is <500 steps / hour indicating low activity.

[0065] For example, a patient's blood glucose value is 11.2 mmol / L, and the number of steps is 300, triggering hyperglycemia and low activity abnormalities, and the data is marked as high priority. High-priority data is processed first to reduce the risk of delay and improve medical response efficiency.

[0066] In one embodiment, according to the data priority and device identification, the device type and patient information are obtained, and a temporary data packet is generated. The device identification CGM123 corresponds to a continuous glucose monitor, and the associated patient information such as name, age and medical history. For example, patient Zhang, 50 years old, with a history of type 2 diabetes, the temporary data packet contains a timestamp (such as 2025-04-27 10:00:00), blood glucose value, activity data and patient information. Patient information facilitates personalized diagnosis, and time stamp ensures data traceability.

[0067] Using geographic location and timestamp, the patient's location change trend is calculated, and location-related data is generated. For example, the patient moves from coordinate A to coordinate B within 1 hour, with a moving distance of about 2 kilometers, indicating normal travel. Location-related data is generated by analyzing latitude and longitude changes, reflecting patient behavior patterns. If blood glucose is abnormal (such as 11.2 mmol / L) and location change (such as in a vigorous exercise area) is related, then location-related data is embedded in the temporary data packet to generate an initial data packet. Location-related data enhances the context analysis of abnormal states, which helps to determine the cause of the abnormality.

[0068] It should be noted that the initial data packet is transmitted to a remote server through a 5G network for data processing. The initial data packet is transmitted in encrypted form to a hospital cloud server through a 5G network, and the server processes the data according to the priority, and high-priority data triggers a doctor's alarm. For example, data packets with blood glucose of 11.2 mmol / L are analyzed first, and the doctor receives an abnormal notification for timely intervention. The reliability of the 5G network ensures that data is not lost, and encrypted transmission protects patient privacy, and the overall process improves the response speed and security of telemedicine.

[0069] The service node matching module matches a service node according to the device type, a preset device-service node mapping table and a data priority;

[0070] In this embodiment, a set of service nodes is obtained, and the set of service nodes includes a plurality of service nodes, each service node having a corresponding service capability parameter and a current load state; according to the service capability parameter and the current load state, the processing priority of each service node is calculated to obtain a node processing sequence; according to the node processing sequence, a set of available data transmission channels is obtained from the set of service nodes, the set of data transmission channels includes a plurality of data transmission channels, each data transmission channel has a corresponding bandwidth parameter and a delay parameter; according to the bandwidth parameter and the delay parameter, a data transmission channel is determined; a set of data to be transmitted is obtained, the set of data to be transmitted includes a plurality of data packets, each data packet has a corresponding data priority; if the data priority is higher than a preset threshold, a low-delay node in the service node is selected, and if the data priority is lower than the preset threshold, a load-balanced node in the service node is selected.

[0071] For example, in a remote medical data processing scenario, the assigned service node is a node in a hospital cloud server cluster, and the service capability parameters of the node include computing power, storage capacity, and network interface bandwidth.

[0072] For example, node A has a computing capacity of 1000 data sets per second, a storage capacity of 1TB, and a bandwidth of 1Gbps. The current load status is obtained by querying the node's real-time task queue. Assuming node A is currently at 60% load, it indicates there is still processing capacity. The processing priority of the computing nodes is determined based on the service capacity parameters and load status.

[0073] Preferably, nodes with a load below 70% are marked as high priority, and a node processing sequence is generated, with node A ranked first.

[0074] In one possible implementation, available data transmission channels are obtained from node A based on the node processing sequence. The bandwidth and latency parameters of each channel are then queried. For example, channel 1 has a bandwidth of 800 Mbps and a latency of 2 ms; channel 2 has a bandwidth of 400 Mbps and a latency of 5 ms. When determining the preferred transmission channel, channel 1, with its higher bandwidth and lower latency, is preferentially selected as the preferred channel.

[0075] Specifically, the transmission order is determined based on data priority. Assuming a preset threshold is high-priority data, blood glucose data has a higher priority than the preset threshold and is assigned to channel 1; activity data has a lower priority than the preset threshold and is assigned to channel 2, thus forming a transmission channel allocation scheme.

[0076] It should be noted that, according to the transmission channel allocation scheme, the device's sending timestamp and data packet identifier are obtained. For example, if a wearable device sends blood glucose data at 10:00:00 on 2025-04-27, the data packet identifier is PKT123. A transmission data packet is generated, containing the timestamp, priority, and high-bandwidth channel information. The transmission data packet is then sent to node A via the 5G network.

[0077] Figure 2 This is a schematic diagram of the path instruction generation module framework in the system of this application, as shown below. Figure 2 As shown, the path instruction generation module includes: extracting the patient's geographical location and data priority from the service node allocation result, obtaining a disease complexity score by combining the patient's medical record, and generating a direct connection path instruction to the specialist team if the disease complexity score is higher than a first threshold, and generating a regular platform path instruction if it is lower than the first threshold.

[0078] In this embodiment, the patient's geographical location and data priority are obtained from the service node allocation results, and the complexity of the condition is calculated by combining the patient's medical record with a preset condition assessment model to obtain a complexity score.

[0079] If the complexity score is higher than the preset first threshold, the system queries the preset specialist team node table based on the patient's geographical location to obtain the nearest specialist team node and generate specialist team path instructions; if the complexity score is lower than the preset first threshold, the system queries the regular platform node table based on the patient's geographical location and generates regular platform path instructions.

[0080] For example, in a telemedicine data processing scenario, the patient's geographical location and data priority are obtained from the service node allocation results. For instance, if the patient is located in City C, the uploaded blood glucose data has a high priority. Combining this with the patient's medical records, such as a history of diabetes and recent complications, a pre-defined medical condition assessment model is used to analyze the data, resulting in a complexity score of 85, which is higher than the first threshold of 80. Based on this, the specialist team node table is queried according to the patient's geographical location to find the nearest diabetes specialist team node in a hospital in City C, generating a direct connection path instruction and forming a path allocation scheme. If the complexity score is lower than 80, for example, a score of 60, the regular platform node table is queried, a regional cloud platform node is selected, and a routing instruction is generated.

[0081] In one possible implementation, the load status and processing capacity of the target node are obtained based on the path allocation scheme. For example, the current load of the specialist team node is 50%, and its processing capacity includes processing 800 sets of data per second, 2TB of storage capacity, and 2Gbps bandwidth. Using a load balancing algorithm, this node is assigned a high priority, and a node processing sequence is generated, prioritizing this node. Its available data transmission channels are then queried: Channel 1 has a bandwidth of 1.5Gbps and a latency of 1ms, while Channel 2 has a bandwidth of 1Gbps and a latency of 3ms.

[0082] Preferably, after comparing parameters, channel 1 is selected as the preferred transmission channel.

[0083] Specifically, based on the transmission channel scheme and the high priority of blood glucose data, a transmission data packet is generated, containing the patient's location in City C, priority identifier, and high-bandwidth channel information. The data packet is sent to the target node via the 5G network, the reception status is queried, the integrity is verified using a CRC check algorithm, and if the data is confirmed to be correct, a data reception confirmation is generated.

[0084] Understandably, based on the confirmation result, the target node's processing log is retrieved to determine whether the blood glucose data analysis is complete. If complete, a processing completion identifier, such as CMP789, is generated; if incomplete due to network fluctuations, an abnormal status, such as ERR012, is recorded.

[0085] In one embodiment, the disease assessment model generates a disease complexity score based on the patient's historical data and real-time monitoring indicators through weighted analysis. For example, a history of diabetes accounts for 40% of the weight, complications for 30%, and real-time blood glucose levels for 30%, ensuring that the disease complexity score accurately reflects the complexity of the disease.

[0086] The path allocation scheme prioritizes the nearest specialist node through geographical matching, reducing transmission latency. The load balancing algorithm considers node load and processing capacity, dynamically adjusting priorities to avoid resource waste. Channel selection considers both bandwidth and latency to ensure rapid transmission of high-priority data. For example, patient blood glucose data is quickly delivered to the specialist node via a high-bandwidth channel, analyzed, and used to generate treatment suggestions that are promptly relayed to the patient. Dynamic adjustment of node processing sequences ensures load balancing and extends server lifespan. Recording abnormal states facilitates subsequent network configuration optimization and reduces the risk of data loss. These measures collectively ensure the timely processing of high-priority data, improving the patient's treatment experience.

[0087] Figure 3 This is a schematic diagram of the network slice allocation and adjustment module in the system of this application, as shown below. Figure 3 As shown, the allocation and adjustment of network slices involves allocating dedicated slices using 5G network slicing technology according to the path instructions, dynamically adjusting slice bandwidth based on data priority to form an optimized transmission channel, monitoring the real-time status of the optimized transmission channel and calculating the transmission delay in conjunction with the timestamp, and readjusting slice bandwidth and service node allocation if the transmission delay is higher than a second threshold.

[0088] In this embodiment, data priority and target node information are parsed from path instructions, and a dedicated network slice is generated using a preset slice allocation rule to obtain a dedicated slice configuration. Based on the dedicated slice configuration, the bandwidth parameters and channel status of the current network are obtained, and a dynamic adjustment algorithm is used to calculate the slice bandwidth allocation ratio to obtain a bandwidth allocation scheme. Based on the bandwidth allocation scheme and the data priority, a priority identifier is generated and embedded into the dedicated network slice to obtain an optimized slice transmission configuration. The real-time status of the transmission channel is obtained through channel monitoring, and status data including timestamps is extracted to obtain channel status data. Based on the channel status data, the transmission delay is calculated using timestamps to obtain a delay value. If the delay value is higher than a preset second threshold, a dynamic adjustment algorithm is used to calculate the slice bandwidth allocation ratio to obtain a bandwidth adjustment scheme. Based on the bandwidth adjustment scheme, the current load status of the service nodes is queried, and an allocation algorithm is used to reallocate the service nodes to obtain a node allocation scheme. Based on the bandwidth adjustment scheme and the node allocation scheme, an optimized configuration including slice bandwidth and service node parameters is generated to obtain an optimized transmission configuration. Through the optimized transmission configuration, the bandwidth parameters and node allocation of the transmission channel are updated, and a verification algorithm is used to verify the configuration availability to obtain an available transmission channel.

[0089] For example, in a remote medical data processing scenario, data priority and target node information are parsed from the path instruction. Assume the instruction indicates that the patient's heart rate data has a high priority, and the target node is the cardiovascular specialist node of a hospital in City D. The parsing process involves reading the instruction header metadata, extracting the priority field and the high 16-bit address code, and confirming the node's identity. It should be noted that priorities are divided into three levels: high, medium, and low. High priority corresponds to real-time monitoring data, ensuring rapid processing.

[0090] In one possible implementation, dedicated network slices are generated using preset slice allocation rules. These rules allocate independent slice resources within the 5G network based on data priority and node type. For example, high-priority heart rate data is allocated slice ID SL001, with 2Gbps of reserved bandwidth, isolating other low-priority traffic. The dedicated slice configuration includes slice ID, bandwidth range, and priority identifier to ensure data transmission stability.

[0091] Specifically, based on the slice configuration, the current network bandwidth parameters and channel status are obtained. Assuming the total network bandwidth is 10Gbps, currently 60% utilized, channel 1 latency is 2ms, and channel 2 latency is 5ms, the dynamic adjustment algorithm allocates 70% of the available bandwidth (2.8Gbps) to slice SL001 based on priority and channel status. The bandwidth allocation scheme prioritizes channel 1 by comparing channel latency.

[0092] Preferably, a priority identifier PRI-H is generated based on the bandwidth allocation scheme and high priority, and embedded into the SL001 slice to obtain an optimized slice transmission configuration. The identifier is embedded in the slice header to notify the network device to prioritize scheduling. The target node channel status is queried, and the channel availability is verified using a CRC check algorithm. If it is confirmed that there is no packet loss in channel 1, it is selected as an available transmission channel.

[0093] For example, in a remote medical data transmission scenario, channel monitoring is used to obtain the real-time status of the transmission channel. Channel monitoring collects data periodically through network devices, generating status data containing timestamps. For instance, the monitoring device records the channel's latency and throughput once per second, generating status data including the timestamp 2025-04-27 10:00:00, latency 3ms, and bandwidth utilization 70%. This data reflects the channel's real-time performance, providing a basis for subsequent optimization.

[0094] In one possible implementation, the transmission delay is calculated based on channel status data. Timestamp calculation is performed by comparing the time difference between data transmission and reception to determine the delay value. For example, if the transmission time is 10:00:00.000 and the reception time is 10:00:00.003, the delay is 3ms. If a preset second threshold is 5ms, and the current delay is below the threshold, it indicates good channel performance. If the delay reaches 6ms, exceeding the threshold, bandwidth adjustment is triggered.

[0095] Specifically, the dynamic adjustment algorithm calculates the slice bandwidth allocation ratio based on the latency value. The algorithm comprehensively considers latency and current bandwidth utilization to generate a bandwidth adjustment scheme. For example, if the total channel bandwidth is 10Gbps, and slice SL002 currently occupies 3Gbps with a latency of 6ms, the algorithm adjusts the bandwidth of SL002 to 2Gbps, freeing up resources for high-priority data. This scheme ensures the stability of critical data transmission.

[0096] It's important to note that querying the load status of service nodes is crucial for optimizing configurations. Load status includes the node's CPU utilization and memory usage.

[0097] For example, in a hospital in City D, node A has a CPU utilization of 60%, while node B has 80%. The allocation algorithm prioritizes node A, generating a node allocation scheme to avoid overloading nodes and improve response speed.

[0098] Preferably, the optimized transmission configuration is generated based on a bandwidth adjustment scheme and a node allocation scheme, including slice bandwidth and service node parameters. For example, the configuration specifies a slice bandwidth of 2Gbps for SL002 and a service node of node A. This configuration is issued through the network controller to update the channel parameters. The verification algorithm uses CRC checksum to verify that the configuration is correct and ensure the channel is available.

[0099] Understandably, these measures, through real-time monitoring, dynamic adjustment, and encryption protection, ensure the efficient transmission of remote medical data. Accurate acquisition of channel status and flexible bandwidth allocation reduce latency risks. Reasonable allocation of node load improves processing efficiency. Encrypted logs ensure data security and traceability, providing reliable protection for medical data transmission.

[0100] The customized health guidance generation module generates warning probability values ​​based on the initial data packet using a pre-trained analysis model, and then generates customized health guidance based on the warning probability values.

[0101] In this embodiment, time-series data is obtained from blood glucose change trends, and the time-series data is processed using a data standardization algorithm to obtain standardized trend data. Based on the standardized trend data, a warning signal analysis method is used to calculate the probability of a warning signal to obtain a warning probability value. If the warning probability value is higher than a preset threshold, an intervention instruction generation algorithm is used to generate a specialist intervention instruction to obtain intervention instruction data. If the warning probability value is lower than the preset threshold, a suggestion generation algorithm is used to generate routine health suggestions to obtain health suggestion data. Based on the intervention instruction data or the health suggestion data, a data integration algorithm is used to generate a health management plan to obtain health plan data. Personalized features are extracted from the health plan data, and a feature matching algorithm is used to generate customized health guidance to obtain customized guidance data.

[0102] For example, in telemedicine scenarios, obtaining time-series data from blood glucose trend analysis requires ensuring data continuity and accuracy. Time-series data is typically generated by continuous glucose monitoring devices, such as recording blood glucose levels hourly, generating data points like a blood glucose level of 6.7 mmol / L at 09:00:00 on April 27, 2025, and 7.0 mmol / L at 10:00:00. These data reflect the patient's blood glucose fluctuations, laying the foundation for subsequent analysis.

[0103] In one possible implementation, a data standardization algorithm processes time-series data to eliminate dimensional differences. Z-score standardization is used to convert blood glucose values ​​into standardized values ​​with a mean of 0 and a standard deviation of 1. For example, for blood glucose sequences of 6.7, 7.0, and 6.5 mmol / L, the calculated mean is 6.73 and the standard deviation is 0.25, resulting in standardized values ​​such as -0.12, 1.08, and -0.92. Standardized trend data facilitates cross-patient comparisons and model input.

[0104] Specifically, the early warning signal analysis method calculates the early warning probability based on standardized trend data. It employs a logistic regression model, taking standardized blood glucose levels, fluctuation frequency, and other features as input, and outputting the early warning probability. For example, if the model analyzes 6 hours of standardized data and the fluctuation frequency is higher than the normal range, it outputs an early warning probability of 0.78. A preset threshold of 0.7 is set; exceeding this probability triggers subsequent intervention.

[0105] Preferably, if the probability of an early warning is higher than a threshold, the intervention instruction generation algorithm generates a specialist intervention instruction. The algorithm combines the patient's medical history, such as a history of type 2 diabetes, to generate an instruction such as "It is recommended to adjust the insulin dose to 12 units." If the probability is lower than the threshold, the suggestion generation algorithm generates health recommendations, such as "Walk for 30 minutes daily." These instructions or recommendations form the basis for personalized intervention.

[0106] In one embodiment, the data integration algorithm integrates intervention instructions or health recommendations into a health management plan. For example, it integrates insulin adjustment instructions, dietary recommendations such as "reduce the intake of high-sugar foods," and exercise plans to generate health plan data. The plan data includes a schedule, such as "take medication at 08:00 daily and exercise at 18:00 daily."

[0107] It's important to note that personalized features are extracted from health plan data, and a feature matching algorithm is used to generate customized health guidelines. For example, if a patient prefers a low-carbohydrate diet, the generated guideline might be "Consume 50g of carbohydrates daily, prioritizing whole grains." This customized guideline data is more closely aligned with the patient's lifestyle.

[0108] In this embodiment, the system further includes an update module, which includes: acquiring patient feedback data and blood glucose data from the patient terminal; formatting and denoising the patient feedback data using a data cleaning algorithm to obtain standardized feedback data; calculating a blood glucose change curve using a trend analysis algorithm based on the standardized feedback data and blood glucose data to generate blood glucose trend data; and integrating the blood glucose trend data and standardized feedback data into the patient's medical record using a record update rule to obtain an updated patient medical record.

[0109] Based on the updated patient medical records, a mapping adjustment algorithm is used to reallocate the relationship between devices and service nodes, generating an adjustment mapping table. If the load of service nodes in the adjustment mapping table is higher than a preset threshold, a load balancing algorithm is used to reallocate nodes, resulting in an optimized mapping table.

[0110] For example, obtaining patient feedback data and blood glucose data from patient terminals is typically done through smart devices such as blood glucose meters or mobile applications. Patient feedback data may include dietary habits, exercise patterns, etc., such as "200 grams of rice and 30 minutes of walking per day." Blood glucose data is measured using a blood glucose meter, such as "fasting blood glucose 6.5 mmol / L." This data provides the basis for subsequent analysis, ensuring the accuracy of the health management plan.

[0111] In one possible implementation, a data cleaning algorithm formats and denoises patient feedback data. Formatting standardizes data from different sources, such as converting "200 grams of rice" to "200 grams of staple food intake." Denoising removes outliers, such as eliminating obviously erroneous entries like "300 minutes of walking per day." The processed data generates standardized feedback data, such as "200 grams of staple food intake, 30 minutes of exercise," providing reliable input for trend analysis.

[0112] Specifically, the trend analysis algorithm calculates a blood glucose change curve based on standardized feedback data and blood glucose data. The algorithm combines time-series data to analyze the relationship between blood glucose fluctuations and diet and exercise. For example, if a patient's three-day data are "6.5 mmol / L, 6.8 mmol / L, 6.3 mmol / L", and combined with feedback that "daily staple food intake decreased by 50 grams", the algorithm generates blood glucose trend data showing a "blood glucose trend is decreasing". This provides a basis for updating the patient's medical record.

[0113] Preferably, the record update rules integrate blood glucose trend data and standardized feedback data into the patient's medical record. The update rules may include classification criteria, such as "a decrease in blood glucose is classified as a stable condition." For example, the original record might state "significant blood glucose fluctuations," which could be updated to "stable blood glucose, reduced staple food intake." Updating the record data ensures comprehensive patient information and facilitates subsequent resource allocation.

[0114] It should be noted that the mapping adjustment algorithm reallocates the relationship between devices and service nodes based on updated file data. Service nodes refer to cloud servers or medical institution interfaces. The algorithm analyzes the priority of the condition in the file; for example, "stable blood sugar" is assigned to a low-priority node, generating an adjustment mapping table, such as "Device ID_001 corresponds to node A". This optimizes resource allocation efficiency.

[0115] In one embodiment, if the mapping table shows that the service node is overloaded, such as "Node A is handling 1000 devices, exceeding the threshold of 800", the load balancing algorithm reallocates nodes. The algorithm may allocate some devices to node B, generating an optimized mapping table, such as "Device ID_001 to node B". This ensures stable system operation.

[0116] Understandably, the path generation algorithm calculates the optimal communication path based on the optimized mapping table. The algorithm evaluates network latency, bandwidth, etc., such as "Node B latency 15ms, bandwidth 20Mbps," and generates the optimized path rule "Device ID_001 communicates via channel C." This improves data transmission efficiency.

[0117] For example, the data distribution algorithm sends updated service configurations, such as "monitor blood glucose twice daily," to the patient's terminal by optimizing path rules. After sending, the terminal returns an acknowledgment signal, such as "Configuration received successfully." This confirmation of transmission completion ensures that patients receive services promptly, improving the reliability of health management.

[0118] The above description is merely a preferred embodiment of one or more embodiments of this specification and is not intended to limit the scope of one or more embodiments of this specification. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of one or more embodiments of this specification should be included within the protection scope of one or more embodiments of this specification.

Claims

1. A collaborative remote monitoring system for diabetes based on 5G and the Internet of Things, characterized in that, The system includes: The initial data packet generation module generates an initial data packet containing device type, data priority, patient geographical location and timestamp based on blood glucose data, activity data, device identification and geographical location information collected by the monitoring device; The service node matching module matches service nodes based on the device type, a preset device-service node mapping table, and data priority. The path instruction generation module extracts the patient's geographical location and data priority from the service node allocation results, combines the patient's medical record with a preset medical condition assessment model to calculate the complexity of the condition, and obtains a medical condition complexity score. If the medical condition complexity score is higher than a preset first threshold, the module queries a preset specialist team node table based on the patient's geographical location to obtain the nearest specialist team node and generates a specialist team path instruction; if the medical condition complexity score is lower than the preset first threshold, the module queries a regular platform node table based on the patient's geographical location and generates a regular platform path instruction. The network slice allocation and adjustment module allocates dedicated slices using 5G network slicing technology according to the path instructions, dynamically adjusts the slice bandwidth based on data priority to form an optimized transmission channel, monitors the real-time status of the optimized transmission channel and calculates the transmission delay in combination with the timestamp, and readjusts the slice bandwidth and service node allocation if the transmission delay is higher than a second threshold. The customized health guidance generation module generates warning probability values ​​based on the initial data packet using a pre-trained analysis model, and generates customized health guidance based on the warning probability values. The initial data packet generation module includes: The system uses a 5G network to acquire blood glucose data, activity data, device identification, and geolocation information from monitoring devices to generate a raw dataset. If the original dataset is complete, abnormal states are determined based on preset thresholds for blood glucose and activity data, and data priorities are determined based on the abnormal states. Based on data priority and device identifier, obtain device type and patient information, and generate a temporary data packet containing a timestamp; Using geographic location and timestamps, the trend of patient location changes is calculated to generate location-related data; If the location-related data is associated with abnormal blood glucose or activity data, the location-related data is embedded into a temporary data packet to generate an initial data packet.

2. The system according to claim 1, characterized in that, The service node matching module includes: Obtain a set of service nodes, which includes multiple service nodes, each of which has corresponding service capability parameters and current load status; Based on the service capability parameters and the current load status, the processing priority of each service node is calculated to obtain the node processing sequence; According to the node processing sequence, a data transmission channel set is obtained from the service node set. The data transmission channel set includes multiple data transmission channels, each with corresponding bandwidth parameters and delay parameters. Based on the bandwidth parameters and delay parameters, a data channel to be transmitted is selected. The data set to be transmitted includes multiple data packets, each with a corresponding data priority. If the data priority is higher than the preset threshold, the low-latency node among the service nodes is selected; if it is lower than the preset threshold, the load balancing node among the service nodes is selected.

3. The system according to claim 1, characterized in that, The network slice allocation and adjustment module includes: Data priority and target node information are parsed from the path instructions, and dedicated network slices are generated using preset slice allocation rules to obtain dedicated slice configurations; Based on the dedicated slice configuration, the bandwidth parameters and channel status of the current network are obtained, and a dynamic adjustment algorithm is used to calculate the slice bandwidth allocation ratio to obtain a bandwidth allocation scheme. Based on the bandwidth allocation scheme and the data priority, a priority identifier is generated and embedded into the dedicated network slice to obtain an optimized slice transmission configuration; The real-time status of the transmission channel is obtained by channel monitoring, and the status data containing timestamps is extracted to obtain the channel status data. Based on the channel status data, the transmission delay is calculated using the timestamps to obtain the delay value. If the latency value is higher than the preset second threshold, a dynamic adjustment algorithm is used to calculate the allocation ratio of slice bandwidth to obtain a bandwidth adjustment scheme. Based on the bandwidth adjustment plan, query the current load status of the service nodes, and use the allocation algorithm to reallocate the service nodes to obtain the node allocation plan; Based on the bandwidth adjustment scheme and node allocation scheme, an optimized configuration including slice bandwidth and service node parameters is generated to obtain the optimized transmission configuration; By optimizing the transmission configuration, updating the bandwidth parameters and node allocation of the transmission channel, and using a verification algorithm to verify the availability of the configuration, an available transmission channel is obtained.

4. The system according to claim 1, characterized in that, The customized health guidance generation module includes: Time series data is obtained from the blood glucose change trend, and the time series data is processed using a data standardization algorithm to obtain standardized trend data; Based on the standardized trend data, the probability of the early warning signal is calculated using the early warning signal analysis method to obtain the early warning probability value; If the warning probability value is higher than the preset threshold, an intervention instruction generation algorithm is used to generate a specialist intervention instruction to obtain intervention instruction data. If the warning probability value is lower than a preset threshold, a suggestion generation algorithm is used to generate regular health suggestions, and health suggestion data is obtained. Based on the intervention instruction data or the health recommendation data, a health management plan is generated using a data integration algorithm to obtain health plan data. Personalized features are extracted from the health plan data, and a feature matching algorithm is used to generate customized health guidance, resulting in customized guidance data.

5. The system according to claim 1, characterized in that, The system further includes an update module, which includes: Patient feedback data and blood glucose data are obtained from the patient terminal. The patient feedback data is formatted and denoised using a data cleaning algorithm to obtain standardized feedback data. Based on the standardized feedback data and blood glucose data, a trend analysis algorithm is used to calculate the blood glucose change curve and generate blood glucose trend data. The blood glucose trend data and standardized feedback data are integrated into the patient's medical record using the record update rules to obtain the updated patient medical record. Based on the updated patient medical records, a mapping adjustment algorithm is used to reallocate the relationship between devices and service nodes, generating an adjustment mapping table. If the load of the service nodes in the adjusted mapping table is higher than a preset threshold, a load balancing algorithm is used to redistribute the nodes to obtain an optimized mapping table.

Images