Internal medicine emergency patient medical assistant system and method based on artificial intelligence

Abstract

The invention discloses an internal medicine emergency patient medical assistant system and method based on artificial intelligence, and relates to the technical field of medical treatment.The method comprises the steps that multiple sensor sets are arranged, and infusion physiological state data are collected in real time; carrying out data preprocessing and standardized normalization processing through an AI preprocessing strategy to form a standard infusion physiological state data set; inputting the standard infusion physiological state data set into the infusion time prediction model in the constructed infusion time prediction model, realizing high-precision prediction of the infusion completion time prediction value Tinf of the patient in combination with nonlinear attenuation modeling, and dividing the internal medicine emergency patients who are about to complete infusion through a preliminary comparison evaluation mechanism so as to obtain an infusion completion time prediction value Tinf; therefore, a patient group about to complete infusion is recognized in advance, accurate early warning and pre-planning of a liquid changing task are achieved, and the occurrence rate of adverse events such as infusion empty passing, blood return and vein blockage is remarkably reduced.

Description

Technical Field

[0001] The present invention relates to the field of medical technology, and particularly to an artificial intelligence-based medical assistant system and method for emergency internal medicine patients. Background Art

[0002] It relates to the field of artificial intelligence medical assistance, especially within the scope of research on intelligent medical scheduling systems, and more specifically to an artificial intelligence-based medical assistant method for emergency internal medicine patients. This method is mainly applied to the management of high-frequency nursing operation processes mainly based on intravenous infusion treatment in the emergency internal medicine scenario, especially in situations where there are problems of tight nursing resource scheduling due to concentrated patient numbers and numerous infusion tasks. Through this assistant method, intelligent collection of patients' physiological state data during the emergency internal medicine infusion process, dynamic prediction of the infusion process, reasonable allocation of nurses' tasks, and system-level load assessment can be achieved.

[0003] At the current stage, in the emergency internal medicine infusion area of the hospital, nurses need to frequently travel back and forth to the patient's bedside to perform tasks such as infusion rounds, changing infusion solutions, and handling abnormalities. Especially during peak hours, due to concentrated patient numbers and asynchronous infusion cycles, there is a problem of excessive local nursing load caused by some nurses being called by multiple patients in a short period of time. In addition, the existing infusion round scheduling usually relies on manual experience judgment and lacks real-time data support, making it impossible to effectively predict which patients will finish their infusions soon or whether there are potential risks, which is likely to lead to delays in changing solutions, running out of infusion, and even secondary medical risks such as venous reflux. This "task stacking" problem seriously affects the quality and efficiency of emergency nursing, and also increases the human pressure on nurses.

[0004] The emergence of the above phenomena mainly stems from the following three core deficiencies: First, there is a lack of real-time and quantifiable monitoring means for the current infusion state of patients; second, it is impossible to extract effective indicators from historical nursing behaviors to achieve dynamic assessment of nurses' tasks; third, the current nurses' task allocation lacks a system-wide perspective and does not establish an intelligent scheduling optimization mechanism. When these factors are superimposed, it often causes an instantaneous concentration of infusion task peaks during certain time periods, thereby inducing the following abnormal consequences: Some patients are not visited for a long time after the infusion is completed, which is likely to cause blood return or catheter blockage; some nurses are overloaded with work, resulting in increased operation fatigue and error rates; the overall system scheduling is unbalanced, manifested as local overload and global inefficiency. These consequences not only affect the continuity and safety of emergency treatment, but may also lead to medical disputes due to delayed handling, restricting the overall improvement of the hospital's service quality. Summary of the Invention

[0005] In view of the deficiencies of the prior art, the present invention provides an artificial intelligence-based medical assistant system and method for emergency internal medicine patients, which solves the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: including the following steps:

[0007] S1. By setting up a sensor group, the infusion physiological state data is collected in real time, and a wireless communication network is set up to transmit the infusion physiological state data to the AI scheduling server;

[0008] S2. Connect the nursing task platform to the AI scheduling server, extract the nursing behavior data in real time, and preprocess the nursing behavior data and the infusion physiological state data to obtain a standard nursing behavior data set and a standard infusion physiological state data set;

[0009] S3. Build an infusion time prediction model, extract the standard infusion physiological state data set and input it into the infusion time prediction model for calculation to output the predicted value Tinf of the infusion completion time, and set up a preliminary evaluation mechanism to trigger the scheduling optimization mechanism;

[0010] S4. After triggering the scheduling optimization mechanism, calculate and output the nursing scheduling interference function Hs based on the predicted value Tinf of the infusion completion time and the standard nursing behavior data set, set up a nursing threshold F2 for nursing evaluation, and execute the task shunting strategy based on the evaluation results;

[0011] S5. After the task shunting strategy is executed, calculate and output the global pressure index Osys based on the nursing scheduling interference function Hs of all current nurses, set up a pressure threshold Oth for secondary comparison and evaluation, and judge the scheduling load balance situation.

[0012] Preferably, the S1 includes S11 and S12;

[0013] S11. Install a sensor group on the medical emergency infusion patients and infusion devices to collect the infusion physiological state data of the medical emergency infusion patients in real time;

[0014] The sensor group includes a wristband pulse sensor, an infrared thermal imager, a drug physical database API, and an intelligent infusion pump sensor;

[0015] The infusion physiological state data includes a micro-pressure fluctuation curve, a blood flow area heat map sequence, a pharmacological feature data set, a total set volume Vset, and an injected volume Vinjected;

[0016] The micro-pressure fluctuation curve is obtained by wearing a wristband pulse sensor on the medical emergency infusion patient and attaching it near the patient's vein puncture area, and detecting the micro-pressure fluctuation of the local tissue every 5 seconds during the infusion process;

[0017] The blood flow area heat map sequence is collected by fusing near-infrared reflectance blood flow imaging NIRS with an infrared thermal imager;

[0018] The pharmacological characteristic data set automatically captures the drug code in the doctor's order through the drug physical database API, and extracts the pharmacological characteristic data set based on the drug code, wherein the pharmacological characteristic data set includes the molecular weight M of the drug used by the i-th patient, the solvent dynamic viscosity u of the drug used by the i-th patient, and the drug molecular formula of the drug used by the i-th patient;

[0019] The set volume Vset and the injected volume Vinjected are collected and acquired by the intelligent infusion pump sensor;

[0020] S12. All sensor groups’ self-provided network communication modules are connected to the intranet LoRa / MQTT gateway to wirelessly transmit the real-time collected infusion physiological status data to the constructed AI scheduling server.

[0021] Preferably, S2 includes S21 and S22;

[0022] S21. Set up a medical care platform API in the AI scheduling server, integrate the AI scheduling server with the nursing task platform, and extract nursing behavior data in real time;

[0023] The nursing behavior data includes the path complexity Lj between the i-th patient and the j-th nurse ij , the task operation time of the i-th patient and the j-th nurse Cz ij and the number of abnormal states of the i-th patient Yc i ;

[0024] The path complexity Lj between the i-th patient and the j-th nurse ij Through the RFID tags worn by nurses and UWB indoor positioning, the shortest time path between nurses and patients is calculated in combination with the path node graph algorithm;

[0025] The task operation time of the i-th patient and the j-th nurse is Cz ij The difference between the task start time and the completion feedback time is automatically recorded and then analyzed for regression analysis;

[0026] The number of abnormal states of the i-th patient Yc i The number of times the infusion blockage, reflux and alarm conditions are detected by the infusion pump is obtained;

[0027] S22. Construct an AI preprocessing strategy in the AI scheduling server, and preprocess the acquired nursing behavior data and infusion physiological status data based on the AI preprocessing strategy to obtain a standard nursing behavior data set and a standard infusion physiological status data set, respectively;

[0028] The AI preprocessing strategy includes filtering, image preprocessing, graph neural network, machine algorithm and AI-driven normalization processing;

[0029] The standard infusion physiological state data set includes the local vein permeability index A of the i-th patient i , the end-vein return compression factor Ys of the i-th patient i , the drug molecule diffusion retardation constant P of the i-th patient i and the current remaining infusion volume V of the i-th patient i ;

[0030] The standard nursing behavior data set includes the path complexity Lj between the i-th patient and the j-th nurse after normalization ij , the task operation time Cz of the i-th patient and the j-th nurse ij and the number of abnormal status times Yc of the i-th patient i .

[0031] Preferably, the S3 includes S31 and S32;

[0032] S31. Construct an infusion time prediction model through a machine algorithm, extract the standard infusion physiological state data set and input it into the infusion time prediction model, and calculate and output the predicted value Tinf of the infusion completion time of each internal medicine emergency patient to predict the infusion completion time of each internal medicine emergency patient;

[0033] The predicted value Tinf of the infusion completion time is calculated and output through the following infusion time prediction model;

[0034] ;

[0035] In the formula, Tinf i The predicted value of the infusion completion time of the i-th patient, represents a minimum constant, with a value of 1×10 -3 , and e represents an exponential function.

[0036] Preferably, S32. After obtaining the predicted value Tinf of the infusion completion time of each patient, a preliminary evaluation mechanism is executed. The preliminary evaluation mechanism sets the infusion completion warning time threshold △Tth by the user, and at the same time extracts the current system time tnow and the predicted value Tinf of the infusion completion time of each patient for preliminary comparison and evaluation, and aggregates the internal medicine emergency infusion patients whose infusions are about to be completed to construct an infusion completion set Ct. After the infusion completion set Ct exceeds the resource upper limit threshold Nth preset by the hospital resources, a scheduling optimization mechanism is triggered;

[0037] The infusion completion set Ct is constructed in the following manner;

[0038] ;

[0039] When the collection Ct of completed infusions exceeds the resource upper limit threshold Nth, the scheduling optimization mechanism is triggered;

[0040] When the collection Ct of completed infusions is less than or equal to the resource upper limit threshold Nth, it is not triggered.

[0041] Preferably, S4 includes S41 and S42;

[0042] S41. After the preliminary evaluation mechanism triggers the scheduling optimization mechanism, the predicted infusion completion time Tinf of the patients in the collection Ct of completed infusions and the standard nursing behavior data set are extracted, and the nursing scheduling interference function Hs is calculated and output through summary calculation, and the pressure degree of the tasks to be completed is analyzed;

[0043] The nursing scheduling interference function Hs is calculated and output through the following algorithm formula;

[0044] ;

[0045] In the formula, Hs j represents the nursing scheduling interference function of the j-th nurse, represents the minimum constant, with a value of 1×10 -3 .

[0046] Preferably, S42. Every 60 seconds, the nursing scheduling interference functions Hs of all nurses are traversed, and the actual average response time of each operation task of each nurse is extracted from the nursing records, the critical moment of nurse task overload is marked, the nursing scheduling interference function Hs of the current nurse is back-projected by the random forest of the AI model, and the 85th percentile is taken as the nursing threshold F2, and the nursing scheduling interference functions Hs of all nurses obtained in real time are evaluated with the nursing threshold F2, and the task diversion strategy is triggered based on the evaluation results. The specific evaluation content is as follows;

[0047] When the nursing scheduling interference function Hs of the j-th nurse j > the nursing threshold F2, it means that the nurse task load is overloaded, and at this time, the task diversion strategy is triggered;

[0048] When the nursing scheduling interference function Hs of the j-th nurse j ≤ the nursing threshold F2, it means that the nurse task load is normal, and at this time, priority is given to allocation;

[0049] The task diversion strategy finds the nurses with overloaded task loads that meet the nursing scheduling interference function Hs of the j-th nurse j > the nursing threshold F2 by traversing all nurses j, and searches for the minimization allocation function Min(L ij + Cz ij ) in the collection Ct of completed infusions, and transfers the current task to the nurse j with the shortest path and the fastest task operation time first;

[0050] Among them, the minimization allocation function Min(L ij + Cz ij ) represents the path complexity Lj between the ith patient and the jth nurse ij and the task operation time Cz between the ith patient and the jth nurse ij both take the minimum value, that is, the nurse j with the shortest path and the fastest operation is assigned to the ith patient in the infusion completion set Ct.

[0051] Preferably, the S5 includes S51 and S52;

[0052] S51. After the task shunting strategy is executed, recalculate the nursing scheduling interference function Hs of all nurses, calculate the standard deviation based on the nursing scheduling interference function Hs of all nurses, output the global pressure index Osys, and conduct a global analysis of the task load of all nurses;

[0053] The global pressure index Osys is calculated and output through the following algorithm formula;

[0054] ;

[0055] In the formula, n represents the total number of nurses participating in the scheduling, represents the nursing scheduling interference function of all nurses.

[0056] Preferably, S52. Collect the nurse scheduling records of the past 30 days, mark the normal distribution and abnormal distribution, conduct a statistical analysis of the global pressure index Osys in these states, set the critical point where abnormalities start to occur frequently as the pressure threshold Oth, and conduct a secondary comparison and evaluation of the globally obtained global pressure index Osys and the pressure threshold Oth to judge the overall scheduling load balance situation. The specific evaluation content is as follows;

[0057] When the global pressure index Osys < the pressure threshold Oth, it means that the current scheduling is effective, and the current scheduling is maintained;

[0058] When the global pressure index Osys ≥ the pressure threshold Oth, it means that the global scheduling is unbalanced. At this time, based on the current scheduling, continue to execute the task shunting strategy until the scheduling is effective and stop the execution.

[0059] The medical assistant system for internal medicine emergency patients based on artificial intelligence includes an infusion perception module, an AI preprocessing module, an infusion time prediction module, a scheduling intervention module, and a comprehensive evaluation module;

[0060] The infusion perception module sets up a sensor group to collect infusion physiological state data in real time, and sets up a wireless communication network to transmit the infusion physiological state data to the AI scheduling server;

[0061] The AI preprocessing module accesses the nursing task platform in the AI scheduling server, extracts nursing behavior data in real time, and preprocesses the nursing behavior data and infusion physiological state data to obtain a standard nursing behavior data set and a standard infusion physiological state data set;

[0062] The infusion time prediction module constructs an infusion time prediction model, extracts the standard infusion physiological state data set and inputs it into the infusion time prediction model for calculation to output the predicted value Tinf of the infusion completion time, and sets a preliminary evaluation mechanism to trigger the scheduling optimization mechanism;

[0063] After triggering the scheduling optimization mechanism, the scheduling intervention module calculates and outputs a nursing scheduling interference function Hs based on the predicted value Tinf of the infusion completion time and the standard nursing behavior data set, sets a nursing threshold F2 for nursing evaluation, and executes a task diversion strategy based on the evaluation result;

[0064] After the task diversion strategy is executed, the comprehensive evaluation module calculates and outputs a global pressure index Osys based on the nursing scheduling interference function Hs of all current nurses, sets a pressure threshold Oth for secondary comparison evaluation, and judges the scheduling load balance situation.

[0065] The present invention provides an internal medicine emergency patient medical assistant system and method based on artificial intelligence. It has the following beneficial effects:

[0066] (1) By setting a variety of sensor groups including a wristband pulse sensor, an infrared thermal imager, a drug physical database API, and an intelligent infusion pump sensor in step S1, the method collects infusion physiological state data including a micro-pressure fluctuation curve, a blood flow area thermal map sequence, a pharmacological feature data set, and an infusion volume; and in step S2, through an AI preprocessing strategy for standardization and normalization processing, a standard infusion physiological state data set is formed; in step S3, the constructed infusion time prediction model uses the local vein permeability index A, the terminal vein return compression factor Ys, the drug molecule diffusion lag constant P, and the current remaining infusion volume V, and combines non-linear attenuation modeling to achieve high-precision prediction of the patient's Tinf infusion completion time, thereby identifying in advance the group of patients whose infusion is about to be completed, realizing accurate early warning and pre-planning of the fluid replacement task, and significantly reducing the incidence of adverse events such as over-empty infusion, blood return, and venous catheter blockage.

[0067] (2)In step S4 of this method, after the preliminary evaluation mechanism triggers the scheduling optimization mechanism, based on the Tinf prediction results of the patients in the infusion completion set Ct and the standard nursing behavior data set, the nursing scheduling interference function Hs is calculated and generated as an evaluation parameter reflecting the current task pressure level of the j-th nurse. The nursing scheduling interference function Hs is compared and evaluated with the dynamically set nursing threshold F2. When the nursing scheduling interference function Hs > the nursing threshold F2, the task diversion strategy is triggered. By traversing the minimization allocation function Min(L ij + Cz ij ), the current task is reassigned to the nurse j with the shortest path and the lowest time consumption, realizing the dynamic balance of nursing resources without interrupting the original task process, improving the timeliness of the nurse's task response and the smoothness of the overall nursing operation, especially showing significant scheduling flexibility and improved nursing efficiency during the peak infusion period.

[0068] (3)In step S5 of this method, after the task diversion strategy is executed, the nursing scheduling interference function Hs of each nurse is recalculated, and the global pressure index Osys is calculated based on it, that is, the standard deviation of the task pressures of all nurses; then this index is compared and evaluated with the pressure threshold Oth obtained through training historical scheduling records to form a complete load balancing feedback closed loop. If the global pressure index Osys < the pressure threshold Oth, it means that the current scheduling resource configuration is reasonable and the nursing task distribution is balanced; if the global pressure index Osys ≥ the pressure threshold Oth, the task diversion strategy is automatically continued until the load difference is reduced. Through this strategy, the continuous tracking and optimal control of the global distribution state of nursing tasks are realized, effectively preventing problems such as delayed operations, service interruptions, or decreased nursing satisfaction caused by overloading of individual nurses' tasks, and significantly improving the stability and sustainability of emergency nursing scheduling. Description of the Drawings

[0069] Figure 1 Schematic diagram of the steps of the medical assistant method for internal medicine emergency patients based on artificial intelligence of the present invention;

[0070] Figure 2 Schematic diagram of the system process of the medical assistant for internal medicine emergency patients based on artificial intelligence of the present invention;

[0071] Figure 3 Curve graph of the change of the global pressure index Osys of the present invention. Detailed Embodiment

[0072] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0073] Example 1, please refer to Figure 1 and Figure 3 , the present invention provides a medical assistant method for internal medicine emergency patients based on artificial intelligence. To achieve the above objectives, the present invention is realized through the following technical solutions: including the following steps:

[0074] S1. By setting up a sensor group, the infusion physiological state data is collected in real time, and a wireless communication network is set to transmit the infusion physiological state data to the AI scheduling server;

[0075] S2. In the AI scheduling server, a nursing task platform is accessed, the nursing behavior data is extracted in real time, and the nursing behavior data and the infusion physiological state data are preprocessed to obtain a standard nursing behavior data set and a standard infusion physiological state data set;

[0076] S3. Build an infusion time prediction model, extract the standard infusion physiological state data set and input it into the infusion time prediction model, calculate and output the predicted value Tinf of the infusion completion time, and set up a preliminary evaluation mechanism to trigger the scheduling optimization mechanism;

[0077] S4. After triggering the scheduling optimization mechanism, calculate and output the nursing scheduling interference function Hs based on the predicted value Tinf of the infusion completion time and the standard nursing behavior data set, set a nursing threshold F2 for nursing evaluation, and execute a task diversion strategy based on the evaluation result;

[0078] S5. After the task diversion strategy is executed, calculate and output the global pressure index Osys based on the nursing scheduling interference function Hs of all current nurses, set a pressure threshold Oth for secondary comparison evaluation, and judge the scheduling load balance situation.

[0079] In this embodiment, the method collects infusion physiological state data by deploying multiple sensor groups, and transmits it to the AI scheduling server in real time via LoRa and MQTT wireless communication modules, realizing high-frequency and low-latency collection and uploading of data; subsequently, nurse behavior data is obtained by integrating with the nursing task platform, and data normalization, standardization, and structuring are performed based on the AI preprocessing strategy to generate a standard nursing behavior data set and a standard infusion physiological state data set, and a unified feature expression method is established. A nonlinear-structured infusion time prediction model is constructed using the standard infusion physiological state data set, and an individualized predicted value Tinf of the infusion completion time is output. Based on this prediction result, a preliminary evaluation mechanism is constructed. By comparing with a preset infusion completion warning time threshold ΔTth, a set Ct of high-risk patients whose infusions are about to be completed is identified in a timely manner; when Ct exceeds a preset load threshold Nth, a scheduling optimization mechanism is automatically triggered, and step S4 is entered. At this time, combining the predicted value Tinf of the infusion completion time with the standard nursing behavior data set, a nurse care scheduling interference function Hs is calculated, and a nurse care resource overload determination is made based on the nursing threshold F2 obtained by AI learning. A task diversion strategy is triggered for nurses exceeding the nursing threshold F2, and by optimizing the minimum path complexity Lj and the task operation time Cz, a low-pressure transfer and reassignment of patient tasks are realized. Finally, in step S5, the nurse care scheduling interference functions Hs of each nurse are re-aggregated, and the standard deviation is calculated and output as a global pressure index Osys, so as to analyze the balance of the overall task allocation, and compare it with a pressure threshold Oth obtained by training with historical data, forming a global-level dynamic scheduling optimization feedback mechanism to ensure that the system scheduling achieves an optimal balance between efficiency and fairness. In summary, the present invention realizes the whole-process closed-loop control from infusion state perception, prediction, evaluation to dynamic task optimization by introducing a multi-dimensional perception and AI-driven decision-making mechanism, not only improving the scheduling efficiency and response timeliness of the internal medicine emergency nursing tasks, but also significantly reducing infusion abnormalities and nursing delays caused by task stacking and uneven resource allocation, and improving the intelligent level of the overall nursing process and the stability and safety of medical services.

[0080] Example 2, please refer to Figure 1 , specifically: S1 includes S11 and S12;

[0081] S11. Install a sensor group on the internal medicine emergency infusion patients and infusion devices to collect the infusion physiological state data of the internal medicine emergency infusion patients in real time;

[0082] The sensor group includes a wristband pulse sensor, an infrared thermal imager, a drug physical database API, and an intelligent infusion pump sensor;

[0083] The infusion physiological status data include micro-pressure fluctuation curve, blood flow area heat map sequence, pharmacological characteristic data set, total set volume Vset and injected volume Vinjected;

[0084] The micro-pressure fluctuation curve is obtained by having the internal medicine emergency infusion patient wear a wristband pulse sensor, which is attached near the patient's venous puncture area, and detecting the micro-pressure fluctuation of the local tissue every 5 seconds during the infusion process;

[0085] The blood flow area heat map sequence is acquired by infrared thermal imager fused with near-infrared reflective blood flow imaging NIRS;

[0086] The pharmacological characteristic data set automatically captures the drug code in the doctor's order through the drug physical database API, and extracts the pharmacological characteristic data set based on the drug code. The pharmacological characteristic data set includes the molecular weight M of the drug used by the i-th patient, the solvent dynamic viscosity u of the drug used by the i-th patient, and the drug molecular formula of the drug used by the i-th patient;

[0087] The set volume Vset and the injected volume Vinjected are collected and obtained through the intelligent infusion pump sensor;

[0088] S12. All sensor groups’ self-provided network communication modules are connected to the intranet LoRa / MQTT gateway to wirelessly transmit the real-time collected infusion physiological status data to the constructed AI scheduling server.

[0089] In this embodiment, the method constructs a complete infusion physiological state data perception and transmission mechanism through step S1. Among them, in S11, a variety of heterogeneous sensor groups are deployed on the infusion patients in the internal medicine emergency department and the infusion equipment to realize multi-dimensional collection of the patient's individual physiological state and pharmacological information. Specifically, the wristband pulse sensor is attached near the venous puncture area to collect the micro-pressure fluctuation curve at a high frequency of 5 seconds / time, effectively reflecting the local venous permeability and elastic changes during the infusion process; the infrared thermal imager combines the near-infrared reflectance blood flow imaging NIRS technology to collect a sequence of thermal maps of the blood flow area, realizing the visual acquisition of the blood flow dynamics and heat diffusion state in the infusion area; the drug physical database API docks with the hospital pharmacy system to extract the molecular weight M, solvent dynamic viscosity u and drug molecular formula composition of the corresponding drugs for each patient in real time, forming a high-dimensional pharmacological feature data set; in addition, the intelligent infusion pump sensor obtains the set volume Vset and the injected volume Vinjected in real time, providing key parameter support for subsequent infusion volume modeling. Subsequently, in S12, using the communication module preset in each sensor group, the multi-dimensional infusion physiological state data is stably transmitted to the constructed AI scheduling server in real time by means of the low-power wide-area network LoRa and the lightweight message transmission protocol MQTT wirelessly. This design not only ensures the continuity and robustness of data transmission in the emergency complex environment, but also effectively reduces data latency and system load through the edge-deployed network structure.

[0090] Example 3, please refer to Figure 1 , specifically: S2 includes S21 and S22;

[0091] S21. Set up the medical staff platform API in the AI scheduling server, integrate and connect the AI scheduling server with the nursing task platform, and extract nursing behavior data in real time;

[0092] The nursing behavior data includes the path complexity Lj between the ith patient and the jth nurse ij , the task operation time Cz of the ith patient and the jth nurse ij and the number of abnormal state times Yc of the ith patient i ;

[0093] The path complexity Lj between the ith patient and the jth nurse ij is obtained by calculating the shortest time path between the nurse and the patient through the RFID nurse-worn tag and UWB indoor positioning, combined with the path node map algorithm;

[0094] The task operation time Cz of the ith patient and the jth nurse ij is obtained by performing a regression analysis on the difference between the automatically recorded task start time and the completion feedback time;

[0095] The number of abnormal state times Yc of the ith patienti Obtained by detecting the number of times of infusion blockage, backflow, and alarm through an infusion pump;

[0096] S22. Construct an AI preprocessing strategy in the AI scheduling server, and preprocess the obtained nursing behavior data and infusion physiological state data based on the AI preprocessing strategy to obtain a standard nursing behavior data set and a standard infusion physiological state data set respectively;

[0097] The AI preprocessing strategy includes filtering processing, image preprocessing, graph neural network, machine algorithm, and AI-driven normalization processing;

[0098] The standard infusion physiological state data set includes the local vein permeability index A of the i-th patient i , the end vein backflow compression factor Ys of the i-th patient i , the drug molecule diffusion retardation constant P of the i-th patient i and the current remaining infusion volume V of the i-th patient i ;

[0099] The standard nursing behavior data set includes the path complexity Lj between the i-th patient and the j-th nurse after normalization processing ij , the task operation time Cz of the i-th patient and the j-th nurse ij and the number of abnormal states Yc of the i-th patient i

[0100] The local vein permeability index A of the i-th patient i Perform low-pass filtering on the micro-pressure fluctuation curve through the filtering process of the AI analysis strategy to remove high-frequency jitter, perform wavelet decomposition on the waveform of the filter, extract the low-frequency rebound response curve, extract the molecular recovery time, pressure rise slope, and relative amplitude based on the low-frequency rebound response curve, and after eliminating the dimension through AI-driven normalization processing, use the machine algorithm to construct a non-linear polynomial fitting model to obtain the local vein permeability index A of the i-th patient by weighted fitting of the extracted molecular recovery time, pressure rise slope, and relative amplitude. i The larger the value of A of the i-th patient, the better the blood vessel elasticity and the lower the infusion resistance. Among them, the weights in the formula are obtained through training by using the AI analysis to label the true blood vessel rebound level.

[0101] The end vein backflow compression factor Ys of the i-th patient i Through the image preprocessing in the AI analysis strategy, perform background difference on the sequence of blood flow area heat maps based on sparse principal component analysis SPCA, extract the regional temperature gradient ▽T, temperature texture change bHT, and infrared region conduction delay time △tther, and after eliminating the dimension through AI-driven normalization processing, use the Bayesian network to establish the relationship between the temperature change pattern and the blockage degree, and output to obtain the blockage factor in the sense of probability;

[0102] The drug solution molecular diffusion retardation constant P of the i-th patient i Through the obtained pharmacological feature dataset, use a graph neural network to extract features of the drug molecular formula used by the i-th patient, predict the drug diffusion coefficient D of the drug used by the i-th patient and the molecular structure polarity index C of the drug used by the i-th patient. Combine the pharmacological feature dataset, use AI-driven normalization processing, eliminate the dimension, and then use a machine algorithm to construct a ratio calculation formula to calculate and obtain. The specific calculation formula is: ;

[0103] The current remaining infusion volume V of the i-th patient i Obtained by constructing a difference model of the total set volume Vset and the injected volume Vinjected using a machine algorithm and performing difference calculation output;

[0104] AI-driven normalization processing constructs an Autoencoder-style normalization network, inputs the nursing behavior data and infusion physiological state data into the Autoencoder-style normalization network, performs preliminary normalization processing, and performs secondary normalization processing after outputting the standard nursing behavior dataset and the standard infusion physiological state dataset to eliminate the dimensional influence of all parameters.

[0105] In this embodiment, in step S21 of the method, by setting a medical care platform API in the AI scheduling server, three types of core data closely related to patient interaction in the nursing task platform are extracted in real time: path complexity, task operation time consumption, and patient abnormal state records. Among them, the path complexity Lj is obtained by combining the RFID nurse-worn tag with the UWB indoor high-precision positioning system and the path node map algorithm, realizing the quantification of the path cost between the nurse and the patient; the task operation time consumption Cz extracts the objective time consumption of each nursing task through an automated time recording and regression analysis mechanism; and the number of abnormal states Yc is detected by the infusion pump for infusion abnormal events in real time, realizing a high-sensitivity early warning mechanism. This process effectively establishes a digital model between the "nursing resource utilization efficiency" and the "task risk factor". Subsequently, in step S22, the present invention uses an AI preprocessing strategy to fuse and standardize the collected nursing behavior data and infusion physiological state data. Specifically, the AI preprocessing strategy combines multiple intelligent computing modules such as filtering processing, wavelet decomposition, image preprocessing, graph neural network, and Autoencoder-based normalization network to realize the extraction, cleaning, and normalization of multi-dimensional features of the data. In terms of processing physiological state data, by filtering and wavelet decomposing the micro-pressure fluctuation curve, key parameters are extracted and then non-linearly fitted to generate the local venous permeability index A; by processing the blood flow heat map through sparse principal component analysis SPCA and Bayesian network, the regional heat response is extracted, and the terminal venous return compression factor Ys is calculated; by analyzing the drug molecular structure and pharmacological properties through a graph neural network, the drug solution molecular diffusion retardation constant P is calculated; a difference model is constructed by combining the set volume and the injected volume to obtain the current remaining infusion volume V. In terms of nursing behavior data, the scale is unified through AI-driven normalization processing, and a standard nursing behavior data set is output. In summary, step S2 constructs a multi-source heterogeneous data fusion and standardization processing mechanism with AI analysis as the core, which not only realizes the structured expression of traditional nursing behaviors and patient physiological data, but also provides a unified, computable, and reliable input data basis for subsequent prediction models and scheduling optimization algorithms, significantly improving the intelligent level and overall operation efficiency of the system, while enhancing the dynamic perception and accurate response capabilities for nursing resource allocation and infusion status early warning.

[0106] Example 4, please refer to Figure 1 , specifically: S3 includes S31 and S32;

[0107] S31. Construct an infusion time prediction model through a machine algorithm, extract the standard infusion physiological state data set and input it into the infusion time prediction model for calculation, and output the predicted value Tinf of the infusion completion time for each internal medicine emergency patient to predict the infusion completion time of each internal medicine emergency patient;

[0108] The predicted infusion completion time value Tinf is calculated and output through the following infusion time prediction model;

[0109] ;

[0110] In the formula, Tinf i The predicted value of the infusion completion time for the i-th patient, represents the minimum constant value, with a value of 1×10 -3 , and e represents the exponential function;

[0111] Among them: represents the venous condition correction coefficient, which is a coefficient that simulates the influence of the individual venous state of the patient on the actual flow rate. The larger the value, the smoother the infusion and the shorter the time. Mathematical meaning: If the terminal venous return compression factor Ys i of the i-th patient approaches 0, it means the passage is very smooth; if it approaches 1, it indicates the existence of peripheral resistance. Physical meaning: This ratio controls the "difficulty of entering a unit of liquid" and is used to correct the ideal flow rate;

[0112] represents the drug - blood vessel interaction composite index. This index term simulates the decrease in flow rate caused by the movement, diffusion, and retardation of the drug in the microcirculation system. The smaller the index term, the more serious the "flow rate limitation";

[0113] The overall denominator of the calculation logic of the formula represents the "correction of the amount of liquid injected per unit time": Traditionally, only the drip rate is considered, while here the blood vessel state and drug characteristics are considered. The index term is a non - linear control term, introducing a dynamic simulation of the diffusion retardation phenomenon. For example, the drip rate is fast at the front of the drug and slow at the back;

[0114] Specific example: Assume V i = 100, A i = 0.9, Ys i = 0.3, P i = 2.0;

[0115] , although it is set to finish dripping per hour, due to blood vessel and drug reasons, the system is expected to take 74 minutes.

[0116] S32. After obtaining the predicted infusion completion time value Tinf for each patient, a preliminary evaluation mechanism is executed. The preliminary evaluation mechanism sets the infusion completion warning time threshold △Tth by the user, and at the same time extracts the current system time tnow and the predicted infusion completion time value Tinf for each patient for preliminary comparison and evaluation, and summarizes the medical emergency infusion patients whose infusions are about to be completed to construct the infusion completion set Ct. After the infusion completion set Ct exceeds the resource upper - limit threshold Nth preset by the hospital resources, the scheduling optimization mechanism is triggered;

[0117] The infusion completion set Ct is constructed in the following way;

[0118] ;

[0119] If the infusion completion set Ct > the resource upper limit threshold Nth, the scheduling optimization mechanism is triggered;

[0120] If the infusion completion set Ct ≤ the resource upper limit threshold Nth, it is not triggered;

[0121] The significance of the infusion completion set Ct is that by traversing all patients in the infusion and calculating the difference between the time when the patient completes the infusion and the current system time tnow, it is the remaining infusion time of the infusion patient. The infusion completion warning time threshold △Tth is set by the user, such as 5 minutes or 3 minutes. If the remaining infusion time of the infusion patient is lower than this time range, it means that the patient's infusion is about to be completed. At the same time, the infusion completion set Ct is evaluated with the resource upper limit threshold Nth to analyze how many patients are about to complete the infusion at the current moment, and the scheduling optimization mechanism is triggered.

[0122] In this embodiment, the method constructs an infusion time prediction and warning evaluation mechanism, and establishes an intelligent infusion task recognition and scheduling pre-perception ability driven by AI. Specifically, in S31, by extracting multiple key parameters from the standard infusion physiological state dataset, including the current remaining infusion volume V of the patient, the local vein permeability index A, the terminal vein return compression factor Ys, and the drug molecule diffusion retardation constant P, and inputting them into the infusion time prediction model constructed by the machine algorithm, the infusion completion time Tinf of each patient is accurately predicted. The mathematical model adopted is jointly constructed based on the venous correction coefficient and the diffusion lag factor in the form of a non-linear exponent, fully simulating the combined effects of vascular state and drug properties on the flow rate change, significantly superior to the traditional static estimation method based on flow rate or drip rate, and can dynamically feedback real physiological effects such as late-stage deceleration of the drug solution and vascular blockage, so as to achieve high-precision prediction of the infusion duration. Further in S32, based on the difference between the predicted infusion completion time Tinf of each patient and the current system time tnow, the infusion completion set Ct is constructed, and a dual comparison evaluation is carried out with the infusion completion warning time threshold △Tth and the resource upper limit threshold Nth set by the hospital. When the number of patients in the infusion completion set Ct exceeds Nth, it means that a large number of infusion end events will occur concentrated in a short time, and the system immediately triggers the subsequent scheduling optimization mechanism to achieve early identification and response to peak loads. The core value of this mechanism is to convert the passive waiting nursing task into an active perception scheduling guidance, greatly reducing the nursing resource conflict caused by the concentrated arrival of tasks.

[0123] Example 5, please refer to Figure 1, specifically: S4 includes S41 and S42;

[0124] S41. After the preliminary evaluation mechanism triggers the scheduling optimization mechanism, the predicted infusion completion time Tinf of patients in the infusion completion set Ct and the standard nursing behavior data set are extracted, and the nursing scheduling interference function Hs is calculated and output through summary calculation to analyze the pressure degree of the tasks to be completed;

[0125] The nursing scheduling interference function Hs is calculated and output through the following algorithm formula;

[0126] ;

[0127] In the formula, Hs j represents the nursing scheduling interference function of the j-th nurse, represents the minimum constant, with a value of 1×10 -3 ;

[0128] represents the task urgency factor of the i-th patient. The larger it is, the more urgent and important the task of this patient is, which comes from the comprehensive analysis of the expected infusion end and abnormal status of the patient;

[0129] The denominator represents the "unit cost" for the nurse to complete this task, that is, the path cost and the cost of the average processing time;

[0130] The ratio calculation is to analyze the weight of urgent tasks under the unit action cost. Urgent tasks, short paths, and fast operations, that is, a large ratio, which means high pressure.

[0131] S42. Traverse the nursing scheduling interference function Hs of all nurses every 60 seconds, extract the actual average response time of each operation task of each nurse from the nursing records, mark the critical moment of nurse task overload, use the AI model random forest to back-calculate the nursing scheduling interference function Hs of the current nurse, and take the 85th percentile as the nursing threshold F2, and conduct nursing evaluation on the nursing scheduling interference function Hs of all nurses obtained in real time and the nursing threshold F2, and trigger the task shunting strategy based on the evaluation results. The specific evaluation content is as follows;

[0132] When the nursing scheduling interference function Hs of the j-th nurse j > the nursing threshold F2, it means that the nurse's task load is overloaded, and at this time, the task shunting strategy is triggered;

[0133] When the nursing scheduling interference function Hs of the j-th nurse j ≤ the nursing threshold F2, it means that the nurse's task load is normal, and at this time, priority is given to allocation;

[0134] The task shunting strategy finds all nurses j that satisfy the nursing scheduling interference function Hs of the j-th nurse j>Nurses with a task load exceeding the care threshold F2 search for the minimization allocation function Min(L ij + Cz ij ) in the collection Ct of completed infusions, and transfer the current task with priority to the nurse j with the shortest path and the fastest task operation time;

[0135] Among them, the minimization allocation function Min(L ij + Cz ij ) represents the path complexity Lj between the i-th patient and the j-th nurse ij and the task operation time Cz between the i-th patient and the j-th nurse ij both take the minimum value, that is, the nurse j with the shortest path and the fastest operation is assigned to the i-th patient in the collection Ct of completed infusions.

[0136] In this embodiment, the method extracts, in S41, the predicted infusion completion time Tinf of all patients whose infusions are about to be completed in the collection Ct of completed infusions and their corresponding standard care behavior datasets by the AI scheduling server, and constructs a care scheduling interference function Hs. This function calculates the ratio of the task urgency factor of each patient, covering infusion end prediction, abnormal status, etc. to the current path complexity and task processing time of the nurse, so as to realize the quantitative evaluation of the task pressure borne by the nurse under the unit work cost. The care scheduling interference function Hs constructed here not only considers the urgency of the task itself, but also integrates the differences in the nurse's task execution path and efficiency, making the evaluation result more realistically reflective. In S42, every 60 seconds, the care scheduling interference function Hs of all nurses is cyclically traversed, and combined with an AI model, such as a random forest, the task response time in the historical care records is back-predicted and modeled, and the 85th percentile is dynamically extracted as the care threshold F2. By comparing the care scheduling interference function Hs obtained in real time with the care threshold F2, it can be determined whether the current nurse has a task overload. If the care scheduling interference function Hs exceeds the care threshold F2. It means that the task pressure of this nurse is too high, and the system immediately starts the task shunting strategy. The task shunting logic traverses all current nurses, and based on the weighted value of the minimum path complexity Lj and the task operation time Cz, re-selects the nurse with the shortest path and the fastest task processing to take over the task, ensuring that the task flows to the optimal execution path.

[0137] Example 6, please refer to Figure 1 and Figure 3 , specifically: S5 includes S51 and S52;

[0138] S51. After the task shunting strategy is executed, recalculate the care scheduling interference function Hs of all nurses, and calculate the standard deviation based on the care scheduling interference function Hs of all nurses, output the global pressure index Osys, and conduct a global analysis of the task load of all nurses;

[0139] The global pressure index Osys is calculated and output through the following algorithm formula;

[0140] ;

[0141] In the formula, n represents the total number of nurses participating in the scheduling, represents the nursing scheduling interference function of all nurses, that is, the average value of the task pressure of all nurses;

[0142] The overall calculation logic of the formula is the square of a standard deviation, which is used to describe the distribution dispersion degree of the task loads of all nurses after the task shunting strategy is executed.

[0143] S52. Collect the nurse scheduling records of the past 30 days, mark the normal distribution and abnormal distribution, such as whether there are delays, nurse complaints, etc., conduct statistical analysis on the global pressure index Osys in these states, and set the critical point where abnormalities start to occur frequently as the pressure threshold Oth. Then, conduct a secondary comparison and evaluation of the real-time obtained global pressure index Osys and the pressure threshold Oth to judge the overall scheduling load balance situation. The specific evaluation content is as follows;

[0144] When the global pressure index Osys < the pressure threshold Oth, it means that the current scheduling is effective, and the current scheduling is maintained;

[0145] When the global pressure index Osys ≥ the pressure threshold Oth, it means that the global scheduling is unbalanced. At this time, based on the current scheduling, continue to execute the task shunting strategy until the scheduling is effective and stop executing;

[0146] Specific example: The nursing scheduling interference function Hs of nurse A is 2.1, that of nurse B is 0.9, and that of nurse C is 1.6;

[0147] = (2.1 + 0.9 + 1.6) / 3 = 1.53;

[0148] Osys = 1 / 3[(2.1 - 1.53) 2 + (0.9 - 1.53) 2 + (1.6 - 1.53) 2 = 0.242;

[0149] If Oth = 0.2, then the global pressure index Osys ≥ the pressure threshold Oth, indicating that the tasks are unbalanced, and continue to execute the task shunting strategy.

[0150] In this embodiment, after completing the task shunting strategy, the method realizes the macroscopic analysis and closed-loop regulation of the overall nursing resource scheduling status by constructing a global pressure assessment mechanism. In the specific implementation, in S51, the latest nursing scheduling interference function Hs of all nurses is first recalculated, and standard deviation analysis is performed based on these values to output the global pressure index Osys. This index reflects the distribution dispersion degree of the current nurses' task pressure. The larger the standard deviation, the more unbalanced the task distribution, and the more likely it is to occur local nursing resource overload or scheduling imbalance. The global pressure index Osys, as a quantitative manifestation of the scheduling health, provides a decision-making basis for the subsequent scheduling strategy. In S52, to scientifically set the judgment criteria, a historical data comparison mechanism is introduced. The normal and abnormal distributions in the nurses' scheduling records in the past 30 days, such as delays, complaints, etc., are collected and analyzed. The critical point where the global pressure index Osys frequently appears abnormally is determined through data backtracking, and this point is set as the pressure threshold Oth. The system continuously performs a secondary comparison and evaluation on the current global pressure index Osys and the pressure threshold Oth: if the global pressure index Osys < the pressure threshold Oth, it means that the current task scheduling has tended to be reasonable and balanced, and the system can stop shunting and maintain the status quo; if the global pressure index Osys ≥ the pressure threshold Oth, it indicates that there is still a relatively strong scheduling imbalance, and the task shunting should continue until the overall load of the system stably drops back to the safe range. The implementation of this strategy constructs a complete closed-loop from "local task pressure identification" to "overall scheduling status" to "dynamic strategy continuous iteration". It not only ensures the realization of local optimization in a short time, but also ensures the realization of continuous scheduling balance in the long-term operation. Through the precise comparison between the global pressure index Osys and the pressure threshold Oth, it can dynamically respond to sudden nursing peaks, effectively reduce the nurse load difference, reduce operation errors and nursing delays, thereby significantly improving the intelligent management level and overall operation efficiency of emergency nursing services.

[0151] Embodiment 7, please refer to Figure 1 and Figure 2 , an internal medicine emergency patient medical assistant system based on artificial intelligence, including an infusion perception module, an AI preprocessing module, an infusion time prediction module, a scheduling intervention module, and a comprehensive evaluation module;

[0152] The infusion perception module collects infusion physiological state data in real time by setting a sensor group, and sets up a wireless communication network to transmit the infusion physiological state data to the AI scheduling server;

[0153] The AI preprocessing module accesses the nursing task platform in the AI scheduling server, extracts nursing behavior data in real time, and preprocesses the nursing behavior data and the infusion physiological state data to obtain a standard nursing behavior data set and a standard infusion physiological state data set;

[0154] The infusion time prediction module constructs an infusion time prediction model, extracts the standard infusion physiological state data set and inputs it into the infusion time prediction model for calculation to output the predicted infusion completion time Tinf, and sets up a preliminary evaluation mechanism to trigger the scheduling optimization mechanism;

[0155] The scheduling intervention module, after triggering the scheduling optimization mechanism, calculates and outputs the nursing scheduling interference function Hs based on the predicted infusion completion time Tinf and the standard nursing behavior data set, sets a nursing threshold F2 for nursing evaluation, and executes a task diversion strategy based on the evaluation results;

[0156] The comprehensive evaluation module, after the task diversion strategy is executed, calculates and outputs the global pressure index Osys based on the nursing scheduling interference function Hs of all current nurses, sets a pressure threshold Oth for secondary comparison and evaluation, and judges the scheduling load balance situation.

[0157] Although the embodiments of the present invention have been shown and described, those of ordinary skill in the art can understand that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the present invention.

Claims

1. A medical assistant method for emergency internal medicine patients based on artificial intelligence, characterized in that: The following steps are involved: S1. By setting up a sensor group, the infusion physiological status data is collected in real time, and a wireless communication network is set up to transmit the infusion physiological status data to the AI scheduling server; S2. Connect the nursing task platform to the AI scheduling server, extract nursing behavior data in real time, and pre-process the nursing behavior data and infusion physiological status data to obtain a standard nursing behavior data set and a standard infusion physiological status data set; S3. Construct an infusion time prediction model, extract a standard infusion physiological state data set and input it into the infusion time prediction model, calculate and output the infusion completion time prediction value Tinf, and set a preliminary evaluation mechanism to trigger the scheduling optimization mechanism; S4. After the scheduling optimization mechanism is triggered, the nursing scheduling interference function Hs is calculated and output based on the predicted value of the infusion completion time Tinf and the standard nursing behavior data set, and the nursing threshold F2 is set for nursing evaluation, and the task diversion strategy is executed based on the evaluation results; S5. After the task diversion strategy is executed, the global pressure index Osys is calculated and output based on the nursing scheduling interference function Hs of all current nurses, and the pressure threshold Oth is set for secondary comparative evaluation to determine the scheduling load balance situation.

2. The method for the medical assistant for internal medicine emergency patients based on artificial intelligence according to claim 1, wherein: Said S1 includes S11 and S12; S11. Install a sensor group on the emergency infusion patients and infusion equipment of the internal medicine department to collect the infusion physiological status data of the emergency infusion patients of the internal medicine department in real time; The sensor group includes a wristband pulse sensor, an infrared thermal imager, a drug physical database API, and a smart infusion pump sensor; The infusion physiological state data includes micro-pressure fluctuation curve, blood flow area heat map sequence, pharmacological characteristic data set, total set volume Vset and injected volume Vinjected; The micro-pressure fluctuation curve is obtained by having an internal medicine emergency infusion patient wear a wristband pulse sensor, attached near the patient's venous puncture area, and detecting the micro-pressure fluctuation of the local tissue every 5 seconds during the infusion process; The blood flow area heat map sequence is collected and acquired by an infrared thermal imager fused with near-infrared reflective blood flow imaging NIRS; The pharmacological characteristic data set automatically captures the drug code in the doctor's order through the drug physical database API, and extracts the pharmacological characteristic data set based on the drug code, wherein the pharmacological characteristic data set includes the molecular weight M of the drug used by the i-th patient, the solvent dynamic viscosity u of the drug used by the i-th patient, and the drug molecular formula of the drug used by the i-th patient; The set volume Vset and the injected volume Vinjected are collected and acquired by the intelligent infusion pump sensor; S12. All sensor groups’ self-provided network communication modules are connected to the intranet LoRa / MQTT gateway to wirelessly transmit the real-time collected infusion physiological status data to the constructed AI scheduling server.

3. The method for the medical assistant for internal medicine emergency patients based on artificial intelligence according to claim 2, wherein: The S2 includes S21 and S22; S21. Set up a medical care platform API in the AI scheduling server, integrate the AI scheduling server with the nursing task platform, and extract nursing behavior data in real time; The nursing behavior data includes the path complexity Lj between the i-th patient and the j-th nurse ij , the task operation time Cz of the i-th patient and the j-th nurse ij , and the number of abnormal statuses Yc of the i-th patient i ; The path complexity Lj between the i-th patient and the j-th nurse ij Obtained by calculating the shortest time path between the nurse and the patient through RFID nurse-worn tags and UWB indoor positioning, combined with the path node graph algorithm; The task operation time Cz between the i-th patient and the j-th nurse ij Obtained by performing regression analysis on the difference between the automatically recorded task start time and the completion feedback time; The number of abnormal states Yc of the i-th patient i is obtained by detecting the number of times of infusion blockage, backflow, and alarm through an infusion pump; S22. Construct an AI preprocessing strategy in the AI scheduling server, and preprocess the obtained nursing behavior data and infusion physiological state data based on the AI preprocessing strategy to obtain a standard nursing behavior data set and a standard infusion physiological state data set respectively; The AI preprocessing strategy includes filtering processing, image preprocessing, graph neural network, machine algorithm, and AI-driven normalization processing; The described standard infusion physiological state data set includes the local vein permeability index A of the i-th patient i , the end-vein return compression factor Ys of the i-th patient i , the drug molecule diffusion retardation constant P of the i-th patient i and the current remaining infusion volume V of the i-th patient i ; The standard nursing behavior dataset includes the path complexity Lj between the i-th patient and the j-th nurse after normalization processing ij , the task operation time Cz of the i-th patient and the j-th nurse ij and the number of abnormal states Yc of the i-th patient i .

4. The method for a medical assistant for internal medicine emergency patients based on artificial intelligence according to claim 3, wherein: S3 includes S31 and S32; S31. Construct an infusion time prediction model through a machine algorithm, extract the standard infusion physiological state data set and input it into the infusion time prediction model for calculation to output the predicted value Tinf of the infusion completion time for each internal medicine emergency patient, and predict the infusion completion time for each internal medicine emergency patient; The predicted value Tinf of the infusion completion time is calculated and output through the following infusion time prediction model; ； where, Tinf i predicted value of the infusion completion time of the i-th patient represents a minimum constant with a value of 1×10 -3 , and e represents the exponential function.

5. The method for a medical assistant for internal medicine emergency patients based on artificial intelligence according to claim 4, wherein: S32. After obtaining the predicted value Tinf of the infusion completion time for each patient, execute a preliminary evaluation mechanism. The preliminary evaluation mechanism sets an infusion completion warning time threshold △Tth by the user, and at the same time extracts the current system time tnow and the predicted value Tinf of the infusion completion time for each patient for preliminary comparison and evaluation, and summarizes the internal medicine emergency infusion patients whose infusions are about to be completed to construct an infusion completion set Ct. After the infusion completion set Ct exceeds the resource upper limit threshold Nth preset by the hospital resources, trigger a scheduling optimization mechanism; The infusion completion set Ct is constructed in the following way; ； If the infusion completion set Ct > the resource upper limit threshold Nth, trigger a scheduling optimization mechanism; If the infusion completion set Ct ≤ the resource upper limit threshold Nth, then do not trigger.

6. The method for the medical assistant for internal medicine emergency patients based on artificial intelligence according to claim 5, characterized in that: S4 includes S41 and S42; S41. After the preliminary evaluation mechanism triggers the scheduling optimization mechanism, extract the predicted value Tinf of the infusion completion time of the patients in the infusion completion set Ct and the standard nursing behavior data set, perform summary calculation to output a nursing scheduling interference function Hs, and analyze the pressure degree of the tasks to be completed; The nursing scheduling interference function Hs is calculated and output through the following algorithm formula; ； where, Hs j represents the nursing scheduling interference function of the j-th nurse, represents the minimum constant, with a value of 1×10 -3 .

7. The method for a medical assistant for internal medicine emergency patients based on artificial intelligence according to claim 6, wherein: S42. Traverse the nursing scheduling interference function Hs of all nurses every 60 seconds, extract the actual average response time of each operation task of each nurse from the nursing records, mark the critical moment of nurse task overload, use the AI model random forest to back-calculate the nursing scheduling interference function Hs of the current nurse, and take the 85th percentile as the nursing threshold F2, and perform a nursing evaluation on the nursing scheduling interference function Hs of all nurses obtained in real time and the nursing threshold F2, and trigger a task diversion strategy based on the evaluation result. The specific evaluation content is as follows; When the nursing scheduling interference function Hs of the j-th nurse j > the nursing threshold F2, it indicates that the nurse's task load is overloaded, and at this time, the task diversion strategy is triggered; When the nursing scheduling interference function Hs of the j-th nurse j ≤ the nursing threshold F2, it indicates that the nurse's task load is normal, and at this time, priority is given to allocation; The task diversion strategy traverses all nurses j to find the task load overloaded nurses who satisfy the nursing scheduling interference function Hs of the j-th nurse j > the nursing threshold F2, and searches for the minimization allocation function Min(L ij + Cz ij ) in the infusion completion set Ct, and preferentially transfers the current task to the nurse j with the shortest path and the fastest task operation time Among them, the minimization allocation function Min(L ij + Cz ij ) represents the path complexity Lj between the i-th patient and the j-th nurse ij and the task operation time Cz between the i-th patient and the j-th nurse ij both take the minimum value, that is, the nurse j with the shortest path and the fastest operation is assigned to the i-th patient in the infusion completion set Ct 8. The method for the medical assistant for emergency internal medicine patients based on artificial intelligence according to claim 6, wherein: S5 includes S51 and S52; S51. After the task diversion strategy is executed, recalculate the nursing scheduling interference function Hs of all nurses, perform a standard deviation calculation based on the nursing scheduling interference function Hs of all nurses, output a global pressure index Osys, and perform a global analysis of the task loads of all nurses; The global pressure index Osys is calculated and output through the following algorithm formula; ； Where n represents the total number of nurses participating in the scheduling, represents the nursing scheduling interference function of all nurses.

9. The method for a medical assistant for internal medicine emergency patients based on artificial intelligence according to claim 8, wherein: S52. Collect the nurse scheduling records of the past 30 days, mark the normal distribution and abnormal distribution, conduct statistical analysis on the global pressure index Osys in these states, set the critical point where abnormalities start to occur frequently as the pressure threshold Oth, and conduct a secondary comparison and evaluation between the globally obtained real-time pressure index Osys and the pressure threshold Oth to determine the overall scheduling load balance. The specific evaluation content is as follows; When the global pressure index Osys < the pressure threshold Oth, it indicates that the current scheduling is effective, and the current scheduling is maintained; When the global pressure index Osys ≥ the pressure threshold Oth, it indicates that the global scheduling is unbalanced. At this time, based on the current scheduling, continue to execute the task diversion strategy until the scheduling is effective and stop the execution.

10. An AI-based medical assistant system for emergency internal medicine patients, which is applied to the AI-based medical assistant method for emergency internal medicine patients according to any one of claims 1-9, and is characterized in that: It includes an infusion sensing module, an AI preprocessing module, an infusion time prediction module, a scheduling intervention module, and a comprehensive evaluation module; The infusion sensing module collects the infusion physiological state data in real time by setting up a sensor group, and sets up a wireless communication network to transmit the infusion physiological state data to the AI scheduling server; The AI preprocessing module accesses the nursing task platform in the AI scheduling server, extracts the nursing behavior data in real time, and preprocesses the nursing behavior data and the infusion physiological state data to obtain a standard nursing behavior data set and a standard infusion physiological state data set; The infusion time prediction module constructs an infusion time prediction model, extracts the standard infusion physiological state data set and inputs it into the infusion time prediction model, calculates and outputs the predicted infusion completion time Tinf, and sets up a preliminary evaluation mechanism to trigger the scheduling optimization mechanism; The scheduling intervention module, after triggering the scheduling optimization mechanism, calculates and outputs the nursing scheduling interference function Hs based on the predicted infusion completion time Tinf and the standard nursing behavior data set, sets a nursing threshold F2 for nursing evaluation, and executes the task diversion strategy based on the evaluation result; The comprehensive evaluation module, after the task diversion strategy is executed, calculates and outputs the global pressure index Osys based on the nursing scheduling interference function Hs of all current nurses, sets a pressure threshold Oth for secondary comparison and evaluation, and determines the scheduling load balance.

Images