Intelligent accompanying calling system and method based on AI model and storage medium

Abstract

The invention discloses an intelligent accompanying calling system and method based on an AI model, and a storage medium, relates to the technical field of artificial intelligence, and solves the problem that when a guardian needs to leave a hospital temporarily due to personal affairs, the guardian cannot flexibly arrange return time to consider the medical treatment demand of a patient, and the medical treatment efficiency is improved. And thus, the personal transaction processing efficiency of the guardian is reduced. The total time from leaving to returning to a hospital of a guardian is calculated according to guardian information; the patient registers through the intelligent wheelchair, and the predicted time for waiting for treatment is calculated; calling a time prediction model to predict the doctor-seeing time of the patient, and obtaining the doctor-seeing condition of the patient in the doctor-seeing process; judging whether the guardian can return to the hospital before the doctor-seeing time is over or not; classifying the doctor seeing conditions and calculating advanced calling time under different types of doctor seeing conditions; according to the invention, the guardian can be ensured to return to the hospital before the doctor-seeing time ends, and the patient can be ensured to get necessary care in the doctor-seeing process.

Description

Technical Field

[0001] The present invention belongs to the technical field of medical devices and relates to remote calling technology. Specifically, it is an intelligent escort calling system, method and storage medium based on an AI model. Background Art

[0002] Since when the guardian needs to leave temporarily due to personal affairs, it is impossible to flexibly arrange the return time to take into account the medical needs of the patient. With the rise of the escort industry and the emergence of various intelligent devices in the future, it can solve the trouble of patients needing to know the locations of various departments in the hospital and hospital procedures when seeing a doctor. However, new problems will arise from this situation. For example, after the patient completes procedures such as seeing a doctor under the accompaniment of an escort or an intelligent device, they often fall into a passive waiting state.

[0003] Currently, for the above situation, only the escort can guess the time when the consultation is completed and call the guardian's phone. When the intelligent device monitors the patient, it can only wait for the guardian to finish handling personal affairs and then pick up the patient. Once the guardian leaves the waiting area, they fall into an information blind spot. The guardian faces the binary opposition of "rigid medical time" and "flexible personal affairs", which directly leads to a decrease in the efficiency of the guardian's handling of personal affairs. At the same time, it may also bring unnecessary waiting anxiety and psychological burden to the patient.

[0004] Therefore, there is an urgent need for an intelligent escort calling system, method and storage medium based on an AI model. Summary of the Invention

[0005] The present invention aims to solve at least one of the technical problems existing in the prior art. For this purpose, the present invention provides an intelligent escort calling system, method and storage medium based on an AI model, which is used to solve the technical problem that when the guardian needs to leave the hospital temporarily due to personal affairs, it is impossible to flexibly arrange the return time to take into account the medical needs of the patient, resulting in a decrease in the efficiency of the guardian's handling of personal affairs.

[0006] To achieve the above object, the first aspect of the present invention provides an intelligent escort calling system based on an AI model, including: a data acquisition module, a time prediction module and a calling module;

[0007] Data acquisition module: Obtain the travel information of the guardian and the medical information of the patient in the intelligent wheelchair system; wherein, the medical information is the medical process input by the diagnosing doctor when the patient sees a doctor;

[0008] Time prediction module: Calculate the total time for the guardian to travel to and from the hospital according to the travel information;

[0009] The patient registers through the intelligent wheelchair, calculates the estimated waiting time for seeing a doctor based on the real-time queuing number of the department where the patient is registered; inputs the medical appointment information into the time prediction model to obtain the appointment time; among them, the time prediction model is constructed based on the artificial intelligence model.

[0010] Call module: Determine the early call time according to the time required for the guardian to travel to and from the hospital, the appointment time, and the estimated waiting time, and call the guardian according to the early call time.

[0011] Preferably, the intelligent escort call system based on the AI model further includes: a system interaction module, which is used to retrieve the guardian information of the intelligent wheelchair; and is used for the intelligent wheelchair to retrieve the real-time queuing number of each medical appointment item in the hospital appointment system and the average medical appointment time of the corresponding department; among them, the patient information includes: medical record information, the department where the patient sees a doctor, and the time for the patient to pick up medicine and have an infusion.

[0012] Preferably, calculating the total time for the guardian to leave and return to the hospital according to the guardian information includes:

[0013] S1: Extract the travel information of the guardian; among them, the guardian information includes the location of the guardian, the estimated processing time, and the destination of the guardian; judge whether the intelligent wheelchair can detect the Bluetooth signal of the guardian's mobile phone within the preset time threshold; if yes, it means the guardian has not left the hospital, and mark the call time as 0; if not, obtain the GPS coordinates of the guardian's mobile phone, and judge whether the GPS coordinates exceed the hospital electronic fence; if yes, mark the guardian as having left the hospital and jump to S2; if not, mark the guardian as not having left the hospital and mark the call time as 0.

[0014] S2: Obtain the round-trip time of the guardian through the map API.

[0015] S3: Obtain all feasible paths between the guardian and the hospital, sort the travel times of all feasible paths in descending order; sum the longest travel time, the shortest travel time, and the travel time that is the median in the sequence and obtain the average travel time to get the path buffer time.

[0016] S4: Record the time when the guardian leaves the hospital, and obtain the total time based on the sum of the estimated processing time, the round-trip time, and the path buffer time.

[0017] It should be noted that when the call time is 0, it means that there is no need to call the guardian or call after the patient finishes seeing a doctor.

[0018] By comprehensively considering the estimated processing time, round-trip time, and path buffer time, the system can relatively accurately estimate the total time for the guardian to leave and return to the hospital; by calculating the average of the longest time, shortest time, and median time as the path buffer time, the system can more comprehensively consider the time differences that different paths may bring, improving the accuracy of the estimation.

[0019] Preferably, the calculation of the estimated waiting time for seeing a doctor includes:

[0020] Retrieve the current real-time queuing number of the department where the current patient needs to see a doctor and the average consultation time of the department from the hospital's consultation system;

[0021] Judge whether a new number needs to be retrieved for a follow-up consultation in the corresponding department;

[0022] If yes, take the product of the real-time queuing number and the average consultation time of the department as the estimated waiting time for seeing a doctor;

[0023] If no, preset a historical period, obtain the number of days when there are no remaining registered numbers in the corresponding department, and count the number of registered numbers released by the department every day during the corresponding days; mark the difference between the department working hours and the product of the average consultation time and the number of registered numbers as the follow-up consultation time; take the quotient of the follow-up consultation time and the department working hours as the average follow-up consultation time per hour; calculate and round up the product of the real-time queuing number and the average consultation time of the department to obtain the estimated consultation hours, and mark the sum of the product of the real-time queuing number and the average consultation time of the department and the product of the hours and the average follow-up consultation time as the estimated waiting time for seeing a doctor.

[0024] For follow-up patients, the present invention can automatically judge whether a new number needs to be retrieved through the process, avoiding repeated queuing caused by misunderstanding or information lag; by intelligently calculating the estimated waiting time, patients can arrange their time more reasonably and reduce the ineffective waiting in the hospital.

[0025] Preferably, the retrieval of the time prediction model to predict the patient's consultation time includes:

[0026] Obtain the patient's medical record information and analyze the complexity of the condition based on the medical record information. Among them, the complexity of the condition includes: simple first consultation, simple follow-up consultation, multiple-symptom first consultation, multiple-symptom follow-up consultation;

[0027] Obtain the patient's age and set an understanding degree value according to the patient's age;

[0028] Integrate the complexity of the condition and the understanding degree value into a time prediction set, and mark the time prediction set as a time prediction recognition sequence;

[0029] Retrieve the time prediction model, input the time prediction recognition sequence into the time prediction model to obtain a time prediction label; match the patient's visit time according to the time prediction label; wherein, the time prediction model is constructed based on an artificial intelligence model; the time prediction label is set as a positive integer.

[0030] In the present invention, the disease complexity and the understanding degree value are integrated into a time prediction set, and the time prediction set is marked as a time prediction recognition sequence; wherein, the higher the disease complexity, the longer the time prediction set; the larger the understanding degree value, the shorter the time prediction set.

[0031] Preferably, determining the call time according to the time required for the guardian to travel to and from the hospital, the visit time, and the estimated waiting time includes:

[0032] Calculate the difference between the sum of the visit time and the estimated waiting time and the time required for the guardian to travel to and from the hospital; when the difference is greater than 0, mark the call time as 0; when the difference is less than or equal to 0, classify the visit situation and calculate the early call time under different types of visit situations.

[0033] Preferably, classifying the visit situation and calculating the early call time under different types of visit situations includes:

[0034] Retrieve the visit situation, and divide the visit situation into: post-visit situations with guardian accompaniment and post-visit situations without guardian accompaniment; wherein, post-visit situations with guardian accompaniment include: physical examination, hospitalization, and going home; post-visit situations without guardian accompaniment include: getting medicine and infusion.

[0035] Judge whether the patient's current visit situation is a post-visit situation without guardian accompaniment.

[0036] If yes, retrieve the time for the patient to get medicine and have an infusion; judge whether the total time for the guardian to leave the hospital and return to the hospital is less than the sum of the estimated patient waiting time, the patient's visit time, and the time for the patient to get medicine and have an infusion; if yes, mark it as calling the guardian immediately after the visit ends; if not, mark the absolute value of the difference between the sum of the guardian's round-trip time and the path buffer time and the sum of the estimated patient waiting time and the patient's visit time as the early call time after the visit ends.

[0037] If not, mark it as calling the guardian immediately after the visit ends.

[0038] Construct a correction factor to perform real-time correction on the early call time to obtain a real-time corrected call time, and call the guardian according to the real-time corrected call time.

[0039] In the case of post - diagnosis without the accompaniment of a guardian, the system of the present invention can intelligently judge whether the round - trip time of the guardian is sufficient to ensure that the patient can receive timely help or accompaniment when needed, improving the safety and convenience of the patient; by accurately calculating the round - trip time of the guardian and the expected medical treatment process time of the patient, the system can reasonably arrange the return time of the guardian, avoiding the long - time waiting of the guardian or the unaccompanied situation of the patient, thus optimizing the time management of the guardian.

[0040] Preferably, constructing a correction factor to perform real - time correction on the early call time to obtain the real - time corrected call time includes:

[0041] Retrieve the early call time, count the total number of trips of the guardian within a set period and the number of times the guardian fails to return to the hospital within the early call time;

[0042] Calculate the round - trip delay rate Dl of the guardian by dividing the number of times the guardian fails to return to the hospital within the early call time by the total number of trips;

[0043] Obtain the real - time weather conditions of the current period, and assign values to the weather types respectively according to the preset weather type classification table to obtain the weather impact weight W;

[0044] Through the formula Calculate to obtain the correction factor x;

[0045] Mark the product of the early call time and the correction factor as the real - time corrected call time.

[0046] It should be noted that if the real - time corrected call time exceeds the sum of the expected waiting time of the patient for medical treatment, the medical treatment time of the patient, and the time for the patient to get medicine and have an infusion, then the difference between the real - time corrected call time and the sum of the times is used as the new real - time corrected call time.

[0047] To achieve the above - mentioned purpose, the second aspect of the present invention provides an intelligent escort call method based on an AI model, including:

[0048] Obtain the guardian information and patient information in the intelligent wheelchair system;

[0049] Calculate the total time for the guardian to leave and return to the hospital according to the guardian information;

[0050] The patient registers through the intelligent wheelchair and calculates the expected waiting time for medical treatment;

[0051] Retrieve the time prediction model to predict the medical treatment time of the patient and obtain the medical treatment situation of the patient during the medical treatment; wherein, the time prediction model is constructed based on an artificial intelligence model;

[0052] Determine whether the guardian can return to the hospital before the end of the consultation time; if so, there is no need to call; if not, classify the consultation situation and calculate the early call time for different types of consultation situations.

[0053] The third aspect of the present invention provides an intelligent escort call storage medium based on an AI model, on which a computer-readable storage medium is stored. When the computer-readable storage medium is executed by a processor, it implements an intelligent escort call system based on an AI model as described in the first aspect above.

[0054] Compared with the prior art, the beneficial effects of the present invention are:

[0055] 1. Through the time prediction module, the present invention can accurately calculate the total time for the guardian to leave and return to the hospital, helping the guardian better plan the itinerary and ensuring timely return when necessary; after the patient registers through the intelligent wheelchair, the system can calculate the expected waiting time for the consultation, reducing the ineffective waiting of the patient in the hospital and improving the efficiency of the overall medical service; the time prediction model is constructed based on artificial intelligence and can more accurately predict the patient's consultation time, enabling the patient and the guardian to make preparations in advance and reducing waiting. During the consultation process, the system can obtain the patient's consultation situation in real time, provide timely information updates for the guardian, and enhance the transparency of information; the call module can determine whether the guardian can return to the hospital before the end of the consultation time and make an early call when necessary, ensuring that the patient receives necessary company and care during the consultation process, especially in case of emergency, it can respond quickly.

[0056] 2. The system of the present invention can also classify and calculate the early call time according to the consultation situation, ensuring that different types of consultation situations can be properly handled, further improving the safety guarantee level; data-driven decision-making: the entire system is based on data collection and analysis, and can form data-driven decision support, helping medical institutions better understand the needs of patients and the situation of guardians, so as to optimize resource allocation and service processes. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0058] Figure 1 It is a schematic diagram of the module relationship included in the present invention;

[0059] Figure 2 It is a schematic diagram of the specific steps of the time model prediction of the present invention;

[0060] Figure 3 Schematic diagram of the specific steps for determining the call time of the present invention;

[0061] Figure 4 Schematic diagram of the process of intelligent escort call of the present invention. Specific embodiments

[0062] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0063] Please refer to Figure 1 , an embodiment of the first aspect of the present invention provides an intelligent escort call system based on an AI model, including: a data collection module, a time prediction module, and a call module;

[0064] Data collection module: Obtain the guardian information and patient information in the intelligent wheelchair system;

[0065] Time prediction module: Calculate the total time for the guardian to leave and return to the hospital according to the guardian information;

[0066] The patient registers through the intelligent wheelchair, calculates the estimated waiting time for seeing a doctor; retrieves the time prediction model to predict the patient's seeing a doctor time, and obtains the patient's seeing a doctor situation during the seeing a doctor; wherein, the time prediction model is constructed based on an artificial intelligence model;

[0067] Call module: Judge whether the guardian can return to the hospital before the end of the seeing a doctor time; if yes, no call is required; if not, classify the seeing a doctor situation and calculate the early call time under different types of seeing a doctor situations.

[0068] Please refer to Figure 2 , the specific steps of time model prediction, judge whether the intelligent wheelchair can detect the Bluetooth signal of the guardian's mobile phone within a preset time threshold; if yes, the guardian has not left the hospital; if not, obtain the GPS coordinates of the guardian's mobile phone and judge whether the GPS coordinates exceed the hospital electronic fence; if yes, the guardian has left the hospital; if not, the guardian has not left the hospital;

[0069] Record the time when the guardian leaves the hospital and retrieve the guardian information; wherein the guardian information includes: the destination of the guardian, the travel mode, and the estimated processing time;

[0070] Estimate the round-trip time of the guardian through the map API;

[0071] Sum up the estimated processing time, round-trip time, and path buffer time to obtain the total time for the guardian to leave and return to the hospital.

[0072] The method for obtaining the path buffer time: Obtain all feasible paths between the guardian and the hospital, and sort the travel times of all feasible paths in descending order; Sum up the longest travel time, the shortest travel time, and the travel time that is the median in the sequence and obtain the average travel time; Use the average travel time as the path buffer time.

[0073] Retrieve the patient information of the patient, and through the intelligent wheelchair, retrieve the real-time queuing number of the corresponding medical treatment items of the patient in the hospital visit system, and obtain the average consultation time of the department.

[0074] Judge whether it is necessary to re-queue for a follow-up visit to the corresponding department.

[0075] If yes, then take the product of the real-time queuing number and the average consultation time of the department as the estimated waiting time for the consultation.

[0076] If no, then preset a historical period, obtain the number of days when there are no remaining queuing numbers in the corresponding department, and count the number of queuing numbers released by the department every day during the corresponding days; Take the difference between the department working hours and the product of the average consultation time of the department and the number of queuing numbers released as the follow-up visit time; Take the quotient of the follow-up visit time and the department working hours as the average follow-up visit time per hour; Calculate the product of the real-time queuing number and the average consultation time of the department and round it to obtain the estimated consultation hours, and take the sum of the product of the real-time queuing number and the average consultation time of the department and the product of the hours and the average follow-up visit time as the estimated waiting time for the consultation.

[0077] Retrieve the medical record information of the patient, and analyze the complexity of the condition according to the medical record information. Among them, the complexity of the condition includes: simple initial diagnosis, simple follow-up visit, multi-symptom initial diagnosis, multi-symptom follow-up visit.

[0078] Obtain the patient's age, and set the comprehension level value according to the patient's age.

[0079] Integrate the complexity of the condition and the comprehension level value into a time prediction set, and mark the time prediction set as a time prediction recognition sequence.

[0080] Retrieve the time prediction model, input the time prediction recognition sequence into the time prediction model to obtain a time prediction label; Match the patient's consultation time according to the time prediction label.

[0081] For example, there is currently a guardian A who leaves after sending patient B to the hospital. Retrieve the preset destination of the guardian from the intelligent wheelchair, the travel mode is by car, and the estimated processing time is 30 minutes.

[0082] Estimate through the map API that the total round-trip time of the guardian on the shortest path is 30 minutes.

[0083] Obtain all 5 feasible paths between the guardian and the hospital, namely L1, L2, L3, L4, and L5; among them, L1 takes 15 minutes, L2 takes 16 minutes, L3 takes 26 minutes, L4 takes 19 minutes, and L5 takes 22 minutes

[0084] Through the formula The path buffer time is calculated to be 20 minutes by the formula of minutes;

[0085] Sum up the estimated processing time, round-trip time, and path buffer time to obtain the total time for the guardian to leave and return to the hospital, which is 80 minutes;

[0086] Retrieve the patient information of the patient, and through the intelligent wheelchair, retrieve the real-time queuing number of the corresponding medical treatment items of the patient in the hospital medical treatment system, which is 10 people, and obtain the average medical treatment time of the department as 8 minutes;

[0087] For the review of the orthopedic department of patient B, a new number needs to be taken. Then, multiply the real-time queuing number by the average medical treatment time of the department as the estimated waiting time for medical treatment, which is 80 minutes;

[0088] Obtain the medical record information of patient B. Patient B is for a simple review; obtain the patient's age as 48 years old, and according to the relationship table between age and understanding degree value (Table 1), obtain the understanding degree value of patient B as 0.8;

[0089] Table 1 Relationship table between age and understanding degree value

[0090]

[0091]

[0092] Integrate the disease complexity and the understanding degree value into a time prediction set, and mark the time prediction set as a time prediction recognition sequence; retrieve the time prediction model, input the time prediction recognition sequence into the time prediction model to obtain a time prediction label; match the patient's medical treatment time as 10 minutes according to the time prediction label.

[0093] Please refer to Figure 3 , for the specific steps of determining the call time, retrieve the total time for the guardian to leave and return to the hospital, the patient's medical treatment time, and the estimated waiting time for medical treatment;

[0094] Determine whether the total time is greater than the sum of the consultation time and the estimated waiting time for consultation; if yes, mark that the guardian can return to the hospital before the end of the consultation time; if no, mark that the guardian cannot return to the hospital before the end of the consultation time; retrieve the consultation situation and divide the consultation situation into: post-consultation situations with guardian accompaniment and post-consultation situations without guardian accompaniment; among them, the post-consultation situations with guardian accompaniment include: physical examination, hospitalization, and going home; the post-consultation situations without guardian accompaniment include: getting medicine and infusion;

[0095] Determine whether the current consultation situation of the patient is a post-consultation situation without guardian accompaniment;

[0096] If yes, retrieve the time for the patient to get medicine and infusion; determine whether the total time from the guardian leaving the hospital to returning to the hospital is less than the sum of the estimated waiting time for the patient to consult, the patient's consultation time, and the time for the patient to get medicine and infusion; if yes, call the guardian immediately after the consultation ends; if no, take the absolute value of the difference between the sum of the guardian's round-trip time and the path buffer time minus the sum of the estimated waiting time for the patient to consult and the patient's consultation time as the early call time after the consultation ends;

[0097] If no, call the guardian immediately after the consultation ends;

[0098] Construct a correction factor to perform real-time correction on the early call time to obtain the real-time corrected call time, and call the guardian according to the real-time corrected call time;

[0099] Retrieve the early call time, count the total number of trips of the guardian within the set period and the number of times the guardian fails to return to the hospital within the early call time;

[0100] Obtain the round-trip delay rate Dl of the guardian by dividing the number of times the guardian fails to return to the hospital within the early call time by the total number of trips;

[0101] Obtain the real-time weather condition of the current period, and assign values to the weather types respectively according to the preset weather type grading table to obtain the weather influence weight W;

[0102] Through the formula Calculate to obtain the correction factor x;

[0103] Mark the product of the early call time and the correction factor as the real-time corrected call time.

[0104] For example, retrieve the total time of 80 minutes from guardian A leaving the hospital to returning to the hospital, the consultation time of 10 minutes for patient B, and the estimated waiting time for consultation of 80 minutes;

[0105] Since the total time of 80 minutes is greater than the sum of the consultation time and the estimated waiting time for consultation of 90 minutes, retrieve the consultation situation of patient B, and the consultation situation is to get medicine; among them, the time for getting medicine is 10 minutes;

[0106] Since the total time of 80 minutes for the guardian to leave and return to the hospital is greater than the sum of the expected waiting time for the patient to see a doctor, the patient's consultation time, and the patient's time to pick up medicine, which is 100 minutes, the absolute value of the difference between the sum of the guardian's round-trip time and the path buffer time (50 minutes) minus the sum of the expected waiting time for the patient to see a doctor and the patient's consultation time (90 minutes) is used as the early call time after the consultation ends;

[0107] Then the intelligent wheelchair calls the guardian A 40 minutes before the end of the consultation.

[0108] Retrieve the early call time of 40 minutes, and count that the total number of times the guardian travels during the set period is 5 times and the number of times the guardian does not return to the hospital within the early call time is 1 time;

[0109] By dividing the number of times the guardian does not return to the hospital within the early call time by the total number of trips, the round-trip delay rate of the guardian is obtained as 0.2;

[0110] Obtain the current real-time weather condition as light rain, and assign a value to the weather type according to the preset weather type classification table (as shown in Table 2) to obtain the weather influence weight W as 0.13;

[0111] Through the formula Calculate to obtain the correction factor x = 1.248;

[0112] Mark the product of the early call time and the correction factor as the real-time corrected call time, that is, 40×1.248 = 49.92 minutes; call the guardian according to the real-time corrected call time (49.92 minutes).

[0113] Table 2 Weather type classification table

[0114] Weather type Subclass and definition Weather impact weight Sunny / Few clouds No precipitation, cloud cover < 30% 0 Cloudy Cloud cover 30% - 70% 0 Overcast Cloud cover > 70%, no precipitation 0.1 Light rain Precipitation 0.1 - 9.9 mm / 24h 0.13 Moderate rain Precipitation 10 - 24.9 mm / 24h 0.2 Heavy rain / Storm Precipitation ≥ 25 mm / 24h 0.25 Thundershower Short-term heavy precipitation accompanied by lightning 0.5 Light snow Snowfall ≤ 2.4 mm / 24h 0.5 Moderate snow Snowfall 2.5 - 4.9 mm / 24h 0.7 Heavy snow / Blizzard Snowfall ≥ 5 mm / 24h 0.9 Fog Visibility 500 m - 1 km 0.5 Dense fog / Heavy dense fog Visibility < 500 m 0.9

[0115] Please refer to Figure 4 , an intelligent escort call method based on an AI model provided by the second aspect embodiment of the present invention includes:

[0116] Obtain the guardian information and patient information in the intelligent wheelchair system;

[0117] Calculate the total time for the guardian to leave and return to the hospital according to the guardian information;

[0118] The patient registers through the intelligent wheelchair and calculates the expected waiting time for seeing a doctor;

[0119] Retrieve the time prediction model to predict the patient's consultation time and obtain the patient's consultation situation during the consultation; among them, the time prediction model is constructed based on the artificial intelligence model;

[0120] Determine whether the guardian can return to the hospital before the end of the consultation time; if so, there is no need to call; if not, classify the consultation situation and calculate the early call time under different types of consultation situations.

[0121] The third aspect of the present invention provides an intelligent escort call storage medium based on an AI model, on which a computer-readable storage medium is stored. When the computer-readable storage medium is executed by a processor, it implements an intelligent escort call system based on an AI model as described in the first aspect above.

[0122] Some of the data in the above formula is calculated by removing the dimension and taking its numerical value. The formula is obtained by software simulation of a large amount of collected data to get a formula closest to the actual situation; the preset parameters and preset thresholds in the formula are set by those skilled in the art according to the actual situation or obtained through simulation of a large amount of data.

[0123] The working principle of the present invention: The present invention obtains the guardian information and patient information in the intelligent wheelchair system; calculates the total time for the guardian to leave and return to the hospital according to the guardian information; the patient registers through the intelligent wheelchair and calculates the estimated waiting time for the consultation; retrieves the time prediction model to predict the patient's consultation time and obtains the patient's consultation situation during the consultation; among them, the time prediction model is constructed based on an artificial intelligence model; determines whether the guardian can return to the hospital before the end of the consultation time; if so, there is no need to call; if not, classify the consultation situation and calculate the early call time under different types of consultation situations.

[0124] The above embodiments are only used to illustrate the technical method of the present invention and not to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical method of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical method of the present invention.

Claims

1. An AI model-based intelligent escort call system, characterized in that: include: A time prediction module and a data acquisition module and a call module connected thereto; Data collection module: obtain the guardian's travel information and the patient's medical information; the medical information is the medical process and medical situation input by the diagnosing doctor when the patient visits the hospital; Time prediction module:

3. Calculate the total time for the guardian to go to and from the hospital based on the travel information; and, 4. Calculate the estimated waiting time for the consultation based on the real-time number of people waiting in line for the patient's department; 5. Input the consultation information into the time prediction model to obtain the consultation time; wherein the time prediction model is constructed based on the artificial intelligence model; Call module: Determine the advance call time according to the time required for the guardian to travel to and from the hospital, the consultation time and the estimated waiting time, and call the guardian according to the advance call time.

2. According to claim 1, an AI model-based intelligent escort call system is characterized in that: Also includes: Time revision module: used to retrieve the real-time queue number of each medical item in the hospital's medical system, as well as the average consultation time of the corresponding department; among them, patient information includes: medical record information, the patient's medical department, and the time when the patient gets medicine and infusion. It should be noted that the data in this system can be retrieved from the hospital's medical system.

3. According to claim 1, an AI model-based intelligent accompanying call system is characterized in that: The total time for the guardian to travel to and from the hospital is calculated based on the travel information, including: The calculation of the time required for the guardian to travel to and from the hospital based on the travel information includes: S1: Extract the guardian's travel information; the guardian information includes the guardian's location, estimated processing time and the guardian's destination; determine whether the guardian is in the hospital based on the guardian's location; if yes, mark the call time as 0; if no, jump to S2; S2: Obtain the guardian’s round trip time through the map API; S3: Obtain all feasible paths between the guardian and the hospital, and sort the time of all feasible paths in descending order; sum the longest time, the shortest time, and the median time in the sequence and obtain the average time to obtain the path buffer time; S4: Record the time when the guardian leaves the hospital, and obtain the total time based on the sum of the estimated processing time, round-trip time, and path buffer time.

4. According to the AI ​​model-based intelligent accompanying call system of claim 1, it is characterized in that: The estimated waiting time for consultation is calculated based on the real-time number of people in the queue for registration in the department where the patient is located, including: Retrieve the real-time queue number of the department that the current patient wants to visit and the average consultation time of the department; Determine whether a new number is required for follow-up visits to the corresponding department; If yes, the product of the real-time number of people in line and the average consultation time of the department is used as the estimated waiting time for consultation; If not, preset the historical period, obtain the number of days when the corresponding department has no remaining registrations, and count the number of numbers released by the department every day on the corresponding days; mark the difference between the department working hours and the product of the average department consultation time and the number of numbers released as the follow-up time; take the quotient of the follow-up time and the department working hours as the average follow-up time per hour; calculate the product of the real-time number of people in line and the average department consultation time and round it to get the estimated consultation hours, and mark the product of the real-time number of people in line and the average department consultation time plus the product of the number of hours and the average follow-up time as the estimated waiting time for consultation.

5. According to claim 2, an AI model-based intelligent accompanying call system is characterized in that: The step of inputting the consultation information into the time prediction model to obtain the consultation time includes: Obtain the patient's medical record information, and analyze the complexity of the condition based on the medical record information, where the complexity of the condition includes: simple initial visit, simple follow-up visit, multiple symptom initial visit, and multiple symptom follow-up visit; Get the patient's age and set the understanding level value according to the patient's age; Integrate the disease complexity and understanding level values ​​into a time prediction set, and mark the time prediction set as a time prediction recognition sequence; Retrieve the time prediction model, input the time prediction identification sequence into the time prediction model, and obtain the time prediction label; obtain the patient's consultation time according to the time prediction label matching; wherein, the time prediction model is constructed based on the artificial intelligence model; and the time prediction label is set to a positive integer.

6. According to the AI ​​model-based intelligent accompanying call system of claim 1, it is characterized in that: The call time is determined based on the time required for the guardian to travel to and from the hospital, the consultation time and the estimated waiting time, including: Calculate the difference between the sum of the consultation time and the expected waiting time and the time required for the guardian to travel to and from the hospital; when the difference is greater than 0, mark the call time as 0; when the difference is less than or equal to 0, classify the consultation situation and calculate the advance call time for different types of consultation situations.

7. According to claim 6, an AI model-based intelligent accompanying call system is characterized in that: The method of classifying the medical visits and calculating the advance call time for different types of medical visits includes: The medical records were retrieved and divided into two types: post-diagnosis records with guardians accompanying the patient and post-diagnosis records without guardians accompanying the patient. Among them, post-diagnosis records with guardians accompanying the patient included: physical examination, hospitalization, and going home; post-diagnosis records without guardians accompanying the patient included: taking medicine and infusion. Determine whether the patient's current medical condition is a post-diagnosis condition that does not require a guardian to accompany him / her; If yes, retrieve the time for the patient to get medicine and infusion; determine whether the total time from the guardian leaving the hospital to returning to the hospital is less than the sum of the estimated patient waiting time, the patient's consultation time, and the patient's medicine and infusion time; if yes, mark it as calling the guardian immediately after the consultation; if no, mark the absolute value of the difference between the sum of the guardian's round-trip time and the path buffer time minus the sum of the estimated patient waiting time and the patient's consultation time as the advance call time after the consultation; If no, it was marked as calling the guardian immediately after the visit; A correction factor is constructed to correct the advance call time in real time to obtain the real-time corrected call time, and the guardian is called according to the real-time corrected call time.

8. According to claim 7, an AI model-based intelligent accompanying call system is characterized in that: The step of constructing a correction factor to correct the advance call time in real time to obtain a real-time corrected call time includes: Retrieve the advance call time and mark it as T, and count the total number of trips of the guardian within the set period and the number of times the guardian did not return to the hospital within the advance call time; The guardian's round-trip delay rate Dl is obtained by dividing the number of times the guardian did not return to the hospital within the advance call time by the total number of trips; Obtain the real-time weather conditions of the current period, and assign values ​​to the weather types according to the preset weather type classification table to obtain the weather impact weight W; By formula The correction factor x is calculated; The product of the advance call time and the correction factor is labeled as the real-time corrected call time. It should be noted that if the real-time corrected call time exceeds the sum of the estimated patient waiting time, the patient's consultation time, and the time it takes for the patient to get medicine and receive infusion, the difference between the real-time corrected call time and the sum of the time will be used as the new real-time corrected call time.

9. An AI model-based intelligent accompanying calling method, adapted to an AI model-based intelligent accompanying calling system according to any one of claims 1 to 8, characterized in that: include: Obtain guardian information and patient information in the smart wheelchair system; Calculate the total time from when the guardian leaves the hospital to when he / she returns to the hospital based on the guardian's information; Patients use smart wheelchairs to register and calculate the estimated waiting time for consultation; Retrieve the time prediction model to predict the patient's consultation time and obtain the patient's consultation status during the consultation; the time prediction model is built based on the artificial intelligence model; Determine whether the guardian will be able to return to the hospital before the end of the consultation period; If yes, no call is needed; if no, the visits are classified and the advance call time is calculated for different types of visits.

10. An AI model-based intelligent accompanying call storage medium, on which a computer-readable storage medium is stored, characterized in that: When the computer-readable storage medium is executed by a processor, it implements an AI model-based intelligent escort call system as described in any one of claims 1-8.

Images