Single-person first-aid transfer intelligent decision-making system and method based on multi-modal data

Abstract

The invention discloses a single-person first-aid transfer intelligent decision-making system and method based on multi-modal data, and relates to the technical field of first-aid transfer dispatch, and the method comprises the following steps: obtaining the related data of a patient through an integrated first-aid platform; preprocessing the obtained related data, sorting the preprocessed data according to scores from low to high, and generating a patient priority queue; calculating a patient emergency degree index based on the waiting times and duration of the rescue personnel in the waiting area; performing first-aid resource evaluation and adjustment, performing rescue according to the time when the patient arrives at the hospital when the total amount of the on-site deployable first-aid resources is enough for the rescue work of the current patient, performing rescue according to the time when the patient arrives at the hospital when the total amount of the on-site deployable first-aid resources is insufficient, and performing rescue according to the time when the patient waits. And sorting the patient priority queues according to the emergency degree indexes of the patients. According to the invention, first-aid transfer scheduling is realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of emergency transfer scheduling, in particular to a single-person emergency transfer intelligent decision system and method based on multi-modal data. BACKGROUND

[0002] With the increasing demand for refinement of earthquake rescue scenarios for emergency transfer, the limitations of traditional emergency transfer scheduling methods in complex environments are increasingly highlighted. As a core link of earthquake debris rescue, the comprehensiveness of data collection, the accuracy of priority judgment, and the rationality of resource scheduling in the transfer process affect the efficiency of patient treatment.

[0003] However, traditional emergency transfer scheduling often faces the following problems when dealing with multi-modal data processing, dynamic priority adjustment, and resource optimization: First, the emergency priority judgment mechanism is single and lagging, and the traditional method mainly relies on Glasgow coma score sorting, lacking consideration of patient emergency level. Second, the resource assessment and scheduling strategy lacks dynamicity, and the traditional scheduling does not quantitatively assess the resource supply-demand relationship. When the total patient emergency demand exceeds the deployable resources, it cannot quickly calculate the resource gap and start the external support mechanism. Finally, the multi-modal data collection and fusion capability is insufficient, and the traditional method relies on manual intermittent collection of patient signs and trauma data, which cannot integrate geographic location information in real time, resulting in the inability to quickly assess the transfer risk. At the same time, the data format is not unified, making it difficult to form a complete patient state portrait. SUMMARY

[0004] The purpose of the present application is to provide a single-person emergency transfer intelligent decision system and method based on multi-modal data to solve the problems in the prior art.

[0005] To achieve the above-mentioned purpose, the present application provides the following technical solution: a single-person emergency transfer intelligent decision method based on multi-modal data, the method comprising the following steps: Obtaining relevant data of patients through an integrated emergency platform; Preprocessing the obtained relevant data, sorting the data from low to high based on the score after preprocessing, and generating a patient priority queue; Calculating the patient emergency level index based on the waiting number and duration of the rescuer in the waiting area; According to the total amount of current patient emergency needs, the total amount of on-site deployable emergency resources and the minimum amount of emergency resources set to ensure basic rescue ability, the emergency resource evaluation adjustment is carried out, for the total amount of on-site deployable emergency resources sufficient for the current patient rescue work, according to the patient arrival time for rescue, for the total amount of on-site deployable emergency resources cannot meet the current patient rescue work, according to the patient arrival time for rescue, for the existence of patient waiting, according to the patient emergency index, the priority queue of the above patients is sorted, and the patient rescue is carried out according to the adjusted patient priority queue.

[0006] The related data of the patient is obtained through the integrated emergency platform, and the specific steps include: The sign data, state data and geographical position information of the patient are obtained through the integrated emergency platform, wherein the sign data includes blood pressure, heart rate and blood oxygen data of the patient, the state data includes the type of external injury, the injured part, the trauma area, the wound depth, and whether it is active bleeding, the integrated emergency platform includes a wheeled multi-position stretcher, a head and neck hemostasis device, a ventilation module, a fluid replacement module, a monitoring module, a fixing module, a closed drainage module and a rewarming module, the wheeled multi-position stretcher is used as a carrying tool for the wounded to receive treatment, transportation and evacuation during the emergency process; the head and neck hemostasis device is a device for head and neck hemorrhage; the ventilation module includes a blind insertion tracheal intubation guide with a light source and a new type of visual laryngoscope intubation; the fluid replacement module includes a bone marrow cavity infusion puncture gun and a satellite pressurized infusion device; the monitoring module includes an electrocardiogram monitor and an oxygen saturation monitor; the fixing module includes an inflatable cervical spine fixator and a high molecular material fracture fixation bandage; the closed drainage module includes a chest closed drainage device; the rewarming module is used for rewarming the human torso and keeping the limbs warm; The wheeled multi-position stretcher is a rescue equipment integrating transportation, transfer and fixation functions, which is suitable for disaster rescue, pre-hospital emergency and transfer of wounded in complex disaster scenes. Core components: boat stretcher bed, liftable universal wheel, safety belt / fixed belt, storage box; Features: according to the different position forms of the stretcher, it can rotate in multiple degrees of freedom, and one person can perform upright, inclined, plane, upper gangway transfer and wall hanging storage operation; The storage box is internally provided with other module devices; Honeycomb carbon fiber and alloy are adopted, and the material is waterproof and corrosion resistant, and the bearing capacity is ≤1500N; Moving mode: horizontal movement / sliding / vertical movement / universal wheel based; Size (mm): 1912*500*290; weight (kg): 12; suitable environment: -20℃-60℃; The indications of the wheeled multi-position stretcher include: Transportation of wounded in sea and land disaster sites, including ship fire, traffic accident, earthquake, etc. Transportation of wounded with spinal injury, fracture, etc. that need to be fixed and carried; Rescue in narrow space (such as basement, ruins); Batch transportation of wounded or long-distance medical escort; Contraindications of the wheeled multi-position stretcher include: Suspected cervical / thoracolumbar injury patients without fixed spine (need to use neck brace and spine plate first); Severe open chest and abdominal trauma (avoid squeezing the injured part, and prioritize bleeding treatment); Extreme environment (such as fire site where explosion risk has not been ruled out, unstable collapse area); Through the integrated first aid platform, the patient's vital sign data, state data and geographic location information are obtained. Further, a non-invasive blood pressure sensor based on the oscillograph method is installed on one side of the stretcher near the patient's arm to obtain the patient's blood pressure data. An optical heart rate sensor is integrated into the headrest of the stretcher head, which contacts the patient's head skin to obtain the patient's heart rate data. A blood oxygen sensor based on the PPG principle is installed on the stretcher near the patient's finger or toe fixed clip to obtain the patient's blood oxygen data. State data is collected by a combination of high-definition cameras and depth sensors. The high-definition camera is installed above the front end of the stretcher and can rotate 360 degrees, which is used to shoot the overall and wound appearance image of the patient. The depth sensor is installed beside the camera, which uses structured light or time-of-flight (ToF) technology to obtain three-dimensional depth information of the wound. The rescuer can manually operate the camera to aim at the wound on the spot, and the device automatically analyzes the image and depth data to identify the type of injury, locate the injured part, and calculate the wound area and depth. A pressure sensor array is laid on the surface of the stretcher, distributed in the area corresponding to the easy bleeding part of the patient's body, such as the back and buttocks. When blood flows out and contacts the pressure sensor, the sensor will generate an electrical signal due to pressure change, by which it is determined whether there is active bleeding. The stretcher is equipped with GPS positioning module and Beidou positioning module to update the patient's location in real time. In the earthquake ruins rescue scene, the patient data required for calculating the Glasgow coma score is obtained by observation and inquiry, including the patient's eye opening reaction, language reaction and movement reaction. According to the patient data obtained for calculating the Glasgow coma score, the corresponding scoring standard is set, the scores of the patient's eye opening reaction, language reaction and movement reaction are added, and the Glasgow coma score is calculated. The specific scoring criteria are as follows: Eye opening response: 4 points for spontaneous eye opening; 3 points for eye opening after hearing sound stimulation; 2 points for eye opening after giving pain stimulation; 1 point for no eye opening response under any stimulation; Language response: 5 points for being able to answer time and place questions and having normal orientation; 4 points for answering errors but being able to communicate; 3 points for language disorder and being unable to communicate normally; 2 points for being able to only make meaningless sounds; 1 point for no language response; Motor response: 6 points for being able to complete actions as instructed; 5 points for being able to locate pain stimulation; 4 points for having avoidance action to pain stimulation; 3 points for limb flexion (decortication) under pain stimulation; 2 points for limb extension (decerebrate rigidity) under pain stimulation; 1 point for no limb movement response; The Glasgow coma score is synchronized with the obtained patient sign data, state data and patient geographic location information to form a set X, and the definition of set X is as follows: X={(GCS1,s1,t1,l1),(GCS2,s2,t2,l2),...,(GCS q ,s q ,t q ,l q )}, wherein GCS1,GCS2,...,GCS q represent the Glasgow coma scores of the 1st, 2nd,..., qth patients, s1,s2,...,s q represent the sign data sets of the 1st, 2nd,..., qth patients, t1,t2,...,t q represent the state data sets of the 1st, 2nd,..., qth patients, and l1,l2,...,l q represent the geographic location information of the 1st, 2nd,..., qth patients.

[0007] The obtained related data is preprocessed, and the patient emergency treatment priority is determined based on the preprocessed data, and the specific steps include: The obtained related data is cleaned and preprocessed, when the data value at a certain time is missing, the interpolation method based on time series is used to fill in the data value according to the data values of adjacent time points, and the data format is standardized, and all data is stored according to a unified database table structure; The Min-Max normalization algorithm is used to process the obtained light sensing data, and the data is uniformly mapped to the [0,1] interval; Real-time traversal of all patients' Glasgow coma score, according to the score from low to high, generate patient priority queue, when receiving a new patient, judge the new patient's Glasgow coma score is less than or equal to the set score threshold, directly as the highest priority and inserted into the queue first, for the new patient's Glasgow coma score is greater than the set score threshold, compare its score with the existing patient's score in the queue in turn, there are the same score patients, according to the arrival time in turn for secondary sorting, for the same score patients and the same time to queue, combined with the obtained patient blood pressure systolic pressure data for tertiary sorting, after the completion of the insertion, to the field personnel and the hospital update the priority list; The obtained patient blood pressure systolic pressure data for tertiary sorting, the specific steps are as follows: Primary sorting: compare the systolic pressure data with the normal range, the patients with systolic pressure lower than the lower limit of the normal range, according to the principle of the lower the systolic pressure, the higher the priority; Secondary sorting: for patients with systolic pressure in the normal range, further compare the deviation of systolic pressure from the normal range. The greater the deviation, the later the sorting; Tertiary sorting: for patients with systolic pressure higher than the upper limit of the normal range, according to the order of the higher the systolic pressure, the later the order.

[0008] The waiting times and duration of the rescue personnel in the waiting area are used to measure the patient emergency index, the specific steps include: The waiting times and duration of the rescue personnel in the waiting area are used to measure the patient emergency index R j , defined as follows: ; Wherein, R j represents the emergency index of patient j, the value range is between 0 and 1, n represents the waiting times of the rescue personnel in the waiting area during the transfer process, t ji represents the duration of the i-th waiting of patient j, i represents the waiting number label, T represents the total duration from the beginning of the transfer task to the end of the transfer task.

[0009] According to the current total amount of emergency demand, the total amount of on-site deployable emergency resources and the set minimum emergency resource amount to ensure basic rescue capability, the emergency resource evaluation and adjustment, the specific steps include: When DT≤RT-R0, the total amount of on-site deployable emergency resources is enough for the current patient rescue work, according to the patient arrival time for rescue, wherein, DT represents the total amount of current patient emergency demand, RT represents the total amount of on-site deployable emergency resources, R0 represents the set minimum emergency resource amount to ensure basic rescue capability; When DT > RT - R0, the total amount of on-site deployable emergency resources cannot meet the current patient rescue work, the total resource gap amount of DT - RT + R0 is calculated, support is sought from surrounding medical units according to the calculated total resource gap amount, rescue is performed according to the patient arrival time at the hospital, and for the case where there is patient waiting, the patient priority queue is sorted according to the patient emergency index, and patient rescue is performed according to the adjusted patient priority queue. The patient priority queue is sorted according to the patient emergency index, and the specific steps are as follows: for the case where there is patient waiting, the patient queue sequence is reordered according to the corresponding emergency index of the patient from large to small.

[0010] A single-person emergency rescue transfer intelligent decision system based on multi-modal data, the system comprises a data acquisition module, a data preprocessing module, a priority judgment module and a resource evaluation and scheduling module, the data acquisition module is used to acquire the relevant data of the patient through an integrated emergency platform; the data preprocessing module is used to preprocess the acquired relevant data; the priority judgment module is used to sort the preprocessed data according to the score from low to high to generate a patient priority queue; the resource evaluation and scheduling module is used to calculate a patient emergency index based on the waiting number and duration of the rescuer in the waiting area, to perform emergency resource evaluation and adjustment according to the total amount of current patient emergency demand, the total amount of on-site deployable emergency resources and the minimum emergency resource amount set to guarantee basic rescue capability, to perform rescue according to the patient arrival time at the hospital for the case where the total amount of on-site deployable emergency resources is sufficient for the current patient rescue work, to perform rescue according to the patient arrival time at the hospital for the case where the total amount of on-site deployable emergency resources cannot meet the current patient rescue work, to sort the patient priority queue according to the patient emergency index for the case where there is patient waiting, and to perform patient rescue according to the adjusted patient priority queue.

[0011] The data acquisition module comprises a vital sign data acquisition unit, a state data acquisition unit, a geographic position acquisition unit and a consciousness state acquisition unit, the vital sign data acquisition unit is used to acquire blood pressure, heart rate and blood oxygen data of the patient; the state data acquisition unit is used to acquire the type of external injury, the injured part, the trauma area, the wound depth and whether active bleeding of the patient; the geographic position acquisition unit is used to acquire the geographic position information of the patient; the consciousness state acquisition unit is used to acquire the eye opening reaction, language reaction and movement reaction of the patient, to set the corresponding scoring standard according to the patient data required for the acquired Glasgow coma score, to add the scores corresponding to the eye opening reaction, language reaction and movement reaction of the patient, and to calculate the Glasgow coma score.

[0012] The data preprocessing module comprises a data filling unit and a format standardization unit, the data filling unit is used for filling in data values at adjacent time points when data values at a time point are missing, and the format standardization unit is used for standardizing data formats and storing all data in a unified database table structure, and the Min-Max normalization algorithm is used to process the acquired light response data and uniformly map the data to the [0, 1] interval.

[0013] The priority judgment module comprises a GCS sorting unit, a second sorting unit and a third sorting unit, the GCS sorting unit is used for real-time traversal of Glasgow coma scores of all patients, sorting according to scores from low to high, generating a patient priority queue, when a new patient is received, judging the Glasgow coma score of the new patient, if the score is less than or equal to a set score threshold, directly indicating the highest priority and inserting into the first position of the queue, the second sorting unit is used for comparing the score of the new patient with the scores of existing patients in the queue when the score of the new patient is greater than the set score threshold, and when there are patients with the same score, the second sorting is performed according to the arrival time; the third sorting unit is used for third sorting in combination with the systolic blood pressure data of the patients when there are patients with the same score and the same queue time.

[0014] The resource assessment and scheduling module comprises an emergency degree calculation unit, a resource supply and demand assessment unit and a priority dynamic adjustment unit, the emergency degree calculation unit is used for measuring the patient emergency degree index by the waiting times and the waiting time of the rescue personnel in the waiting area; the resource supply and demand assessment unit is used for adjusting the emergency resource assessment according to the total amount of current patient emergency demand, the total amount of on-site deployable emergency resources and the minimum emergency resource amount for guaranteeing basic rescue capability; the priority dynamic adjustment unit is used for performing rescue according to the patient arrival time when the total amount of on-site deployable emergency resources is sufficient for the current patient rescue work, performing rescue according to the patient arrival time when the total amount of on-site deployable emergency resources is insufficient, and performing rescue according to the adjusted patient priority queue when there are patients waiting.

[0015] Compared with the prior art, the beneficial effects of the present application are: 1、The intelligent priority judgment mechanism is constructed, the Glasgow coma score is taken as the core, the patient arrival time and the systolic pressure data are fused for multi-factor sorting, and real-time insertion of a new patient and queue updating are supported; 2. Introducing a dynamic resource assessment and intelligent scheduling module, by quantifying the supply and demand relationship between the total quantity of patient emergency needs and the total quantity of deployable resources, automatically triggering resource gap calculation and external support mechanism, and dynamically adjusting the rescue queue combined with the patient emergency index; 3. Integrating non-invasive blood pressure, heart rate, blood oxygen and other vital sign collection devices, as well as high-definition cameras, pressure sensor arrays and other state monitoring modules, and supporting horizontal / vertical movement modes to adapt to complex terrain, unlike the defects of traditional manual data collection and single-function transfer equipment, realizing the integration and cooperation of data collection, on-site disposal and transfer process. BRIEF DESCRIPTION OF DRAWINGS

[0016] Fig. 1 A flowchart of the method for single-person emergency rescue and transfer intelligent decision-making based on multi-modal data according to the present application; Fig. 2 A structural diagram of the system for single-person emergency rescue and transfer intelligent decision-making based on multi-modal data according to the present application. DETAILED DESCRIPTION

[0017] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0018] In the embodiments, as shown in the drawings, Figs. 1-2 The present application provides a technical solution, a method for single-person emergency rescue and transfer intelligent decision-making based on multi-modal data, which comprises the following steps: Obtaining relevant data of the patient through an integrated emergency platform; Preprocessing the obtained relevant data, sorting the data according to the score from low to high based on the preprocessed data, and generating a patient priority queue; Calculating the patient emergency index based on the waiting number and duration of the rescue personnel in the waiting area; According to the total quantity of current patient emergency needs, the total quantity of on-site deployable emergency resources and the minimum emergency resource quantity set to guarantee the basic rescue capability, performing emergency resource assessment and adjustment. For the on-site deployable emergency resource quantity sufficient for the current patient rescue work, performing rescue according to the patient arrival time at the hospital. For the on-site deployable emergency resource quantity insufficient for the current patient rescue work, performing rescue according to the patient arrival time at the hospital. For the case where there are patients waiting, sorting the above patient priority queue according to the patient emergency index, and performing patient rescue according to the adjusted patient priority queue.

[0019] The relevant data of the patient is acquired through the integrated first-aid platform, and the specific steps include: The sign data, state data and geographical position information of the patient are acquired through the integrated first-aid platform, wherein the sign data includes blood pressure, heart rate and blood oxygen data of the patient, the state data includes the type of trauma, the injured part, the trauma area, the wound depth and whether active bleeding of the patient, and the integrated first-aid platform includes a wheeled multi-position stretcher, a head and neck hemostasis device, a ventilation module, a fluid replacement module, a monitoring module, a fixing module, a closed drainage module and a rewarming module, the wheeled multi-position stretcher is used as a carrying tool for the wounded to receive treatment, carrying and evacuation during the first-aid process; the head and neck hemostasis device is a device for hemostasis of head and neck hemorrhage; the ventilation module includes a blind insertion tracheal intubation guide with a light source and a new type of visual laryngoscope intubation; the fluid replacement module includes a bone marrow cavity infusion puncture gun and a satellite pressurized infusion device; the monitoring module includes an electrocardiogram monitor and an oxygen saturation monitor; the fixing module includes an inflatable cervical spine fixator and a high polymer fracture fixation bandage; the closed drainage module includes a chest closed drainage device; and the rewarming module is used for rewarming the human torso and keeping the limbs warm; The wheeled multi-position stretcher is a rescue device integrating carrying, transferring and fixing functions, and is suitable for disaster rescue, pre-hospital care and complex disaster scene patient transfer; Core components: boat stretcher bed, liftable universal wheel, safety belt / fixed belt, storage box; Features: according to the different position forms of the stretcher, the stretcher can be rotated in multiple degrees of freedom, and a single person can perform upright, inclined, planar, upper deck ladder transfer and wall hanging storage and other operations; The storage box is internally provided with other module devices; Honeycomb carbon fiber and alloy are adopted, and the material is waterproof and corrosion-resistant, and the bearing capacity is ≤1500N; Moving mode: horizontal movement / sliding / vertical movement / universal wheel based; Size (mm): 1912*500*290; weight (kg): 12; adaptive environment: -20℃-60℃; Indications of the wheeled multi-position stretcher include: Transfer of wounded in sea and land disaster scenes, including ship fire, traffic accident, earthquake, etc.; Patients with spinal injury, fracture and other injuries requiring fixed carrying; Rescue in narrow space (such as basement, ruins); Batch patient transfer or long-distance medical escort; Contraindications of the wheeled multi-position stretcher include: Suspected cervical / thoracolumbar injury patients without fixed spine (neck brace and spinal plate fixation should be used first); Severe open chest and abdominal trauma (avoid squeezing the injured part, and bleeding should be treated first); Extreme environments (such as fire scenes where explosion risks have not been eliminated, unstable collapse areas); Before using a wheeled multi-position stretcher, ensure the safety of the scene, clear obstacles along the transfer path, and choose between horizontal and vertical transfer depending on the terrain. The wheeled multi-position stretcher is equipped with a hemostasis device for severe head and neck bleeding, a blind endotracheal intubation guide with a light source, a new visual laryngoscope cannula, a bone marrow infusion puncture gun, a satellite pressurized infusion set, an ECG monitor, a blood oxygen saturation monitor, an inflatable cervical fixator, a polymer fracture fixation bandage, a closed chest drainage device, an oscillometric non-invasive blood pressure sensor, a photoelectric heart rate sensor, a PPG-based blood oxygen sensor, a high-definition camera, a depth sensor, a pressure sensor array, a GPS positioning module, and a Beidou positioning module. Secondly, check whether the pulleys of the wheeled multi-position stretcher and the equipment in the storage box are normal, test whether the safety belts and fixing belts are intact, and rescue the patient if the equipment is normal. If the equipment is abnormal, adjust the equipment to keep it normal. Then, the severity of the injury is assessed, and fatal injuries are treated first. For patients with spinal injuries, 3-4 people are required to coordinate and move them to a wheeled multi-position stretcher. Finally, the rescue of the patient is carried out. The specific rescue work for the patient is as follows: Unfold the wheeled multi-position stretcher horizontally, retract the four main universal wheels, ensure the stretcher is stable, and remove the storage box; The patient is moved horizontally onto the stretcher, with one person securing the head and neck, while two people support the shoulders, waist, hips and legs, and then simultaneously moved horizontally onto the wheeled multi-position stretcher. Restrain the patient's chest, hips, and knees with wide straps, tight enough to fit two fingers. Use a dedicated fixation device for the head and neck (to prevent the neck from hanging in the air). For special injury treatment: Fractured limbs: splint fixation followed by bandaging; Bleeding wound: Place on the non-pressurized side of the stretcher after applying pressure bandage; Select different movement modes to transport patients. The movement modes include: horizontal movement and vertical movement. Horizontal movement: The stretcher can be moved horizontally by lowering the four main universal wheels when lying flat, or it can be folded up and dragged on the ground; Vertical movement: The stretcher can be moved vertically in an upright position through the four small universal wheels under the footrest; After arriving at a safe area or ambulance, unlock the fixation belt, move the injured person to the receiving equipment, clean the stretcher surface, reset the equipment, and check the integrity of the accessories.

[0020] The patient's vital sign data, state data and geographical position information are acquired through the integrated first-aid platform. Further, the non-invasive blood pressure sensor based on the oscillometric method is installed on the stretcher near the patient's arm to acquire the patient's blood pressure data. The photoelectric heart rate sensor is integrated in the headrest part of the stretcher head and contacts the patient's head skin to acquire the patient's heart rate data. The blood oxygen sensor based on the PPG principle is installed on the stretcher near the patient's fingers or toes to acquire the patient's blood oxygen data. The state data is collected by the combination of high-definition cameras and depth sensors. The high-definition camera is installed above the front end of the stretcher and can rotate 360 degrees, which is used to shoot the overall and wound appearance images of the patient. The depth sensor is installed beside the camera and uses the structured light or time-of-flight (ToF) technology to acquire the three-dimensional depth information of the wound. The rescuer can manually operate the camera to aim at the wound on the scene. The device automatically analyzes the image and depth data to identify the injury type, locate the injured part, and calculate the wound area and depth. The pressure sensor array is laid on the surface of the stretcher and distributed in the areas corresponding to the easy bleeding parts of the patient's body, such as the back and hips. When blood flows out and contacts the pressure sensor, the sensor will generate an electrical signal due to the change in pressure, by which it is determined whether there is active bleeding. The GPS positioning module and Beidou positioning module are built-in the stretcher to update the patient's position in real time. In the earthquake debris rescue scene, the patient data required for calculating the Glasgow coma score is obtained by observation and inquiry, including the patient's eye opening reaction, language reaction and movement reaction. According to the patient data obtained for calculating the Glasgow coma score, the corresponding scoring standard is set, the scores corresponding to the patient's eye opening reaction, language reaction and movement reaction are added, and the Glasgow coma score is calculated. The specific scoring standard is as follows: Eye opening reaction: 4 points for spontaneous eye opening; 3 points for eye opening after hearing sound stimulus; 2 points for eye opening after giving pain stimulus; 1 point for no eye opening reaction under any stimulus; Language reaction: 5 points for answering time and place questions with normal orientation; 4 points for answering errors but communicating; 3 points for speech disorder and unable to communicate normally; 2 points for only making meaningless sounds; 1 point for no language reaction; Motor reaction: 6 points for completing actions as instructed; 5 points for locating pain stimulus; 4 points for avoiding action to pain stimulus; 3 points for limb flexion (decortical rigidity) to pain stimulus; 2 points for limb extension (decerebral rigidity) to pain stimulus; 1 point for no limb movement reaction; Synchronize the Glasgow coma score with the acquired patient sign data, state data, and patient geographic location information, combine them into a set X, the definition of set X is as follows: X={(GCS1, s1, t1, l1), (GCS2, s2, t2, l2),..., (GCS q ,s q ,t q ,l q )}, wherein GCS1, GCS2,..., GCS q represents the Glasgow coma score of the 1st, 2nd,..., qth patient, s1, s2,..., s q represents the sign data set of the 1st, 2nd,..., qth patient, t1, t2,..., t q represents the state data set of the 1st, 2nd,..., qth patient, and l1, l2,..., l q represents the geographic location information of the 1st, 2nd,..., qth patient.

[0021] Specifically, in a certain earthquake disaster area, 3 patients (A, B, and C) need to be transferred from the debris site to a field hospital 5 kilometers away. The available resources on site include: 3 stretchers, 3 medical staff, basic emergency medicines and equipment, and the minimum resource reserve threshold R0=1. Through the sensor array of the wheeled multi-position stretcher, real-time multi-modal data of the 3 patients are collected: Patient A: Sign data: blood pressure 80 / 50 mmHg, heart rate 120 beats / min, blood oxygen 85%; State data: open fracture of right lower limb (trauma area 20 cm², wound depth 3 cm, active bleeding), consciousness is unclear; Geographic location: (X1, Y1); GCS score: 9; transfer waiting time: cumulative 20 minutes; waiting times: 1 time; Patient B: Sign data: blood pressure 110 / 70 mmHg, heart rate 90 beats / min, blood oxygen 95%; State data: head trauma (no active bleeding), Glasgow coma score 13, can communicate simply; Geographic location: (X2, Y2); transfer waiting time: cumulative 5 minutes; waiting times: 1 time; Patient C: Sign data: blood pressure 70 / 40 mmHg, heart rate 140 beats / min, blood oxygen 75%; State data: combined chest and abdominal injury (trauma area 30 cm², wound depth 5 cm, active bleeding), no eye opening response, no verbal response; Geolocation: (X3, Y3); GCS score: 5; transport latency: 30 minutes cumulative; number of waiting times: 2.

[0022] The obtained relevant data is preprocessed, and the priority of emergency treatment of the patient is determined based on the preprocessed data. The specific steps include: The obtained relevant data is cleaned and preprocessed. When the data value at a certain moment is missing, a time series-based interpolation method is used to fill in the data value according to the data values at the adjacent moments, and the data format is standardized. All data is stored according to a unified database table structure; The Min-Max normalization algorithm is used to process the obtained light sensing data, and the data is uniformly mapped to the interval [0, 1]; Real-time traversal of all patient Glasgow coma scores, sorted from low to high, generating a patient priority queue. When a new patient is received, if the patient's Glasgow coma score is less than or equal to the set score threshold, it is directly represented as the highest priority and inserted at the beginning of the queue. For patients with a Glasgow coma score greater than the set score threshold, their scores are compared with the existing patient scores in the queue in turn. For patients with the same score, they are sorted by arrival time. For patients with the same score and arrival time, they are sorted by systolic blood pressure data. After insertion, the updated priority list is synchronized to the field personnel and the hospital; The obtained patient systolic blood pressure data is sorted in three steps as follows: First-level sorting: Compare the systolic blood pressure data with the normal range. Patients with systolic blood pressure below the lower limit of the normal range are sorted according to the principle of lower systolic blood pressure being more preferred; Second-level sorting: For patients with systolic blood pressure within the normal range, further compare the deviation of systolic blood pressure from the normal range. The greater the deviation, the lower the sorting; Third-level sorting: For patients with systolic blood pressure higher than the upper limit of the normal range, sort them according to the order of higher systolic blood pressure being more preferred.

[0023] Specifically, the heart rate data of patient C has abnormal values due to device jolting, which is filled in by time series interpolation method; Priority queue generation: First-level sorting (GCS score): C (5 points) < A (9 points) < B (13 points), initial queue: C→A→B Second-level sorting (arrival time): 3 patients arrived at the same time, with the same arrival time; Third-level sorting (systolic blood pressure): C's systolic blood pressure is 70 mmHg (lower than the normal lower limit of 90 mmHg, preferred); A's systolic pressure 80mmHg (below the lower limit of normal, second priority); B's systolic pressure 110mmHg (normal range (90mmHg-110mmHg)); Final priority queue: C→A→B.

[0024] The patient emergency index is measured by the waiting times and durations of the rescuers in the waiting area, and the specific steps include: The patient emergency index R is measured by the waiting times and durations of the rescuers in the waiting area j , which is defined as follows: ; Wherein, R j represents the patient j emergency index, the value range is between 0 and 1, n represents the waiting times of the rescuers in the waiting area during the transfer process, t ji represents the duration of the i-th waiting of the patient j, i represents the waiting time label, and T represents the total duration from the start of the transfer task to the end of the transfer task.

[0025] Specifically, based on the total duration of the waiting of the patient C being 30 minutes and the waiting times being 2, the patient C emergency index is 0.5 according to the emergency index calculation formula; based on the total duration of the waiting of the patient A being 20 minutes and the waiting times being 1, the patient A emergency index is 0.67 according to the emergency index calculation formula; based on the total duration of the waiting of the patient B being 5 minutes and the waiting times being 1, the patient B emergency index is 0.92 according to the emergency index calculation formula.

[0026] According to the current total amount of patient emergency demand, the total amount of on-site deployable emergency resources, and the set minimum emergency resource amount for guaranteeing basic rescue capability, the emergency resource evaluation and adjustment is carried out, and the specific steps include: When DT≤RT-R0, the total amount of on-site deployable emergency resources is sufficient for the current patient rescue work, and rescue is carried out according to the patient arrival hospital time, wherein DT represents the current total amount of patient emergency demand, RT represents the total amount of on-site deployable emergency resources, and R0 represents the set minimum emergency resource amount for guaranteeing basic rescue capability; When DT>RT-R0, the total amount of on-site deployable emergency resources cannot meet the current patient rescue work, the total resource gap amount of DT-RT+R0 is calculated, support is sought from surrounding medical units according to the calculated total resource gap amount, rescue is carried out according to the patient arrival hospital time, for the case that there is patient waiting, the above patient priority queue is sorted according to the patient emergency index, and patient rescue is carried out according to the adjusted patient priority queue; The patient priority queue is sorted according to the patient emergency index, and the specific steps are as follows: for the case where there are patients waiting, the patient queuing sequence is reordered according to the corresponding emergency index of the patient from large to small.

[0027] Specifically, the first aid demand unit of each patient is 1, the on-site allocatable resource is 3, the minimum resource threshold is 1, the resource shortage is calculated according to the total amount of first aid demand and resource supply and demand, and 1 stretcher and 1 medical staff are requested from the surrounding medical units at this time; the patient rescue is arranged according to the patient arrival time at the hospital.

[0028] A single person first aid transfer intelligent decision system based on multi-modal data, the system comprises a data acquisition module, a data preprocessing module, a priority judgment module and a resource evaluation and scheduling module, the data acquisition module is used to obtain the related data of the patient through an integrated first aid platform; the data preprocessing module is used to preprocess the obtained related data; the priority judgment module is used to sort the preprocessed data from low to high according to the score, and generate a patient priority queue; the resource evaluation and scheduling module is used to calculate the patient emergency index based on the waiting number and duration of the rescue personnel in the waiting area, to evaluate and adjust the first aid resources according to the current total amount of first aid demand, the total amount of on-site allocatable first aid resources and the minimum amount of first aid resources set to ensure basic rescue capability, for the case where the total amount of on-site allocatable first aid resources is sufficient for the current patient rescue work, rescue is carried out according to the patient arrival time at the hospital, for the case where the total amount of on-site allocatable first aid resources cannot meet the current patient rescue work, rescue is carried out according to the patient arrival time at the hospital, for the case where there are patients waiting, the patient priority queue is sorted according to the patient emergency index, and the patient rescue is carried out according to the adjusted patient priority queue.

[0029] The data acquisition module comprises a vital sign data acquisition unit, a state data acquisition unit, a geographic position acquisition unit and a consciousness state acquisition unit, the vital sign data acquisition unit is used to acquire blood pressure, heart rate and blood oxygen data of the patient; the state data acquisition unit is used to acquire the type of external injury, the injured part, the trauma area, the wound depth and whether active bleeding; the geographic position acquisition unit is used to acquire the geographic position information of the patient; the consciousness state acquisition unit is used to acquire the eye opening reaction, language reaction and movement reaction of the patient, according to the patient data required by the Glasgow coma score, the corresponding scoring standard is set, the scores of the eye opening reaction, language reaction and movement reaction of the patient are added, and the Glasgow coma score is calculated.

[0030] The data preprocessing module comprises a data filling unit and a format standardization unit, the data filling unit is used for filling in missing data values at a certain moment by using a time series-based interpolation method according to data values at adjacent moments before and after, and the format standardization unit is used for standardizing data formats and storing all data according to a unified database table structure, and a Min-Max normalization algorithm is used to process the obtained light response data and uniformly map the data to the interval [0, 1].

[0031] The priority judgment module comprises a GCS sorting unit, a second sorting unit and a third sorting unit, the GCS sorting unit is used for real-time traversal of Glasgow coma scores of all patients, sorting from low to high according to the scores, and generating a patient priority queue, when a new patient is received, judging that the Glasgow coma score of the new patient is less than or equal to a set score threshold, directly indicating the highest priority and inserting into the first position of the queue, the second sorting unit is used for comparing the score of the new patient with the scores of existing patients in the queue in turn when the Glasgow coma score of the new patient is greater than the set score threshold, and when there are patients with the same score, performing secondary sorting according to the arrival time; the third sorting unit is used for tertiary sorting in combination with the obtained systolic blood pressure data of the patient when there are patients with the same score and the same queue time.

[0032] The resource assessment and scheduling module comprises an emergency degree calculation unit, a resource supply and demand assessment unit and a priority dynamic adjustment unit, the emergency degree calculation unit is used for measuring the patient emergency degree index by the waiting number and duration of the rescuer in the waiting area; the resource supply and demand assessment unit is used for adjusting the emergency resource assessment according to the total quantity of current patient emergency demand, the total quantity of on-site deployable emergency resources and the minimum emergency resource quantity for guaranteeing basic rescue capability; the priority dynamic adjustment unit is used for rescue according to the patient arrival time when the total quantity of on-site deployable emergency resources is sufficient for the current patient rescue work, rescue according to the patient arrival time when the total quantity of on-site deployable emergency resources is insufficient, and sorting the above patient priority queue according to the patient emergency degree index when there is patient waiting, and rescuing the patients according to the adjusted patient priority queue.

[0033] It will be obvious to a person skilled in the art that the application is not limited to the details of the above-described exemplary embodiments but can be implemented in other embodiments without departing from the scope of the application. The scope of the application is defined by the appended claims rather than by the description of the exemplary embodiments above and therefore all changes and modifications that come within the meaning and range of equivalency of the claims are to be embraced by the application. Any reference signs in the claims should not be construed as limiting the claims to the figures in which the reference signs are used.

Claims

1. An intelligent decision-making method for single-person emergency transport based on multimodal data, characterized by: The method comprises the following steps: Obtain relevant patient data through the integrated emergency platform; Preprocess the acquired relevant data, sort them from low to high according to the scores based on the preprocessed data, and generate a patient priority queue; Calculate the patient urgency index based on the number and duration of waits by rescuers in the waiting area; Emergency resources are evaluated and adjusted based on the total emergency needs of current patients, the total amount of emergency resources that can be deployed on site, and the minimum amount of emergency resources set to ensure basic rescue capabilities. If the total amount of emergency resources that can be deployed on site is sufficient for the rescue of current patients, rescue will be carried out based on the time the patients arrive at the hospital. If the total amount of emergency resources that can be deployed on site cannot meet the rescue needs of current patients, rescue will be carried out based on the time the patients arrive at the hospital. In cases where there are patients waiting, the above-mentioned patient priority queues are sorted according to the patient urgency index, and patients are rescued based on the adjusted patient priority queues.

2. The intelligent decision-making method for single-person emergency transport based on multimodal data according to claim 1, characterized in that: Obtain relevant patient data through the integrated emergency platform. The specific steps include: The integrated first aid platform is used to obtain the patient's vital signs data, status data and geographic location information, wherein the vital signs data include the patient's blood pressure, heart rate and blood oxygen data, and the status data include the patient's trauma type, injured part, wound area, wound depth, and whether there is active bleeding. The integrated first aid platform includes a wheeled multi-position stretcher, a head and neck hemostasis device, a ventilation module, a fluid infusion module, a monitoring module, a fixation module, a closed drainage module and a rewarming module. The wheeled multi-position stretcher is used for the wounded to receive treatment and move during the first aid process. The device is a carrier for transport and evacuation; the head and neck hemostasis device is used to stop severe bleeding in the head and neck; the ventilation module includes a blind endotracheal intubation guide with a light source and a new type of visual laryngoscope intubation; the fluid infusion module includes a bone marrow infusion puncture gun and a satellite pressurized infusion device; the monitoring module includes an electrocardiogram monitor and a blood oxygen saturation monitor; the fixation module includes an inflatable cervical fixator and a polymer fracture fixation bandage; the closed drainage module includes a closed chest drainage device; the rewarming module is used to rewarm the human body trunk and keep the limbs warm; Obtaining patient data required for calculating the Glasgow Coma Scale, including the patient's eye opening response, language response, and motor response; setting corresponding scoring criteria based on the acquired patient data required for the Glasgow Coma Scale; and adding the patient's scores for the eye opening response, language response, and motor response to calculate the Glasgow Coma Scale; The Glasgow Coma Scale is synchronously recorded with the acquired patient vital sign data, status data and patient geographic location information, and combined into a set X. The definition of set X is as follows: X={(GCS1,s1,t1,l1),(GCS2,s2,t2,l2),...,(GCS q ,s q ,t q ,l q )}, where GCS1, GCS2, ..., GCS q represents the Glasgow Coma Scale score of the 1st, 2nd, ..., qth patients, s1, s2, ..., s q represents the set of vital sign data of the 1st, 2nd, ..., qth patients, t1, t2, ..., t q represents the state data set of the 1st, 2nd, ..., qth patients, l1, l2, ..., l q Represents the geographic location information of the 1st, 2nd, ..., qth patients.

3. The intelligent decision-making method for single-person emergency transport based on multimodal data according to claim 2, characterized in that: The acquired relevant data is pre-processed, and the priority of emergency treatment of patients is determined based on the pre-processed data. The specific steps include: Clean and pre-process the acquired data. When data values ​​at a certain moment are missing, use time series interpolation to fill in the missing values ​​based on the data values ​​at the previous and next moments. Standardize the data format and store all data in a unified database table structure. The Min-Max normalization algorithm is used to process the acquired light sensing data and uniformly map the data to the [0,1] interval; The Glasgow Coma Scale (GCS) of all patients is traversed in real time, and they are sorted from low to high according to the scores to generate a patient priority queue. When a new patient is received and the Glasgow Coma Scale (GCS) of the new patient is determined to be less than or equal to the set score threshold, it is directly indicated as the highest priority and inserted at the top of the queue. If the Glasgow Coma Scale (GCS) of the new patient is greater than the set score threshold, its score is compared with the scores of existing patients in the queue in sequence. If there are patients with the same score, they are sorted in the second level by arrival time. If there are patients with the same score and the same entry time, they are sorted in the third level based on the systolic blood pressure data of the patients. After the insertion is completed, the updated priority list is synchronized to the on-site personnel and the hospital.

4. The intelligent decision-making method for single-person emergency transport based on multimodal data according to claim 3, characterized in that: The patient urgency indicator is measured by the number and duration of waiting times in the waiting area by rescuers. The specific steps include: The patient urgency index R is measured by the number and duration of waiting of rescuers in the waiting area. j , defined as follows: ； Among them, R j represents the urgency index of patient j, with a value range of 0 to 1, n represents the number of times the rescuer waits in the waiting area during the transfer process, t ji represents the duration of patient j's i-th waiting time, i represents the waiting number label, and T represents the total time from the start to the end of the transfer task.

5. The intelligent decision-making method for single-person emergency transport based on multimodal data according to claim 4, characterized in that: Evaluate and adjust emergency resources based on the total number of current patient emergency needs, the total number of emergency resources available on site, and the minimum number of emergency resources required to ensure basic rescue capabilities. The specific steps include: When DT≤RT-R0, the total amount of emergency resources available on site is sufficient to rescue the current patient, and rescue is carried out based on the patient's arrival time at the hospital. DT represents the total amount of emergency resources currently needed by the patient, RT represents the total amount of emergency resources available on site, and R0 represents the minimum amount of emergency resources set to ensure basic rescue capabilities. When DT>RT-R0, the total amount of emergency resources available on site cannot meet the rescue work of the current patients. The total resource gap of DT-RT+R0 is calculated, and support is sought from surrounding medical units based on the calculated total resource gap. Rescue is carried out based on the time the patient arrives at the hospital. If there are patients waiting, the patient priority queue is sorted according to the patient urgency index, and the patient is rescued according to the adjusted patient priority queue; The above patient priority queues are sorted according to the patient urgency index. The specific steps are as follows: if there are patients waiting, the patient queue sequence is re-sorted from large to small according to the patient's corresponding urgency index.

6. An intelligent decision-making system for single-person emergency transport based on multimodal data, characterized by: The system includes a data acquisition module, a data preprocessing module, a priority judgment module and a resource evaluation and scheduling module. The data acquisition module is used to obtain patient-related data through an integrated emergency platform; the data preprocessing module is used to preprocess the obtained relevant data; the priority judgment module is used to sort the preprocessed data from low to high according to the score to generate a patient priority queue; the resource evaluation and scheduling module is used to calculate the patient urgency index based on the number and duration of waiting of rescue personnel in the waiting area, and evaluate and adjust the emergency resources according to the total amount of current patient emergency needs, the total amount of emergency resources that can be deployed on site and the minimum amount of emergency resources set to ensure basic rescue capabilities. If the total amount of emergency resources that can be deployed on site is sufficient for the rescue of the current patient, rescue is carried out according to the time the patient arrives at the hospital. If the total amount of emergency resources that can be deployed on site cannot meet the rescue of the current patient, rescue is carried out according to the time the patient arrives at the hospital. In the case of patients waiting, the above-mentioned patient priority queues are sorted according to the patient urgency index, and the patients are rescued according to the adjusted patient priority queues.

7. The intelligent decision-making system for single-person emergency transport based on multimodal data according to claim 6, characterized in that: The data acquisition module includes a vital sign data acquisition unit, a state data acquisition unit, a geographic location acquisition unit, and a consciousness state acquisition unit. The vital sign data acquisition unit is used to obtain the patient's blood pressure, heart rate, and blood oxygen data; the state data acquisition unit is used to obtain the patient's trauma type, injury site, wound area, wound depth, and whether there is active bleeding; the geographic location acquisition unit is used to obtain the patient's geographic location information; and the consciousness state acquisition unit is used to obtain the patient's eye opening reaction, language reaction, and motor reaction. According to the obtained patient data required for the Glasgow Coma Scale, corresponding scoring standards are set, and the scores corresponding to the patient's eye opening reaction, language reaction, and motor reaction are added to calculate the Glasgow Coma Scale.

8. The intelligent decision-making system for single-person emergency transport based on multimodal data according to claim 7, characterized in that: The data preprocessing module includes a data filling unit and a format standardization unit. The data filling unit is used to fill in the missing data value at a certain moment using a time series-based interpolation method based on the data values ​​of the previous and next adjacent moments. The format standardization unit is used to standardize the data format, store all data according to a unified database table structure, and use the Min-Max normalization algorithm to process the acquired light sensing data and uniformly map the data to the [0,1] interval.

9. The intelligent decision-making system for single-person emergency transport based on multimodal data according to claim 8, characterized in that: The priority determination module includes a GCS sorting unit, a second sorting unit, and a third sorting unit. The GCS sorting unit is used to traverse the Glasgow Coma Scale (GCS) scores of all patients in real time, sort them from low to high according to the scores, and generate a patient priority queue. When a new patient is received and the Glasgow Coma Scale score of the new patient is determined to be less than or equal to a set score threshold, the patient is directly designated as the highest priority and inserted at the top of the queue. The second sorting unit is used to compare the score of the new patient with the scores of existing patients in the queue in sequence when the Glasgow Coma Scale score of the new patient is greater than the set score threshold. If there are patients with the same score, the patients are sorted in a second level according to their arrival time. The third sorting unit is used to perform a tertiary sorting based on the acquired systolic blood pressure data of the patients when there are patients with the same score and the same entry time.

10. The intelligent decision-making system for single-person emergency transport based on multimodal data according to claim 9, characterized in that: The resource assessment and scheduling module includes an urgency calculation unit, a resource supply and demand assessment unit, and a priority dynamic adjustment unit. The urgency calculation unit is used to measure the patient urgency index by the number of times and duration that rescuers wait in the waiting area; the resource supply and demand assessment unit is used to assess and adjust emergency resources based on the current total emergency demand of patients, the total amount of emergency resources that can be deployed on site, and the set minimum amount of emergency resources to ensure basic rescue capabilities; the priority dynamic adjustment unit is used to rescue the current patient based on the time the patient arrives at the hospital if the total amount of emergency resources that can be deployed on site is sufficient, and to rescue the patient based on the time the patient arrives at the hospital if the total amount of emergency resources that can be deployed on site is insufficient. In the case where there are patients waiting, the above-mentioned patient priority queues are sorted according to the patient urgency index, and the patients are rescued based on the adjusted patient priority queues.

Images