CN116597971B

CN116597971B - Digital twinning-based hospital space optimization simulation method and system

Info

Publication number: CN116597971B
Application number: CN202310875300.3A
Authority: CN
Inventors: 詹海波; 张雪; 郝瑞; 詹海浪; 岳可新
Original assignee: Shandong Xinrui Information Technology Co ltd
Current assignee: Shandong Xinrui Information Technology Co ltd
Priority date: 2023-07-18
Filing date: 2023-07-18
Publication date: 2023-09-19
Anticipated expiration: 2043-07-18
Also published as: CN116597971A

Abstract

The invention provides a hospital space optimization simulation method and system based on digital twinning, and relates to the technical field of artificial intelligence. In the invention, a plurality of actions to be analyzed in hospital space are collected; determining a first class object distribution network based on a plurality of actions to be analyzed in hospital space; performing map key information mining operation on the first class object distribution network to output a first class distribution network description vector; based on the first type of distribution network description vector, obtaining action identification data corresponding to actions to be analyzed in each hospital space; based on the motion identification data, determining the space priority parameters of the simulated medical space, and screening out at least one target simulated medical space based on the space priority parameters after obtaining the space priority parameters of a plurality of simulated medical spaces, wherein the target simulated medical space is used as a medical space model optimization result of the target actual medical space. Based on the above, the reliability of the hospital space optimization simulation can be improved to a certain extent.

Description

Digital twinning-based hospital space optimization simulation method and system

Technical Field

The invention relates to the technical field of artificial intelligence, in particular to a hospital space optimization simulation method and system based on digital twinning.

Background

Digital Twin: the method fully utilizes data such as a physical model, sensor update, operation history and the like, integrates simulation processes of multiple disciplines, multiple physical quantities, multiple scales and multiple probabilities, and completes mapping in a virtual space, thereby reflecting the full life cycle process of corresponding entity equipment. Digital Twin is a beyond-the-reality concept that can be seen as a Digital mapping system of one or more important, mutually dependent equipment systems. For example, the medical space may be simulated based on digital twinning techniques, so that corresponding optimization operations and the like may be facilitated. However, the prior art has a problem that the reliability of the space optimization simulation in the hospital is not high.

Disclosure of Invention

In view of the above, the present invention aims to provide a digital twin-based hospital space optimization simulation method and system, so as to improve reliability of the hospital space optimization simulation to a certain extent.

In order to achieve the above purpose, the embodiment of the present invention adopts the following technical scheme:

a hospital space optimization simulation method based on digital twinning comprises the following steps:

collecting a plurality of hospital space to-be-analyzed actions, wherein the hospital space to-be-analyzed actions are used for reflecting hospital space actions output by patient staff to medical staff, each hospital space to-be-analyzed action is recorded through corresponding text data or image data, and the plurality of hospital space to-be-analyzed actions are formed by performing operation simulation on a simulated medical space;

Determining a first type of object distribution network based on the actions to be analyzed of the plurality of hospital spaces, wherein the first type of object distribution network comprises patient side personnel distribution objects corresponding to the patient side personnel, medical side personnel distribution objects corresponding to the medical side personnel and distribution object association lines for associating the patient side personnel distribution objects with the medical side personnel distribution objects;

performing map key information mining operation on the first type object distribution network to output a first type distribution network description vector corresponding to the first type object distribution network, wherein the first type distribution network description vector comprises patient side personnel key information corresponding to each patient side personnel distribution object, medical side personnel key information corresponding to each medical side personnel distribution object and hospital space action key information corresponding to each distribution object association line;

based on the first type of distribution network description vector, respectively carrying out action abnormality recognition operation on the hospital space to-be-analyzed actions corresponding to each distribution object association line to obtain action recognition data corresponding to each hospital space to-be-analyzed action, wherein the action recognition data are used for reflecting whether the hospital space to-be-analyzed actions belong to suspicious actions or the possibility of the suspicious actions;

Determining a space priority parameter of the simulated medical space based on action identification data corresponding to actions to be analyzed of each hospital space, and screening out at least one target simulated medical space from the simulated medical spaces based on the space priority parameter after obtaining the space priority parameters of the simulated medical spaces, wherein the target actual medical space belongs to the space to be built as a medical space model optimization result of the target actual medical space.

In some preferred embodiments, in the above-mentioned digital twin-based hospital space optimization simulation method, after the step of acquiring a plurality of hospital space actions to be analyzed, the digital twin-based hospital space optimization simulation method further includes:

determining a second class object distribution network based on the actions to be analyzed in the plurality of hospital spaces, wherein the second class object distribution network comprises action class distribution objects and action object association lines;

performing map key information mining operation on the second class object distribution network, and outputting group space motion description vectors corresponding to the motion class distribution objects, wherein the group space motion description vectors belong to motion key information for performing group motion;

The step of performing an action anomaly identification operation on the actions to be analyzed in the hospital space corresponding to the correlation lines of the distributed objects based on the first type of distribution network description vectors to obtain action identification data corresponding to the actions to be analyzed in the hospital space, includes:

based on the first type of distribution network description vector and each group space motion description vector, performing motion anomaly identification operation on the motion to be analyzed in the hospital space corresponding to each distribution object association line to obtain motion identification data corresponding to each motion to be analyzed in the hospital space.

In some preferred embodiments, in the above-mentioned digital twin-based hospital space optimization simulation method, the group space motion description vector is obtained by performing a map key information mining operation by using a key information analysis network;

the network optimization process of the key information analysis network comprises the following steps:

extracting a typical second-class object distribution network, wherein the typical second-class object distribution network comprises typical suspicious action class distribution objects of suspicious hospital space actions and typical action object association lines for associating the two typical action class distribution objects;

performing key information mining operation on each typical suspicious action class distribution object in the typical second class object distribution network to output a typical group action description vector corresponding to each typical suspicious action class distribution object, wherein the typical group action description vector refers to a typical group space action description vector;

According to the typical group motion description vector of each typical suspicious motion class distribution object, analyzing typical motion identification data corresponding to each typical suspicious motion class distribution object by utilizing a candidate key information analysis network;

and carrying out parameter optimization adjustment operation on the candidate key information analysis network based on the typical action identification data corresponding to each typical suspicious action class distribution object and each typical suspicious action class distribution object so as to form a key information analysis network corresponding to the candidate key information analysis network.

In some preferred embodiments, in the above-mentioned digital twinning-based hospital space optimization simulation method, the step of performing a key information mining operation on each of the typical suspicious action class distribution objects in the typical second class object distribution network to output a typical group action description vector corresponding to each of the typical suspicious action class distribution objects includes:

determining a typical global adjacent relation array corresponding to the typical second class object distribution network, and determining a typical local adjacent relation array corresponding to each typical suspicious action class distribution object;

performing adjacent analysis operation on each typical suspicious action class distribution object in the typical second class object distribution network, and analyzing to form a typical suspicious adjacent object cluster corresponding to each typical suspicious action class distribution object, wherein the typical suspicious adjacent object cluster comprises typical suspicious adjacent distribution objects which are associated with the typical action object association line between the corresponding typical suspicious action class distribution objects;

Performing reinforcement operation on the typical local adjacent relation array by using each typical suspicious adjacent object cluster to form a reinforced local adjacent relation array corresponding to each typical suspicious action type distribution object;

performing interval mapping operation of parameters on the reinforced local adjacent relation arrays corresponding to the typical suspicious action type distribution objects to form mapped local adjacent relation arrays corresponding to the typical suspicious action type distribution objects;

and carrying out analysis operation of suspicious parameters based on each mapping local adjacent relation array to form typical group action description vectors corresponding to each typical suspicious action class distribution object.

In some preferred embodiments, in the above-mentioned digital twinning-based hospital space optimization simulation method, the step of performing a graph key information mining operation on the first type object distribution network to output a first type distribution network description vector corresponding to the first type object distribution network includes:

performing object key information mining operation on each patient side personnel distribution object and each medical side personnel distribution object in the first class object distribution network to respectively form corresponding object key information description vectors;

Performing associated line key information mining operation on each distributed object associated line in the first class object distribution network to respectively form corresponding associated line key information description vectors;

and carrying out data merging operation based on the first type object distribution network, each object key information description vector and each associated line key information description vector, and outputting a first type distribution network description vector corresponding to the first type object distribution network.

In some preferred embodiments, in the above-mentioned digital twinning-based hospital space optimization simulation method, the step of performing a data merging operation based on the first-class object distribution network, each of the object key information description vectors, and each of the associated line key information description vectors, and outputting a first-class distribution network description vector corresponding to the first-class object distribution network includes:

determining adjacent relation characterization parameter distribution corresponding to the first class object distribution network, wherein the adjacent relation characterization parameter distribution is used for reflecting adjacent information among distribution objects;

performing parameter analysis operation on the adjacent relation representation parameter distribution, and analyzing and outputting patient relation representation parameter distribution corresponding to each patient side personnel distribution object;

Performing data merging operation based on the object key information description vector of the current iteration number of each patient relation characterization parameter distribution and each patient side personnel distribution object to respectively form the object key information description vector of the next iteration number of each patient side personnel distribution object;

if the current iteration number is equal to the distribution network characterization parameter of the first type object distribution network, marking an object key information description vector of the next iteration number of each patient side personnel distribution object to be marked as each patient side personnel key information in the first type distribution network description vector, wherein the distribution network characterization parameter of the first type object distribution network has a correlation with the number of distribution object association lines in the first type object distribution network;

and if the current iteration number is equal to one, the object key information description vector of the patient side personnel distribution object with the iteration number equal to one is the object key information description vector of the patient side personnel distribution object.

In some preferred embodiments, in the above-mentioned digital twinning-based hospital space optimization simulation method, the step of performing a data merging operation based on the first-class object distribution network, each of the object key information description vectors, and each of the associated line key information description vectors, and outputting a first-class distribution network description vector corresponding to the first-class object distribution network further includes:

Performing parameter analysis operation on the adjacent relation characterization parameter distribution, and analyzing and outputting medical relation characterization parameter distribution corresponding to each medical party personnel distribution object;

performing data merging operation based on the object key information description vector of the current iteration number of each medical relation characterization parameter distribution and each medical staff distribution object to respectively form the object key information description vector of the next iteration number of each medical staff distribution object;

if the current iteration number is equal to the distribution network characterization parameter of the first type of object distribution network, marking an object key information description vector of the next iteration number of each medical party personnel distribution object to be marked as each medical party personnel key information in the first type of distribution network description vector;

and if the current iteration number is equal to one, the object key information description vector of the medical staff distribution object with the iteration number equal to one is the object key information description vector of the medical staff distribution object.

Performing parameter analysis operation on the adjacent relation representation parameter distribution, and analyzing and outputting line adjacent relation representation parameter distribution corresponding to the associated line of each distribution object;

performing data merging operation based on the line adjacent relation characterization parameter distribution and the associated line key information description vector of the current iteration number of the associated line of each distribution object so as to respectively form the associated line key information description vector of the next iteration number of the associated line of each distribution object;

if the current iteration number is equal to the distribution network characterization parameter of the first type object distribution network, marking the associated line key information description vector of the next iteration number of the associated line of each distribution object to be marked as each hospital space action key information in the first type distribution network description vector;

and if the current iteration number is equal to one, the associated line key information description vector of which the iteration number of the associated line of the distributed object is equal to one is the associated line key information description vector of the associated line of the distributed object.

In some preferred embodiments, in the above-mentioned digital twinning-based hospital space optimization simulation method, the step of performing an action anomaly identification operation on the hospital space to-be-analyzed actions corresponding to the distribution object association lines based on the first type of distribution network description vectors to obtain action identification data corresponding to the hospital space to-be-analyzed actions includes:

Based on the patient side personnel key information, the medical side personnel key information and the hospital space action key information, performing evaluation operation of suspicious hospital space actions on the distribution object association lines, and outputting possibility parameters of the distribution object association lines as corresponding lines of suspicious hospital space actions;

and obtaining action identification data corresponding to the actions to be analyzed in the hospital space based on the possibility parameters that the associated lines of the distributed objects are the corresponding lines of the suspicious hospital space actions.

The embodiment of the invention also provides a digital twin-based hospital space optimization simulation system, which comprises a processor and a memory, wherein the memory is used for storing a computer program, and the processor is used for executing the computer program so as to realize the digital twin-based hospital space optimization simulation method.

The hospital space optimization simulation method and system based on digital twinning provided by the embodiment of the invention can collect a plurality of hospital space actions to be analyzed first; determining a first class object distribution network based on a plurality of actions to be analyzed in hospital space; performing map key information mining operation on the first class object distribution network to output a first class distribution network description vector; based on the first type of distribution network description vector, obtaining action identification data corresponding to actions to be analyzed in each hospital space; based on the motion identification data, determining the space priority parameters of the simulated medical space, and screening out at least one target simulated medical space based on the space priority parameters after obtaining the space priority parameters of a plurality of simulated medical spaces, wherein the target simulated medical space is used as a medical space model optimization result of the target actual medical space. Based on the foregoing, because the operation of identifying the motion abnormality of the motion to be analyzed in the hospital space corresponding to the association line of each distribution object is performed based on the first type of distribution network description vector, that is, the motion to be analyzed in the hospital space is independently analyzed, the reliability of the motion identification data corresponding to the motion to be analyzed in each obtained hospital space is higher, so that the reliability of the simulated medical space screening based on the motion identification data can be improved to a certain extent, and the reliability of the optimized simulation in the hospital space is improved.

In order to make the above objects, features and advantages of the present invention more comprehensible, preferred embodiments accompanied with figures are described in detail below.

Drawings

Fig. 1 is a block diagram of a digital twin-based hospital space optimization simulation system according to an embodiment of the present invention.

Fig. 2 is a schematic flow chart of steps included in the digital twin-based hospital space optimization simulation method according to an embodiment of the present invention.

Fig. 3 is a schematic diagram of each module included in the digital twin-based hospital space optimization simulation device according to the embodiment of the present invention.

Detailed Description

For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is apparent that the described embodiments are only some embodiments of the present invention, but not all embodiments of the present invention. The components of the embodiments of the present invention generally described and illustrated in the figures herein may be arranged and designed in a wide variety of different configurations.

As shown in FIG. 1, an embodiment of the present invention provides a digital twinning-based hospital space optimization simulation system, which may include a memory and a processor.

In detail, the memory and the processor are electrically connected directly or indirectly to realize transmission or interaction of data. For example, electrical connection may be made to each other via one or more communication buses or signal lines. The memory may store at least one software functional module (computer program) that may exist in the form of software or firmware. The processor may be configured to execute the executable computer program stored in the memory, so as to implement the digital twin-based hospital space optimization simulation method provided by the embodiment of the present invention.

It should be appreciated that in some possible embodiments, the Memory may be, but is not limited to, random access Memory (Random Access Memory, RAM), read Only Memory (ROM), programmable Read Only Memory (Programmable Read-Only Memory, PROM), erasable Read Only Memory (Erasable Programmable Read-Only Memory, EPROM), electrically erasable Read Only Memory (Electric Erasable Programmable Read-Only Memory, EEPROM), and the like. The processor may be a general purpose processor including a central processing unit (Central Processing Unit, CPU), a network processor (Network Processor, NP), a System on Chip (SoC), etc.; but also Digital Signal Processors (DSPs), application Specific Integrated Circuits (ASICs), field Programmable Gate Arrays (FPGAs) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

It should be appreciated that in some possible embodiments, the digital twinning-based hospital space optimization simulation system may be a server with data processing capabilities.

The embodiment of the invention also provides a digital twin-based hospital space optimization simulation method which can be applied to the digital twin-based hospital space optimization simulation system. The digital twin-based hospital space optimization simulation method can comprise the following steps:

determining a first type of object distribution network based on the collected actions to be analyzed in a plurality of hospital spaces; performing map key information mining operation on the first type object distribution network to output a first type distribution network description vector corresponding to the first type object distribution network; based on the first type of distribution network description vector, performing action abnormality recognition operation to obtain action recognition data corresponding to actions to be analyzed in the hospital space; after motion identification data corresponding to the motion to be analyzed of each hospital space in each of a plurality of simulated medical spaces is obtained, at least one target simulated medical space is screened out of the plurality of simulated medical spaces based on the motion identification data, and the target simulated medical space is used as a medical space model optimization result of the target actual medical space.

It should be appreciated that, in some possible embodiments, the step of determining the first type of object distribution network based on the acquired plurality of actions to be analyzed in the hospital space may include:

collecting a plurality of hospital space to-be-analyzed actions, wherein the hospital space to-be-analyzed actions are used for reflecting hospital space actions output by patient staff to medical staff, each hospital space to-be-analyzed action is recorded through corresponding text data or image data, and the plurality of hospital space to-be-analyzed actions are formed by performing operation simulation on a simulated medical space; based on the actions to be analyzed in the hospital spaces, a first type of object distribution network is determined, wherein the first type of object distribution network comprises patient side personnel distribution objects corresponding to patient side personnel, medical side personnel distribution objects corresponding to medical side personnel and distribution object association lines for associating the patient side personnel distribution objects with the medical side personnel distribution objects.

It should be understood that, in some possible embodiments, after obtaining the motion recognition data corresponding to the motion to be analyzed in each hospital space of each of the plurality of simulated medical spaces, at least one target simulated medical space is screened out of the plurality of simulated medical spaces based on the motion recognition data, so as to be a medical space model optimization result of a target actual medical space, where the target actual medical space belongs to a space to be built, and the method includes:

It should be understood that, in some possible embodiments, the specific details of the steps of the above-described digital twin-based hospital space optimization simulation method may also be referred to in the following description.

With reference to fig. 2, the embodiment of the invention also provides another digital twin-based hospital space optimization simulation method, which can be applied to the digital twin-based hospital space optimization simulation system. The method steps defined by the flow related to the digital twin-based hospital space optimization simulation method can be realized by the digital twin-based hospital space optimization simulation system.

The specific flow shown in fig. 2 will be described in detail.

Step S110, collecting a plurality of actions to be analyzed in the hospital space.

In the embodiment of the invention, the digital twin-based hospital space optimization simulation system can acquire a plurality of hospital space actions to be analyzed. The hospital space actions to be analyzed are used for reflecting hospital space actions output by patient side personnel to medical side personnel, each hospital space action to be analyzed is recorded through corresponding text data or image data, the hospital space actions to be analyzed are formed by operating simulation on a simulated medical space, and the hospital space actions to be analyzed belong to one or more patient side personnel in the simulated medical space and are formed by outputting actions to one or more medical side personnel.

And step S120, determining a first type object distribution network based on the plurality of hospital space actions to be analyzed.

In the embodiment of the invention, the digital twin-based hospital space optimization simulation system can determine the first class object distribution network based on the plurality of hospital space actions to be analyzed. The first class object distribution network comprises patient side personnel distribution objects corresponding to the patient side personnel, medical side personnel distribution objects corresponding to the medical side personnel, and distribution object association lines for associating the patient side personnel distribution objects with the medical side personnel distribution objects. The distribution object association line corresponds to the hospital space to-be-analyzed action, that is, a distribution object association line associated between a patient side personnel distribution object and a medical side personnel distribution object corresponds to a hospital space to-be-analyzed action output by a patient side personnel corresponding to the patient side personnel distribution object to a medical side personnel corresponding to the medical side personnel distribution object, for example, a hospital space to-be-analyzed action a is output by a patient side personnel 1 to a medical side personnel 2, and then a distribution object association line a corresponding to the hospital space to-be-analyzed action a is associated (i.e., connected) between the patient side personnel distribution object 1 corresponding to the patient side personnel 1 and the medical side personnel distribution object 2 corresponding to the medical side personnel 2. In addition, the object attribute information is different between the patient side personnel distribution object and the medical side personnel distribution object, for example, the relevant attribute information corresponding to the patient side personnel is described by the patient side personnel distribution object, which may include identity information of the patient side personnel (patient, patient family, etc.), case information of the patient side personnel, and the like. The medical staff member distribution object is used for describing relevant attribute information corresponding to medical staff members, and may include identity information of medical staff members (doctors, nurses, etc.), occupation information of medical staff members, etc. And, the attribute information of the association line of the distributed object may be text data or image data corresponding to the action to be analyzed in the hospital space.

And step S130, carrying out map key information mining operation on the first-class object distribution network so as to output a first-class distribution network description vector corresponding to the first-class object distribution network.

In the embodiment of the invention, the digital twinning-based hospital space optimization simulation system can perform the map key information mining operation on the first-class object distribution network so as to output the first-class distribution network description vector corresponding to the first-class object distribution network. The first type of distribution network description vector comprises patient side personnel key information (which can be expressed in a vector form) corresponding to each patient side personnel distribution object, medical side personnel key information (which can be expressed in a vector form) corresponding to each medical side personnel distribution object and hospital space action key information (which can be expressed in a vector form) corresponding to each distribution object association line.

Step S140, based on the first type of distribution network description vector, performing an action anomaly identification operation on the actions to be analyzed in the hospital space corresponding to each of the distribution object association lines, so as to obtain action identification data corresponding to the actions to be analyzed in the hospital space.

In the embodiment of the invention, the digital twinning-based hospital space optimization simulation system can respectively perform the action anomaly identification operation on the hospital space to-be-analyzed actions corresponding to the distribution object association lines based on the first type of distribution network description vectors to obtain the action identification data corresponding to the hospital space to-be-analyzed actions. The action recognition data are used for reflecting whether the action to be analyzed in the hospital space belongs to suspicious actions or the possibility of the suspicious actions. The specific type of suspicious actions can be defined according to actual requirements, for example, suspicious action tags can be set for learning.

Step S150, determining a spatial priority parameter of the simulated medical space based on the motion identification data corresponding to the motion to be analyzed of each hospital space, and screening out at least one target simulated medical space from the simulated medical spaces based on the spatial priority parameter after obtaining the spatial priority parameters of the simulated medical spaces, so as to serve as a medical space model optimization result of the target actual medical space.

In the embodiment of the invention, the digital twin-based hospital space optimization simulation system may determine the space priority parameters of the simulated medical space based on the motion identification data corresponding to the motion to be analyzed of each hospital space, and screen out at least one target simulated medical space from the simulated medical spaces based on the space priority parameters after obtaining the space priority parameters of the simulated medical spaces, so as to serve as the medical space model optimization result of the target actual medical space. The target actual medical space belongs to the space to be built. For example, the motion recognition data represents the likelihood of being a suspicious motion, so that the average value or the maximum value of the motion recognition data corresponding to the motion to be analyzed in each hospital space is calculated in a suspicious manner, then the negative correlation coefficient of the average value or the maximum value is used as the space priority parameter, and finally one or more simulated medical spaces with the largest space priority parameter can be used as target simulated medical spaces. Or when the motion identification data represents whether the motion is suspicious, the number ratio of the motion to be analyzed in the hospital space which is not suspicious can be determined, and then, the coefficient of positive correlation is determined based on the number ratio to be used as the space priority parameter.

Based on the foregoing, because the operation of identifying the motion abnormality of the motion to be analyzed in the hospital space corresponding to the association line of each distribution object is performed based on the first type of distribution network description vector, that is, the motion to be analyzed in the hospital space is independently analyzed, the reliability of the motion identification data corresponding to the motion to be analyzed in each obtained hospital space is higher, so that the reliability of the simulated medical space screening based on the motion identification data can be improved to a certain extent, the reliability of the optimized simulation in the hospital space is improved, and the problem of low reliability in the prior art is solved.

It should be understood that, in some possible embodiments, step S130 in the foregoing description, that is, the step of performing the map key information mining operation on the first type of object distribution network to output the first type of distribution network description vector corresponding to the first type of object distribution network may include the following in some specific applications:

performing object key information mining operation on each patient side personnel distribution object and each medical side personnel distribution object in the first class object distribution network to respectively form corresponding object key information description vectors, wherein the object key information mining operation can refer to mining of key information of attribute information of a distributed object, namely feature mining, and can refer to the related prior art;

Performing associated line key information mining operation on each distributed object associated line in the first class object distribution network to respectively form corresponding associated line key information description vectors, wherein the associated line key information mining operation can refer to mining of key information, namely feature mining, of attribute information of the distributed object associated line, and can refer to the related prior art;

performing data merging operation based on the first type object distribution network, each object key information description vector and each associated line key information description vector, and outputting a first type distribution network description vector corresponding to the first type object distribution network; illustratively, the first type of distribution network description vector may include patient side personnel key information corresponding to each patient side personnel distribution object, medical side personnel key information corresponding to each medical side personnel distribution object, and hospital space action key information corresponding to each distribution object association line; wherein, the patient side personnel key information may include: the object key information description vector of the patient side personnel distribution object, the object key information description vector of the distribution object of which the distribution object association line is associated with the patient side personnel distribution object, may be a distribution object associated with the patient side personnel distribution object through one distribution object association line, may also be a distribution object associated with the patient side personnel distribution object through a plurality of distribution object association lines, and the patient side personnel key information may also include hospital space action key information corresponding to the distribution object association line connected with the patient side personnel distribution object, and the like; wherein, the medical staff key information may include: the object key information description vector of the medical staff distribution object and the object key information description vector of the distribution object, which is related to the distribution object association line between the medical staff distribution objects, are distributed objects, which are related to the distribution object association line between the medical staff distribution objects, and can be distributed objects related to the medical staff distribution object through one distribution object association line, can also be distributed objects related to the medical staff distribution object through a plurality of distribution object association lines, and the medical staff key information can also comprise hospital space action key information corresponding to the distribution object association line connected with the medical staff distribution object; the hospital space action key information may include: the association line key information description vector of the association line of the distribution object association line, the association line key information description vector of the association line having a correlation with the distribution object association line, and the association line having a correlation with the distribution object association line may be an association line of a common distribution object connected with the distribution object association line, and the hospital space action key information may further include patient side personnel key information corresponding to a patient side personnel distribution object connected with the distribution object association line and medical side personnel key information corresponding to a connected medical side personnel distribution object.

It should be understood that, in some possible embodiments, the step of performing the data merging operation based on the first type of object distribution network, each of the object key information description vectors and each of the associated line key information description vectors, and outputting the first type of distribution network description vector corresponding to the first type of object distribution network may include the following in some specific applications:

determining an adjacent relation characterization parameter distribution corresponding to the first type of object distribution network, where the adjacent relation characterization parameter distribution is used to reflect adjacent information between distribution objects, and for example, parameters in the adjacent relation characterization parameter distribution may include a first value and a second value, each parameter may characterize whether two distribution objects in the first type of object distribution network are adjacent, that is, whether a distribution object association line is associated, if the distribution object association line is associated, a corresponding parameter may be the first value, no distribution object association line is associated, and a corresponding parameter may be the second value;

Performing a data merging operation based on the object key information description vector of the current iteration number of each patient-side personnel distribution object and each patient-side personnel distribution object to respectively form the object key information description vector of the next iteration number of each patient-side personnel distribution object, for example, performing a data merging operation based on the object key information description vector calculated based on the first iteration of each patient-side personnel distribution object and each patient-side personnel distribution parameter distribution to respectively form the object key information description vector calculated by the second iteration of each patient-side personnel distribution object; performing data merging operation based on the object key information description vectors calculated by the second iteration of each patient relation characterization parameter distribution and each patient side personnel distribution object to respectively form object key information description vectors calculated by the third iteration of each patient side personnel distribution object;

if the current iteration number is equal to the distribution network characterization parameter of the first type of object distribution network, marking an object key information description vector of the next iteration number of each patient side personnel distribution object to be marked as each patient side personnel key information in the first type of distribution network description vector, namely forming each patient side personnel key information in the first type of distribution network description vector, wherein a correlation exists between a distribution network characterization parameter of the first type of object distribution network and the number of distribution object association lines in the first type of object distribution network, for example, the number can be directly determined as the distribution network characterization parameter of the first type of object distribution network;

It should be understood that, in some possible embodiments, the neighboring relationship characterizing parameter distribution may be a matrix, and the step of performing a parameter analysis operation on the neighboring relationship characterizing parameter distribution to analyze and output the patient relationship characterizing parameter distribution corresponding to each patient side personnel distribution object may include the following in some specific applications:

determining a parameter distribution to be processed with consistent size based on the size of the adjacent relation characterization parameter distribution, and performing parameter assignment processing on the parameter distribution to be processed based on the corresponding distribution coordinates of each patient side personnel distribution object in the adjacent relation characterization parameter distribution to form a corresponding target parameter distribution, wherein in the target parameter distribution, the parameter of the corresponding distribution coordinates of each patient side personnel distribution object can be equal to 1, and the parameter of other distribution coordinates can be equal to 0;

and carrying out phase multiplication operation on the adjacent relation representation parameter distribution and the target parameter distribution to form patient relation representation parameter distribution corresponding to each patient side personnel distribution object.

It should be understood, however, that in some possible embodiments, the step of performing a data merging operation based on the object key information description vector of the current iteration number of each patient-side personnel distribution object and each patient-side personnel distribution object to form the object key information description vector of the next iteration number of each patient-side personnel distribution object respectively may include the following in some specific applications:

performing an nth iteration calculation for a patient side personnel distribution object includes:

obtaining an intermediate description vector calculated by the patient side personnel distribution object in the n-1 th iterative computation, wherein n is an integer greater than or equal to 2, and the intermediate description vector calculated by the patient side personnel distribution object in the 1 st iterative computation is a corresponding object key information description vector;

performing multiplication operation on the intermediate description vector calculated by the patient side personnel distribution object in the n-1 th iterative computation based on a predetermined first parameter mapping matrix, and performing parameter interval mapping operation on the result of the multiplication operation, for example, the parameter mapping operation interval 0-1 is included, so as to form the intermediate description vector calculated by the patient side personnel distribution object in the n-th iterative computation, wherein the first parameter mapping matrix can be used as a network parameter of a corresponding neural network to form in the network optimization process;

Determining each other distribution object associated with a distribution object association line associated with the patient side personnel distribution object based on the patient relation characterization parameter distribution corresponding to the patient side personnel distribution object, acquiring an intermediate description vector calculated by each other distribution object in n-1 th iterative computation, performing multiplication operation on the intermediate description vector calculated by each other distribution object in n-1 th iterative computation based on a predetermined second parameter mapping matrix, performing average calculation operation on the result of each multiplication operation, and performing parameter interval mapping operation on the result of the average calculation operation to form associated other distribution object description vectors calculated by the patient side personnel distribution object in n-1 th iterative computation;

determining each distribution object associated line connected with the patient side personnel distribution object, acquiring an intermediate description vector calculated by each distribution object associated line in n-1 th iterative computation, respectively carrying out multiplication operation on the intermediate description vector calculated by each distribution object associated line in n-1 th iterative computation based on a predetermined third parameter mapping matrix, carrying out average computation operation on the result of each multiplication operation, and carrying out parameter interval mapping operation on the result of the average computation operation to form a distribution object associated line description vector calculated by the patient side personnel distribution object in n-1 th iterative computation;

And carrying out fusion operation, such as superposition operation, or cascade combination operation and full connection processing, on the intermediate description vector calculated by the patient side personnel distribution object in the nth iterative computation, the associated other distribution object description vector calculated by the patient side personnel distribution object in the nth iterative computation and the distribution object associated line description vector calculated by the patient side personnel distribution object in the nth iterative computation, so as to form an object key information description vector forming the next iteration number of each patient side personnel distribution object.

It should be understood that, in some possible embodiments, the step of performing the data merging operation based on the first type of object distribution network, each of the object key information description vectors and each of the associated line key information description vectors, and outputting the first type of distribution network description vector corresponding to the first type of object distribution network may further include the following in some specific applications:

performing parameter analysis operation on the adjacent relation representation parameter distribution, and analyzing and outputting medical relation representation parameter distribution corresponding to each medical party personnel distribution object, wherein the medical relation representation parameter distribution can be described by referring to the related description;

Performing data merging operation based on the object key information description vector of the current iteration number of each medical relation characterization parameter distribution and each medical staff distribution object to respectively form the object key information description vector of the next iteration number of each medical staff distribution object, wherein the related description can be referred to;

if the current iteration number is equal to the distribution network characterization parameter of the first type of object distribution network, marking an object key information description vector of the next iteration number of each medical party personnel distribution object to be marked as each medical party personnel key information in the first type of distribution network description vector, wherein the previous related description can be referred to;

Performing parameter analysis operation on the adjacent relation representation parameter distribution, and analyzing and outputting line adjacent relation representation parameter distribution corresponding to the associated line of each distribution object, wherein the related description can be referred to;

performing data merging operation based on the line adjacency relation characterization parameter distribution and the associated line key information description vector of the current iteration number of the associated line of each distribution object so as to respectively form the associated line key information description vector of the next iteration number of the associated line of each distribution object, which can be referred to the related description;

if the current iteration number is equal to the distribution network characterization parameter of the first type object distribution network, marking the associated line key information description vector of the next iteration number of the associated line of each distribution object to be marked as each hospital space action key information in the first type distribution network description vector, wherein the previous related description can be referred to;

It should be understood that, in some possible embodiments, step S140 in the foregoing description, that is, the step of performing the action anomaly identification operation on the actions to be analyzed in the hospital space corresponding to each distribution object association line based on the first type of distribution network description vector, to obtain action identification data corresponding to each action to be analyzed in the hospital space may include the following in some specific applications:

Based on the patient side personnel critical information, the medical side personnel critical information and the hospital space action critical information, performing evaluation operation of suspicious hospital space actions on each distribution object association line, and outputting a possibility parameter of each distribution object association line being a corresponding line of suspicious hospital space actions, for example, for each distribution object association line, cascade combination operation can be performed on the hospital space action critical information corresponding to the distribution object association line, the patient side personnel critical information corresponding to the connected patient side personnel distribution object and the medical side personnel critical information corresponding to the connected medical side personnel distribution object, then full connection processing is performed on a result of the cascade combination operation, and activation processing is performed on a result of the full connection processing to obtain a possibility parameter of the distribution object association line being a corresponding line of suspicious hospital space actions;

and obtaining motion identification data corresponding to the motion to be analyzed in each hospital space based on the probability parameter that each distribution object association line is a corresponding line of the suspicious hospital space motion, for example, determining the probability of the motion to be analyzed represented by the motion identification data as the probability of the motion to be suspicious, or determining the motion identification data corresponding to the probability parameter larger than a preset probability parameter as the motion to be suspicious.

It should be understood that, in some possible embodiments, after the step of acquiring a plurality of actions to be analyzed in hospital space, that is, after the step S110 in the above description, the digital twin-based hospital space optimization simulation method may further include the following:

determining a second class object distribution network based on the plurality of hospital space actions to be analyzed, wherein the second class object distribution network comprises action class distribution objects and action object association lines, and the action class distribution objects correspond to one hospital space action to be analyzed by way of example, and the object attribute information of the action class distribution objects comprises attribute information of corresponding patient personnel and attribute information of corresponding medical personnel; in addition, two action class distribution objects connected by the action object association line are provided with common patient personnel and/or medical personnel;

and carrying out map key information mining operation on the second class object distribution network, and outputting group space motion description vectors corresponding to the motion class distribution objects, wherein the group space motion description vectors belong to the motion key information for carrying out group motion.

Based on the foregoing, in some possible embodiments, the step of performing the operation anomaly identification operation on the motion to be analyzed in the hospital space corresponding to the correlation line of the distribution object based on the first type of distribution network description vector to obtain motion identification data corresponding to the motion to be analyzed in the hospital space may include the following in some specific applications:

Based on the first type of distribution network description vector and each group space motion description vector, performing motion anomaly identification operation on the motion to be analyzed in the hospital space corresponding to each distribution object association line to obtain motion identification data corresponding to each motion to be analyzed in the hospital space, and for each distribution object association line, for example, when performing the cascade combination operation, performing cascade combination operation on the corresponding group space motion description vector, and then performing subsequent processing.

It should be appreciated that in some possible embodiments, the group space motion description vector is obtained by performing a graph key information mining operation using a key information analysis network, and the network optimization process of the key information analysis network may include the following:

extracting a typical second-class object distribution network, wherein the typical second-class object distribution network comprises typical suspicious action class distribution objects corresponding to suspicious hospital space actions and typical action object association lines for associating the two typical action class distribution objects, the typical second-class object distribution network refers to a typical second-class object distribution network, namely a second-class object distribution network as a network optimization basis, and the suspicious hospital space actions can be selectively configured by corresponding users, for example, the actions that a patient needs to trend towards the head of a doctor, possibly representing words of an inaudible doctor, thus, poor spatial sealing of the hospital can be reflected, and excessively noisy sounds and the like are caused;

Performing key information mining operation on each typical suspicious action class distribution object in the typical second class object distribution network, for example, performing feature mining operation on corresponding object attribute information, so as to output a typical group action description vector corresponding to each typical suspicious action class distribution object, wherein the typical group action description vector refers to a typical group space action description vector;

according to typical group motion description vectors of the typical suspicious motion class distribution objects, analyzing typical motion identification data corresponding to the typical suspicious motion class distribution objects by utilizing a candidate key information analysis network, for example, performing full connection processing on the typical group motion description vectors, and performing activation processing on the obtained full connection vectors to obtain corresponding typical motion identification data, wherein the activation processing can be realized through functions such as softmax;

and performing parameter optimization adjustment operation on the candidate key information analysis network based on the typical action identification data corresponding to each typical suspicious action class distribution object and each typical suspicious action class distribution object to form a key information analysis network corresponding to the candidate key information analysis network, namely performing error calculation based on the typical action identification data corresponding to each typical suspicious action class distribution object and each typical suspicious action class distribution object, and performing parameter optimization adjustment operation on the candidate key information analysis network based on the calculated errors.

It should be appreciated that, in some possible embodiments, the step of performing the key information mining operation on each of the exemplary suspicious action class distribution objects in the exemplary second-class object distribution network to output the representative group action description vector corresponding to each of the exemplary suspicious action class distribution objects may include, in some specific applications, the following:

determining a typical global adjacent relation array corresponding to the typical second-class object distribution network, and determining a typical local adjacent relation array corresponding to each typical suspicious action class distribution object, wherein the typical global adjacent relation array is an array describing the relation between each typical action class distribution object in a typical second-class object distribution network pair, such as whether an action object association line is connected between the typical action class distribution objects, and the typical local adjacent relation array is an array describing the relation between each typical suspicious action class distribution object in the typical second-class object distribution network;

performing an adjacent analysis operation on each suspicious action class distribution object in the typical second class object distribution network, analyzing and forming a typical adjacent object cluster corresponding to each suspicious action class distribution object, wherein the typical adjacent object cluster comprises typical adjacent distribution objects which are associated with the typical action object association line between the corresponding suspicious action class distribution objects, the typical adjacent distribution objects belong to typical action class distribution objects, for example, the typical second class object distribution network comprises a typical action class distribution object A1, a typical action class distribution object A2, a typical action class distribution object A3, a typical action class distribution object A4 and a typical action class distribution object A5, a typical action class association line is associated between the typical action class distribution object A1 and the typical action class distribution object A2, a typical action class association line is associated between the typical action class distribution object A4 and the typical action class distribution object A5, and if the suspicious action class distribution object A1 is a typical action class association line, a typical action class association line is also carried out in a typical distribution class cluster corresponding to the typical action class distribution object A2, and a typical action class association line is also carried out in a typical distribution class cluster corresponding to the typical action class distribution object A2;

Performing reinforcement operation on the typical local adjacent relation array by using each typical suspicious adjacent object cluster to form a reinforced local adjacent relation array corresponding to each typical suspicious action class distribution object, that is, adding parameters representing whether typical action class distribution objects included in the typical suspicious adjacent object clusters are connected with action object association lines or not in the typical local adjacent relation array;

performing a section mapping operation on the enhanced local neighboring relation array corresponding to each typical suspicious action class distribution object to form a mapped local neighboring relation array corresponding to each typical suspicious action class distribution object, for example, performing a section mapping operation on parameters based on a sum value of parameters in the enhanced local neighboring relation array to map to a section 0-1 or other sections, or performing a section mapping operation on parameters of each column or row based on a sum value of parameters of the column or row to map to a section 0-1 or other sections;

It should be understood that, in some possible embodiments, the step of performing the analysis operation of the suspicious parameters based on each of the mapping local neighboring relation arrays to form a representative group motion description vector corresponding to each of the representative suspicious motion class distribution objects may include the following in some specific applications:

analyzing suspicious parameters based on the mapping local adjacent relation arrays to analyze undetermined typical group motion description vectors of the typical suspicious motion class distribution objects;

analyzing a fused local adjacent relation array corresponding to each typical suspicious action type distribution object by using the mapped local adjacent relation array, for example, the mapped local adjacent relation array and the typical global adjacent relation array can be overlapped to obtain a corresponding fused local adjacent relation array, and analyzing a correlation line quantity representation array corresponding to each typical suspicious action type distribution object based on each fused local adjacent relation array, wherein in the correlation line quantity representation array, parameters except a main diagonal are equal to 0, the value of a parameter on the main diagonal is equal to the quantity of action object correlation lines connected with a distribution object corresponding to a corresponding array coordinate in the fused local adjacent relation array, or the value of a parameter on the main diagonal is positive relative to the quantity of action object correlation lines connected with a distribution object corresponding to a corresponding array coordinate in the fused local adjacent relation array;

And merging operation is carried out on the basis of each fusion local adjacent relation array, each association line quantity representation array and the to-be-determined typical group motion description vector corresponding to each typical suspicious motion class distribution object so as to form a typical group motion description vector corresponding to each typical suspicious motion class distribution object.

It should be understood that, in some possible embodiments, the step of analyzing the pending typical group action description vector of each typical suspicious action class distribution object based on the analysis operation of suspicious parameters of each mapping local neighboring relation array may include the following in some specific applications:

configuring a reference vector for the typical suspicious action class distribution object, wherein the reference vector can be 1, and extracting a first proportionality coefficient and a second proportionality coefficient which are preset, wherein the first proportionality coefficient and the second proportionality coefficient both belong to 0-1, and the sum value is equal to 1;

multiplying the mapping local adjacent relation array and the intermediate vector to obtain a multiplication result vector, and carrying out weighted superposition on the multiplication result vector and the reference vector based on the first scaling factor and the second scaling factor to output a target vector, wherein the first scaling factor is used as a weighting factor of the multiplication result vector, and the second scaling factor is used as a weighting factor of the reference vector;

A step of performing the multiplication of the mapping local neighboring relation array and the intermediate vector with a rotation to obtain a multiplication result vector by taking the target vector as a new intermediate vector, and performing weighted superposition of the multiplication result vector and the reference vector based on the first scaling factor and the second scaling factor to output a target vector;

and after the step of performing the multiplication on the mapping local adjacent relation array and the intermediate vector for a plurality of times (the specific times are configured according to the requirement) to obtain a multiplication result vector, and performing weighted superposition on the multiplication result vector and the reference vector based on the first scaling factor and the second scaling factor to output a target vector, the target vector obtained by performing the step for the last time can be used as a pending typical group motion description vector of the typical suspicious motion class distribution object, wherein the intermediate vector is used as the reference vector when the step is performed for the first time.

It should be understood, that in some possible embodiments, the step of performing a merging operation based on each of the fused local neighboring relation arrays, each of the correlation line quantity representation arrays, and the to-be-determined typical group motion description vector corresponding to each of the typical suspicious motion class distribution objects to form a typical group motion description vector corresponding to each of the typical suspicious motion class distribution objects may include, in some specific applications, the following:

And performing negative zero point fifth power operation on the association line quantity representation array to output a corresponding power operation array, performing multiplication operation on the power operation array, the fusion local adjacent relation array, the power operation array and the undetermined typical group motion description vector, and performing parameter interval mapping operation on vectors or arrays obtained by the multiplication operation to form the typical group motion description vector corresponding to the typical suspicious motion type distribution object.

With reference to fig. 3, the embodiment of the invention also provides a digital twin-based hospital space optimization simulation device, which can be applied to the digital twin-based hospital space optimization simulation system. The digital twin-based hospital space optimization simulation device can comprise:

the system comprises an action acquisition module, a medical staff management module and a medical staff management module, wherein the action acquisition module is used for acquiring a plurality of hospital space actions to be analyzed, the hospital space actions to be analyzed are used for reflecting the hospital space actions output by the medical staff by the patient staff, each hospital space action to be analyzed is recorded through corresponding text data or image data, and the plurality of hospital space actions to be analyzed are formed by performing operation simulation on a simulated medical space;

The distribution network determining module is used for determining a first type of object distribution network based on the plurality of hospital space actions to be analyzed, wherein the first type of object distribution network comprises patient side personnel distribution objects corresponding to the patient side personnel, medical side personnel distribution objects corresponding to the medical side personnel and distribution object association lines for associating the patient side personnel distribution objects with the medical side personnel distribution objects;

the map key information mining module is used for performing map key information mining operation on the first type object distribution network so as to output a first type distribution network description vector corresponding to the first type object distribution network, wherein the first type distribution network description vector comprises patient side personnel key information corresponding to each patient side personnel distribution object, medical side personnel key information corresponding to each medical side personnel distribution object and hospital space action key information corresponding to each distribution object association line;

the motion anomaly identification module is used for respectively carrying out motion anomaly identification operation on the motion to be analyzed in the hospital space corresponding to each distribution object association line based on the first type of distribution network description vector to obtain motion identification data corresponding to each motion to be analyzed in the hospital space, wherein the motion identification data are used for reflecting whether the motion to be analyzed in the hospital space belongs to suspicious motion or the possibility of the suspicious motion;

The medical space simulation screening module is used for determining space priority parameters of the simulated medical spaces based on action identification data corresponding to actions to be analyzed of the hospital spaces, and screening out at least one target simulated medical space from the simulated medical spaces based on the space priority parameters after obtaining the space priority parameters of the simulated medical spaces, wherein the target actual medical space belongs to a space to be built as a medical space model optimization result of the target actual medical space.

In summary, the digital twin-based hospital space optimization simulation method and system provided by the invention can collect a plurality of hospital space actions to be analyzed first; determining a first class object distribution network based on a plurality of actions to be analyzed in hospital space; performing map key information mining operation on the first class object distribution network to output a first class distribution network description vector; based on the first type of distribution network description vector, obtaining action identification data corresponding to actions to be analyzed in each hospital space; based on the motion identification data, determining the space priority parameters of the simulated medical space, and screening out at least one target simulated medical space based on the space priority parameters after obtaining the space priority parameters of a plurality of simulated medical spaces, wherein the target simulated medical space is used as a medical space model optimization result of the target actual medical space. Based on the foregoing, because the operation of identifying the motion abnormality of the motion to be analyzed in the hospital space corresponding to the association line of each distribution object is performed based on the first type of distribution network description vector, that is, the motion to be analyzed in the hospital space is independently analyzed, the reliability of the motion identification data corresponding to the motion to be analyzed in each obtained hospital space is higher, so that the reliability of the simulated medical space screening based on the motion identification data can be improved to a certain extent, and the reliability of the optimized simulation in the hospital space is improved.

The above description is only of the preferred embodiments of the present invention and is not intended to limit the present invention, but various modifications and variations can be made to the present invention by those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. The hospital space optimization simulation method based on digital twinning is characterized by comprising the following steps of:

determining a first type of object distribution network based on the collected actions to be analyzed in a plurality of hospital spaces;

performing map key information mining operation on the first type object distribution network to output a first type distribution network description vector corresponding to the first type object distribution network;

based on the first type of distribution network description vector, performing action abnormality recognition operation to obtain action recognition data corresponding to actions to be analyzed in the hospital space;

after motion identification data corresponding to the motion to be analyzed of each hospital space in a plurality of simulated medical spaces are obtained, screening at least one target simulated medical space from the plurality of simulated medical spaces based on the motion identification data, and taking the target simulated medical space as a medical space model optimization result of a target actual medical space;

The step of determining the first class object distribution network based on the collected actions to be analyzed in the plurality of hospital spaces comprises the following steps:

after the step of collecting the actions to be analyzed of the plurality of hospital spaces, the digital twinning-based hospital space optimization simulation method further comprises the following steps:

the first type of distribution network description vector includes patient side personnel key information corresponding to each patient side personnel distribution object, medical side personnel key information corresponding to each medical side personnel distribution object, and hospital space action key information corresponding to each distribution object association line, wherein the step of performing action abnormality identification operation based on the first type of distribution network description vector to obtain action identification data corresponding to actions to be analyzed in the hospital space includes:

based on the first type of distribution network description vector and each group space motion description vector, performing motion anomaly identification operation on the hospital space motion to be analyzed corresponding to each distribution object association line to obtain motion identification data corresponding to each hospital space motion to be analyzed, wherein the motion identification data are used for reflecting whether the hospital space motion to be analyzed belongs to suspicious motion or the probability of the suspicious motion;

The first class object distribution network includes a patient side personnel distribution object corresponding to a patient side personnel, a medical side personnel distribution object corresponding to a medical side personnel, and a distribution object association line for associating the patient side personnel distribution object and the medical side personnel distribution object, and the step of performing a map key information mining operation on the first class object distribution network to output a first class distribution network description vector corresponding to the first class object distribution network includes:

performing data merging operation based on the first type object distribution network, each object key information description vector and each associated line key information description vector, and outputting a first type distribution network description vector corresponding to the first type object distribution network;

The step of performing data merging operation based on the first type object distribution network, each object key information description vector and each associated line key information description vector and outputting a first type distribution network description vector corresponding to the first type object distribution network includes:

If the current iteration number is equal to one, the object key information description vector of the patient side personnel distribution object with the iteration number equal to one is the object key information description vector of the patient side personnel distribution object;

the step of performing a data merging operation based on the object key information description vector of the current iteration number of each patient-side personnel distribution object and each patient-side personnel distribution object to respectively form the object key information description vector of the next iteration number of each patient-side personnel distribution object includes:

performing multiplication operation on the intermediate description vector calculated by the patient side personnel distribution object in the n-1 th iterative computation based on a first predetermined parameter mapping matrix, and performing parameter interval mapping operation on the result of the multiplication operation to form the intermediate description vector calculated by the patient side personnel distribution object in the n-1 th iterative computation, wherein the first parameter mapping matrix is used as a network parameter to form in the network optimization process;

And carrying out fusion operation on the intermediate description vector calculated by the patient side personnel distribution object in the nth iterative computation, the associated other distribution object description vector calculated by the patient side personnel distribution object in the nth iterative computation and the distribution object associated line description vector calculated by the patient side personnel distribution object in the nth iterative computation so as to form an object key information description vector of the next iteration number of the patient side personnel distribution object.

2. The digital twinning-based hospital space optimization simulation method according to claim 1, wherein the group space motion description vector is obtained by performing a map key information mining operation by using a key information analysis network;

3. The method for optimizing simulation of hospital space based on digital twinning according to claim 2, wherein the step of performing a key information mining operation on each of the typical suspicious action class distribution objects in the typical second class object distribution network to output a typical group action description vector corresponding to each of the typical suspicious action class distribution objects comprises:

4. The method for simulating spatial optimization of a hospital based on digital twinning according to claim 1, wherein the step of performing data merging operation based on the first-class object distribution network, each of the object key information description vectors, and each of the associated line key information description vectors, and outputting a first-class distribution network description vector corresponding to the first-class object distribution network further comprises:

5. The method for simulating spatial optimization of a hospital based on digital twinning according to claim 1, wherein the step of performing data merging operation based on the first-class object distribution network, each of the object key information description vectors, and each of the associated line key information description vectors, and outputting a first-class distribution network description vector corresponding to the first-class object distribution network further comprises:

6. A digital twin based hospital space optimization simulation system comprising a processor and a memory, the memory for storing a computer program, the processor for executing the computer program to implement the digital twin based hospital space optimization simulation method of any of claims 1-5.