A ward intelligent calling method and system based on dual-mode communication and multi-terminal linkage

Abstract

The application relates to a ward intelligent calling method and system based on dual-mode communication and multi-terminal linkage, which comprises the following steps: a system receives a patient terminal trigger signal and generates a structured calling request data packet. The real-time communication quality of an IP network is monitored, and a PSTN link is activated as a main channel when the quality is lower than a threshold. The data packet is sent to a calling soft switch through the main channel, a nurse scheduling database is queried, a target nurse terminal group list is obtained, a request is broadcast to the target group, and response state data is received. Based on the nursing level and the state data, a nurse response priority sequence is generated through a dynamic weight model. An audio and video connection is initiated to the first nurse in the sequence, and a two-way communication session is established. If the communication signal strength continuously falls below an interruption threshold or the nurse's moving speed exceeds a safety threshold during the session, the system automatically switches to the second nurse in the sequence. The application realizes high-reliable response of patient calling by dynamically selecting a transmission channel, intelligently allocating nurse resources and monitoring the session quality in real time.

Description

Technical Field

[0001] This application relates to the field of medical information technology, and in particular to a method and system for intelligent ward calling based on dual-mode communication and multi-terminal linkage. Background Technology

[0002] With the continuous development of medical informatization, the concept of smart wards has gradually become a key direction for improving the quality and efficiency of medical services. Traditional ward call systems, as the infrastructure for doctor-patient communication, mostly adopt wired analog or digital communication technologies. Their core function is to enable one-way voice calls between the patient's bedside and the nurse's station. These systems have formed the cornerstone of ward nursing communication for the past few decades, and their design philosophy mainly focuses on achieving basic distress signal transmission.

[0003] However, with the increasing demands for precision, timeliness, and humanization in clinical nursing, and the high degree of integration in hospital information systems, the limitations of traditional call systems are becoming increasingly apparent. For example, in the face of network equipment failure, line interruption, or signal interference, system services may be completely interrupted, posing a risk to patient safety. Furthermore, the content of patient calls is often limited, typically only including the bed number, leaving nurses unable to ascertain the patient's identity and specific needs before answering, leading to insufficient preparedness. In addition, the system's scheduling logic usually notifies based on call order or fixed zones, failing to dynamically and intelligently schedule shifts and assign tasks according to the nurse's current status and the urgency of the patient's call. This results in low efficiency in utilizing nursing human resources, with some nurses overburdened while others are unable to fully utilize their abilities, thus reducing the patient's medical experience. Summary of the Invention

[0004] To address the aforementioned technical issues, this application provides a method and system for intelligent ward calling based on dual-mode communication and multi-terminal linkage.

[0005] Firstly, this application provides a ward intelligent calling method based on dual-mode communication and multi-terminal linkage, employing the following technical solution: Receive a trigger signal sent by the patient terminal, and generate a structured call request data packet containing the patient identifier, care level, and call type based on the trigger signal; The system acquires real-time communication quality parameters of the IP network. When the real-time communication quality parameters are detected to be lower than a preset parameter threshold, the PSTN communication link is activated as the main transmission channel. The structured call request data packet is sent to the call softswitch through the main transmission channel, and the call softswitch queries the nurse scheduling database to obtain the target nurse terminal group identifier list. The nursing level and call type in the structured call request data packet are parsed, and combined with the target nurse terminal group identifier list, a nurse terminal broadcast policy instruction is generated; According to the broadcast strategy instructions, the structured call request data packet is broadcast to the target nurse terminal group, and the response status data returned by each nurse terminal is received, including terminal location coordinates, current task count and nursing skill matching degree; Based on the nursing level and response status data, a nurse response priority sequence is generated through a dynamic weighting model; An audio / video connection request is initiated to the target nurse terminal that is first in the nurse response priority sequence, establishing a two-way communication session between the patient terminal and the target nurse terminal; During the duration of the two-way communication session, the communication signal strength and the target nurse terminal's movement speed are monitored in real time. When the communication signal strength is continuously lower than a preset interruption threshold or the terminal's movement speed continuously exceeds a safety threshold, the two-way communication session is switched to the nurse terminal ranked second in the nurse response priority sequence.

[0006] By adopting the above technical solutions, from the structured data encapsulation at the time of call initiation, to the intelligent redundancy switching of the communication link, to the precise targeted broadcasting and multi-dimensional data collection based on scheduling and real-time status, and then to the use of dynamic weight model for optimal nursing resource scheduling, a series of steps are formed to establish a high-quality audio and video session and implement full-process quality monitoring and seamless hot migration. This series of steps forms an adaptive and data-driven complete closed loop.

[0007] Secondly, this application provides a ward intelligent call system based on dual-mode communication and multi-terminal linkage, adopting the following technical solution: The call request standardization module is used to receive trigger signals sent by the patient terminal and generate a structured call request data packet containing the patient identifier, care level and call type based on the trigger signals. The communication link redundancy switching module is used to obtain the real-time communication quality parameters of the IP network. When the real-time communication quality parameters are detected to be lower than the preset parameter threshold, the PSTN communication link is activated as the main transmission channel. The data routing and target group decision module is used to send the structured call request data packet to the call softswitch through the main transmission channel, and the call softswitch queries the nurse scheduling database to obtain the target nurse terminal group identifier list. The broadcast strategy intelligent generation module is used to parse the nursing level and call type in the structured call request data packet, and generate broadcast strategy instructions for the nurse terminal by combining the target nurse terminal group identifier list; The multi-terminal status synchronization acquisition module is used to broadcast the structured call request data packet to the target nurse terminal group according to the broadcast strategy instruction, and to receive the response status data returned by each nurse terminal, including terminal location coordinates, current task count and nursing skill matching degree. The dynamic priority scheduling module is used to generate a nurse response priority sequence based on the nursing level and response status data through a dynamic weight model. The intelligent audio and video session establishment module initiates an audio and video connection request to the target nurse terminal that is first in the nurse response priority sequence, thereby establishing a two-way communication session between the patient terminal and the target nurse terminal. The session switching module is used to monitor the communication signal strength and the target nurse terminal's movement speed in real time during the duration of the two-way communication session. When the communication signal strength is continuously lower than a preset interruption threshold or the terminal's movement speed continuously exceeds a safety threshold, the two-way communication session is switched to the nurse terminal ranked second in the nurse response priority sequence.

[0008] Thirdly, this application provides a computer-readable storage medium, which adopts the following technical solution: A computer-readable storage medium storing a computer program that can be loaded by a processor and executed as in any of the methods in the first aspect.

[0009] In summary, this application includes at least one of the following beneficial technical effects: ensuring reliable transmission of call requests through automatic switching of dual-mode communication, dynamically calculating response priority based on nurse scheduling, real-time location, task load and skill matching degree, and realizing intelligent establishment and fault switching of audio and video communication sessions, thereby improving the reliability, response speed and nursing resource allocation efficiency of the ward call system. Attached Figure Description

[0010] Figure 1 This is a first flowchart of a ward intelligent calling method based on dual-mode communication and multi-terminal linkage, which is one embodiment of this application.

[0011] Figure 2 This is a second flowchart illustrating a ward intelligent calling method based on dual-mode communication and multi-terminal linkage, according to one embodiment of this application.

[0012] Figure 3 This is a schematic diagram of the third process of a ward intelligent calling method based on dual-mode communication and multi-terminal linkage, which is one embodiment of this application.

[0013] Figure 4 This is a schematic diagram of the fourth process of a ward intelligent calling method based on dual-mode communication and multi-terminal linkage, which is one embodiment of this application.

[0014] Figure 5This is a schematic diagram of the fifth process of a ward intelligent calling method based on dual-mode communication and multi-terminal linkage, which is one embodiment of this application.

[0015] Figure 6 This is a schematic diagram of the sixth process of a ward intelligent calling method based on dual-mode communication and multi-terminal linkage, which is one embodiment of this application.

[0016] Figure 7 This is a schematic diagram of the seventh process of a ward intelligent calling method based on dual-mode communication and multi-terminal linkage, which is one embodiment of this application.

[0017] Figure 8 This is the eighth flowchart of a ward intelligent calling method based on dual-mode communication and multi-terminal linkage, which is one embodiment of this application. Detailed Implementation

[0018] To make the purpose, technical solution, and advantages of this application clearer, the following description is provided in conjunction with the appendix. Figures 1-8 The present application will be further described in detail below with reference to embodiments. It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the scope of the application.

[0019] This application discloses a method for intelligent ward calling based on dual-mode communication and multi-terminal linkage.

[0020] Reference Figure 1 A method for intelligent ward calling based on dual-mode communication and multi-terminal linkage, the specific method includes: Step S101: Receive a trigger signal sent by the patient terminal, and generate a structured call request data packet containing the patient identifier, care level, and call type based on the trigger signal; When a patient initiates a physical operation (such as pressing a button) through a terminal device like a bedside screen or emergency call button, a raw trigger signal is generated. Upon receiving this signal, the system does not directly transmit this simple signal, but immediately initiates a data augmentation and structuring process.

[0021] First, the system analyzes the device type from which the trigger signal originates (e.g., a regular bedside screen or a red emergency call button) and the operation type (e.g., a single press or a long press), both of which imply different levels of urgency. Next, based on this, the system queries the backend database, binds and extracts the unique patient identifier corresponding to the bed (usually the hospital number or an identity ID associated with the bed number), and the nursing care level synchronized from the hospital information system (e.g., special care, level one care, representing the frequency and intensity of attention required by the patient).

[0022] Finally, by combining the operation type and device type, the system will determine the call type (e.g., routine assistance, infusion completed, emergency). All this key information is encapsulated into a structured call request data packet with defined fields and format, the purpose of which is to provide a complete and uniformly formatted data entity for all subsequent processing steps.

[0023] Step S102: Obtain the real-time communication quality parameters of the IP network. When the real-time communication quality parameters are detected to be lower than the preset parameter threshold, activate the PSTN communication link as the main transmission channel. The system prioritizes using the IP network (based on the hospital's LAN or the Internet) as a high-speed, low-cost data channel for communication by default. However, the IP network may become unstable due to switch failures, loose network cables, or network congestion. Therefore, the system continuously and proactively monitors the real-time communication quality parameters of the IP link. These parameters typically include network latency (the time it takes for a data packet to be sent and received) and packet loss rate (the proportion of data packets lost in the transmitted data). The system compares these real-time parameters with preset threshold values ​​that represent the minimum availability requirements.

[0024] In this embodiment, communication quality detection is not a one-time instantaneous judgment, but rather employs algorithms such as continuous monitoring or sliding window averaging to avoid erroneous handovers caused by instantaneous network jitter. Once the system confirms that the IP network quality has continuously deteriorated to an unreliable level (i.e., "below the preset parameter threshold"), it will immediately and automatically trigger a control command to activate the PSTN (Public Switched Telephone Network), a backup traditional telephone line, and promote it to the primary transmission channel. The PSTN network is characterized by its independence, dedicated lines, and extreme stability.

[0025] Understandably, a dual-mode heterogeneous communication security network was constructed to ensure that, in the event of any single network failure, critical call request data packets can still be reliably transmitted to the core hub of the system.

[0026] Step S103: The structured call request data packet is sent to the call softswitch through the main transmission channel, and the call softswitch queries the nurse scheduling database to obtain the target nurse terminal group identifier list. Structured data packets are delivered to the system's central processing unit, the call softswitch, via the currently active main transmission channel. The softswitch is not a simple hardware telephone exchange, but a server running dedicated communication software, possessing powerful signal processing, routing control, and business logic processing capabilities.

[0027] Specifically, upon receiving a call request, the softswitch's primary task is not to blindly broadcast, but to perform precise responsibility identification. It queries the nurse scheduling database, a dynamic database recording the responsible nurse relationships for each ward and bed during the current time period. By matching the patient's bed or ward information in the call request, the softswitch can accurately obtain a list of target nurse terminal group identifiers. This list may include the patient's responsible nurse, their nursing team leader, and the terminal identifiers of all nurses currently within the corresponding responsibility area. This step ensures that the call is received by the relevant responsible parties while preventing irrelevant nurses from being interfered with by invalid information, thus improving the efficiency of the response organization.

[0028] Step S104: Parse the nursing level and call type in the structured call request data packet, and generate a nurse terminal broadcast policy instruction by combining it with the target nurse terminal group identifier list; The call softswitch parses key business fields in the data packets, namely the nursing level and call type. An emergency call from a critically ill patient requires a significantly different level of urgency and mobilization scope compared to a routine call from a patient requiring level 3 care. The system combines these two dimensions with a list of target nurse terminal group identifiers to generate a broadcast policy instruction. This instruction may specify the broadcast priority, whether to repeat the broadcast, and whether to expand the broadcast scope (e.g., in an emergency, to notify nurses in adjacent areas of responsibility for backup).

[0029] Step S105: Broadcast a structured call request data packet to the target nurse terminal group according to the broadcast strategy instruction, and receive the response status data returned by each nurse terminal, including terminal location coordinates, current task count and nursing skill matching degree. According to the aforementioned strategy, the system broadcasts data to the target nurse terminals in the list, typically using reliable message queues or multicast protocols to ensure that each target terminal receives the data packet exactly once. Simultaneously with the broadcast, the system actively requests and waits to receive response status data from each nurse terminal. This data includes: terminal location coordinates (obtained via indoor Bluetooth beacons, Wi-Fi positioning, or GPS to determine the physical distance between the nurse and the patient), current task count (representing the number of unfinished calls or nursing tasks currently being processed by the nurse, reflecting their immediate workload), and nursing skill matching degree (calculated by comparing the nurse's professional skill tags with the patient's nursing needs, for example, whether an expert in intravenous puncture is needed).

[0030] Step S106: Based on nursing level and response status data, generate a nurse response priority sequence through a dynamic weight model; Specifically, a multi-factor dynamic assessment model is constructed to achieve optimal matching of nursing resources. After receiving feedback data from all relevant nurse terminals, the system inputs this data, along with the nursing level information included in the call request, into a preset dynamic weight model.

[0031] In this embodiment, the model is a mathematical decision-making algorithm that assigns weight coefficients to multiple factors, including distance, task load, nursing skill matching degree, and call urgency (related to nursing level), and calculates a comprehensive priority score for each nurse. For example, nurses who are closer to the caller will receive a higher score (reducing response time), but if their current task is already at full capacity, this score will be lower (to avoid overloading individual nurses); at the same time, for technically demanding needs, nurses with high nursing skill matching degree will receive a significant bonus.

[0032] The system uses this model to calculate and rank all candidate nurses, ultimately generating a nurse response priority sequence. This step upgrades the traditional dispatching method, which relies on the head nurse's experience or the nurse's initiative, to a real-time data-driven, globally optimized, and objective automated scheduling system. The aim is to minimize overall response time at the system level and assign the most suitable nurse to the call that best matches their capabilities.

[0033] Step S107: Initiate an audio / video connection request to the target nurse terminal that is first in the nurse response priority sequence to establish a two-way communication session between the patient terminal and the target nurse terminal. Specifically, based on the optimal sequence calculated in the previous step, the system automatically initiates an audio / video connection request to the target nurse terminal that ranks first in the sequence, i.e., the one with the highest overall priority. Once the nurse accepts the request on their terminal, the system utilizes the established IP or high-quality communication link to establish a stable, two-way, low-latency audio / video communication session between the patient terminal (such as a bedside screen) and the nurse terminal (such as a PDA or workstation computer). This step not only achieves a qualitative leap from silent alarms to face-to-face communication, allowing nurses to intuitively see the patient's condition and provide reassurance, but also makes it possible to remotely conduct preliminary assessments of the patient's condition. It is a key technological achievement for improving the quality of nursing care and patient experience.

[0034] Step S108: During the two-way communication session, the communication signal strength and the target nurse terminal movement speed are monitored in real time. When the communication signal strength is continuously lower than the preset interruption threshold or the terminal movement speed continuously exceeds the safety threshold, the two-way communication session is switched to the nurse terminal ranked second in the nurse response priority sequence.

[0035] The establishment of a communication session is not the end point; the system monitors the entire session in real time. There are two core monitoring indicators: first, the communication signal strength, which directly affects call quality and video smoothness; and second, the terminal movement speed, which indirectly reflects the nurse's status (for example, running quickly may mean that she is handling an emergency or is about to leave the current communication coverage area).

[0036] In this embodiment, the system continuously samples this data and uses algorithms such as sliding windows for continuous judgment (e.g., multiple consecutive drops below a threshold, rather than a single instantaneous fluctuation). Once it detects that the communication quality is about to be interrupted due to weak signal, or that the nurse may be unable to continue the current session due to rapid movement (e.g., needing to run to the emergency scene), the system will not wait for the communication to be completely interrupted or for the patient to call again. Instead, it will proactively and smoothly switch (hot-migrate) the current audio and video session link, along with the relevant patient context information, to the nurse terminal ranked second in the "nurse response priority sequence".

[0037] Understandably, this step ensures that single points of failure or temporary state changes do not lead to service interruptions, achieving closed-loop management from establishing a connection to maintaining a high-quality connection, reflecting system-level robustness and the ultimate pursuit of continuity of care.

[0038] In the above implementation, from the structured data encapsulation at the time of call initiation, to the intelligent redundancy switching of the communication link, to the precise targeted broadcasting and multi-dimensional data collection based on the scheduling and real-time status, and then to the optimal scheduling of nursing resources using a dynamic weight model, a series of steps are formed to establish a high-quality audio and video session and implement full-process quality monitoring and seamless hot migration. This series of steps forms an adaptive and data-driven complete closed loop.

[0039] In practical applications, this technical solution transcends the simple, passive ringing mode of traditional call systems, evolving into a smart nursing communication hub capable of proactively sensing the network environment, intelligently analyzing nursing resources, dynamically optimizing scheduling decisions, and ensuring communication quality throughout the process. It systematically reduces the missed call rate to an extremely low level, shortens the average time from when a patient seeks help to when effective communication is achieved, realizes precise and dynamic matching of nursing manpower and task requirements, and ensures the continuity of communication services in any unexpected situation, thereby improving the overall efficiency, safety, and patient satisfaction of ward nursing work.

[0040] Reference Figure 2 As one implementation of step S101, the step of receiving a trigger signal sent by the patient terminal and generating a structured call request data packet containing a patient identifier, care level, and call type based on the trigger signal includes: Step S201: Receive the trigger signal sent by the patient terminal, and extract the device type code and operation type code from the trigger signal; When a patient performs an operation (such as pressing a button or touching a specific area of ​​the screen), the patient terminal generates a trigger signal that has undergone preliminary digital processing and contains a specific information structure. This trigger signal embeds the encoding of two key dimensions: device type encoding and operation type encoding.

[0041] Specifically, the device type code identifies the physical device from which the signal originates; for example, it might represent a standard call screen installed at the bedside, an emergency pull cord in the bathroom, or a remote control held by the patient. The operation type code records the patient's specific interaction with the device, such as whether it is a single short press, a long press, or multiple consecutive presses.

[0042] Understandably, the logic behind extracting these two codes lies in separating the two fundamental attributes of a call's physical carrier (what kind of device) and the user's intent (what kind of operation) at the very beginning of the process, and assigning them unambiguous numerical codes that can be recognized by the machine. This lays a precise data foundation for subsequent differentiated processing based on different device characteristics (such as emergency call buttons usually having the highest priority) and different operational meanings (such as a long press possibly representing an emergency).

[0043] Step S202: Determine the calling device category identifier based on the device type code and the preset device type mapping table; This can be achieved by using a lookup table to convert generic hardware identifiers into category labels with specific business logic meaning. The device type code is a relatively primitive number defined by the underlying hardware or driver protocol. The system maintains a preset device type mapping table, which is essentially a set of key-value pairs. Its function is to map various possible hardware codes to the call device category identifier defined and understood by the system's business layer. For example, the code 0x10 might be mapped to BEDSIDE_NORMAL (bedside regular screen), and the code 0xF1 might be mapped to EMERGENCY_PULL (emergency pull cord).

[0044] Understandably, this step masks the diversity and complexity of the underlying hardware. Regardless of the different manufacturers or hardware models of the devices connected in the future, as long as their drivers can output a pre-defined code, the system can query this maintainable and scalable mapping table to uniformly categorize them into a limited set of device categories that can be processed by the business logic.

[0045] Step S203: Obtain the current timestamp of the system clock and generate a globally unique call request identifier by combining it with the patient's bed number; Specifically, the strategy for generating call request identifiers integrates both time and space (business location) elements. The current timestamp of the system clock (usually accurate to milliseconds or even microseconds) provides a guarantee of time uniqueness. The patient bed number provides a stable spatial or business anchor point. Combining these two (e.g., through string concatenation, or more complex bitwise operations and hash algorithms) generates an identifier that not only guarantees global uniqueness but also directly or indirectly associates it with the time of the call and the physical bed location.

[0046] Step S204: Query the patient database of the hospital information system to obtain the patient identifier and nursing level bound to the patient's bed number; The core focus of nursing care is the patient, not the bed. By using the patient's bed number as a key index, a real-time query is proactively initiated into the patient database of the Hospital Information System (HIS) to obtain two key pieces of data: first, the patient identifier (such as hospital number or medical record number), which is a unique and official identity identifier for the patient in all information systems within the hospital. This identifier can be used to link the call with all information such as the patient's electronic medical record and doctor's orders; second, the nursing level (such as special care or level one care), which is a key business label set by medical staff based on the patient's condition assessment, reflecting the required intensity and frequency of care.

[0047] Understandably, by transforming a relatively abstract signal triggered by hardware into a business event closely tied to a specific patient and their care needs, all subsequent processing (such as prioritization and nurse skill matching) can be based on accurate and personalized patient information. This forms the data foundation for achieving precise and personalized nursing responses.

[0048] Step S205: Generate the corresponding call type identifier according to the operation type code and the preset call type mapping rule; Among them, the operation type code only represents a physical action (such as key press mode), while the call type identifier defines the type of business service that this action is expected to trigger. The system can complete this translation according to the preset "call type mapping rules".

[0049] For example, the rule might be defined as follows: a short press (0x01) is mapped to VOICE_CALL (normal voice call), a long press for three seconds (0x02) is mapped to EMERGENCY_CALL (emergency call), and tapping the "IV drip complete" icon on the screen (0x03) is mapped to INFUSION_FINISHED (specific event reporting).

[0050] Understandably, the same device (such as a bedside screen) can trigger different types of calls through different operating methods without the need for additional hardware buttons. Mapping rules can be customized and extended according to clinical needs, enabling the system to support a wide variety of business scenarios with a unified hardware interface, clearly and unambiguously conveying the user's operational intent to subsequent processing logic.

[0051] Step S206: Encapsulate the call request identifier, patient identifier, care level, call device category identifier, and call type identifier into a structured call request data packet.

[0052] Among them, five key data elements are encapsulated: call request identifier (unique ID), patient identifier (core object), care level (business priority), call device category identifier (source characteristic), and call type identifier (business intent). These elements are organized in a structured and standardized protocol format (such as JSON, XML, or a more efficient binary TLV format) to ensure data integrity and consistency.

[0053] Furthermore, the receiving party (such as a call softswitch) can strictly parse according to the format specifications and extract each field without error, which improves the efficiency of data transmission and processing and provides support for the scalability of the system. If new information fields (such as patient allergy history identifiers) need to be added to the call request in the future, it is only necessary to extend the new fields in the definition of the structured data unit without overturning the entire communication protocol.

[0054] The above implementation overcomes the shortcomings of traditional call systems, which rely on simple and vague information. Each patient call, from the moment it is initiated, becomes a unique, clearly identified, and clearly defined intelligent data object deeply tied to the patient's personalized care needs. This not only provides a high-quality data foundation for subsequent intelligent priority determination and optimized allocation of nursing resources, but also fundamentally improves the reliability, accuracy, and business relevance of the entire intelligent call system.

[0055] Reference Figure 3 As one implementation of step S102, the step of obtaining real-time communication quality parameters of the IP network and activating the PSTN communication link as the main transmission channel when the real-time communication quality parameters are detected to be lower than a preset parameter threshold includes: Step S301: Periodically collect real-time communication quality datasets of the IP network through the system network interface, including network latency measurements and packet loss statistics; The system does not passively respond after communication is interrupted, but actively initiates probes through the network interface at fixed time intervals (e.g., per second), forming a continuous observation of network quality changes over time, avoiding the randomness of single instantaneous measurements. The collected communication quality dataset focuses on two of the most critical indicators affecting real-time communication (especially audio and video): network latency measurement (usually referring to the round-trip time from sending to receiving data packets, RTT) and packet loss statistics (the proportion of a certain number of probe packets sent that were not successfully received).

[0056] Understandably, high latency can cause audio and video conversations to stutter and become out of sync; high packet loss rates can directly lead to intermittent voice communication, video glitches, or even disconnections. By continuously collecting these two key parameters, the system can build a dynamic data profile of the current performance and stability of the IP network, providing quantitative and factual evidence for subsequent intelligent decision-making, rather than relying on guesswork or simple connectivity judgments.

[0057] Step S302: Input the communication quality dataset into the preset network quality assessment model to generate network status assessment results; Using the raw latency and packet loss data stream directly to trigger a switch is unstable, because any network has normal, instantaneous fluctuations (for example, a large file transfer may cause a microsecond-level latency spike).

[0058] In this embodiment, the pre-defined quality assessment model acts as an intelligent filter and trend analyzer. This model may include a series of algorithms, such as: calculating the average and variance of latency within a recent window to determine its stability; statistically analyzing the frequency and duration of packet loss events; setting dynamic thresholds based on different service tolerances (e.g., for voice calls, latency exceeding 150 milliseconds may be unacceptable); and even employing more complex machine learning models to identify early patterns of network quality degradation. After processing the input dataset, the model outputs a highly generalized qualitative or semi-quantitative conclusion for guiding action, such as "excellent," "good," "slightly degraded," "severely degraded," or "unavailable."

[0059] Step S303: When the network status assessment result indicates that the communication quality parameters are lower than the preset parameter threshold, a link activation command is sent to the PSTN gateway. The core logic of this step is a condition-triggered mechanism: the system will initiate the handover process only if the generated network status assessment clearly indicates that the communication quality is below a preset parameter threshold. This parameter threshold is a predefined threshold representing the minimum acceptable service level for the business. It is not a single numerical value, but a set of standards that match the assessment model (for example, a severely degraded assessment result is considered to be below the threshold). This judgment is the sole basis for the handover decision, ensuring the objectivity and necessity of the handover behavior and preventing unnecessary frequent handovers.

[0060] Specifically, once the conditions are met, the system sends a link activation command to the PSTN gateway, which is a physical or virtual device connecting the IP network to the traditional telephone exchange network. Sending the activation command means that the system actively commands the gateway to prepare to activate the backup PSTN line. This step marks the formal transition of the system from the "monitoring and early warning" state to the "fault switching execution" state, and is a key node in the transfer of control from the monitoring module to the execution module.

[0061] Step S304: Receive the link ready status signal returned by the PSTN gateway and set the current main transmission channel as a PSTN communication link.

[0062] Upon receiving the activation command from the system, the PSTN gateway needs to perform internal preparations (such as occupying a telephone line and establishing a signaling channel), a process that takes time and may fail. Therefore, the system must receive a link readiness status signal from the PSTN gateway to ensure that the backup link is truly ready for data transmission, avoiding service interruptions caused by blindly switching before the link is ready. Only after confirming receipt of this readiness signal will the system set the current primary transmission channel as the PSTN communication link.

[0063] Specifically, this configuration change is a system-level change, and its underlying implementation may include: modifying the default routing table to direct all outgoing call signaling and data packets to the PSTN gateway interface; marking the PSTN link as the preferred path in the communication protocol stack; and potentially marking problematic IP network links as "standby" or "failed." After this step is completed, all subsequent communication traffic (such as new call requests and ongoing audio / video streams) will automatically be transmitted through a stable and reliable PSTN link, thus achieving a smooth switch of the communication backbone with minimal or no impact on the user, ensuring service continuity.

[0064] The above implementation upgrades the crude handover method, which often relies on manual intervention or simple heartbeat detection in traditional communication systems, to a refined management process based on continuous performance perception and intelligent analysis. This not only allows for earlier and more accurate prediction of IP network quality degradation trends, thus initiating handover preparation before call quality is significantly impaired, but also ensures the safety and success of the handover action itself through rigorous readiness verification.

[0065] In practical applications, this technical solution enables the entire intelligent call system to have telecommunications-grade communication reliability. It ensures that when a local or temporary failure occurs in the complex wireless or wired network environment of the ward, the communication channel that maintains the lifeline between patients and nurses can be switched to the backup PSTN line instantly and without interruption. This fundamentally eliminates the risk of the call system being paralyzed due to a single network problem, and provides underlying communication protection for ward safety and nursing efficiency.

[0066] Reference Figure 4 As one implementation of step S104, the step of parsing the nursing level and call type in the structured call request data packet, and generating a nurse terminal broadcast policy instruction in conjunction with the target nurse terminal group identifier list includes: Step S401: Parse the protocol header fields of the structured call request data packet and extract the care level and call type identifier; The protocol header is the management information area in the data packet used to describe the attributes of the data itself. The extracted "Nursing Level Identifier" represents the urgency and resource demand of the patient associated with this call; it is a business priority label derived from the hospital's nursing policy. The "Call Type Identifier" directly represents the specific intent of the patient initiating the call, such as a regular voice call request, a video call requiring face-to-face visual communication, or a highest-priority emergency alarm.

[0067] Understandably, the two extracted identifiers together constitute the most essential business attribute profile of this call event: one is the "demand-side" feature (care level) based on the patient's status, and the other is the "intent-side" feature (call type) based on the patient's active expression.

[0068] Step S402: Query the preset broadcast rule library according to the nursing level and match the corresponding basic broadcast strategy template; Specifically, the system maintains a pre-defined broadcast rule base, which is essentially a lookup table mapping different "nursing level identifiers" to corresponding "basic broadcast strategy templates". For example, the rule base might define: the identifier "P0" (special care) maps to the "all-staff emergency broadcast template", the core strategy of which is "notify all on-duty nurses in the ward and preempt communication resources with the highest priority"; the identifier "P1" (level 1 care) might map to the "responsible ward broadcast template", the strategy of which is "notify all nurses in the patient's nursing ward"; and the identifier "P2" (level 2 care) might map to the "dedicated responsible nurse broadcast template".

[0069] Understandably, transforming the clinical nursing guidelines' provisions regarding the scope of attention and response levels for patients at different nursing levels into deterministic strategy instruction templates that can be automatically executed by computers ensures that the system's most basic response behavior is strictly aligned with the management requirements of medical care, providing a standardized and safe strategy baseline for subsequent flexible optimization.

[0070] Step S403: Obtain the distribution data of terminal device types in the target nurse terminal group identifier list; The target nurse terminal group identifier list specifies the group of nurses who should, in principle, receive notifications (such as all nurses within a responsibility area). The system needs to further obtain device type distribution data for each terminal included in this list. This includes counting the number and proportion of different types of terminals, for example: how many smart PDAs support high-definition video calls, how many ordinary telephones support only voice calls, and how many door screens can only receive text prompts.

[0071] Understandably, real-time statistical analysis of hardware resources within the target notification range provides an objective resource capability profile for dynamic strategy optimization. Terminal device type distribution data reveals the capabilities and resource bottlenecks of currently available communication media. For example, if the list mostly consists of terminals that only support voice, then an attempt to make a video call may face obstacles.

[0072] Step S404: Combining the call type identifier and terminal device type distribution data, perform strategy optimization calculation to generate broadcast strategy adjustment parameters; The system compares and calculates the extracted call intent with the current resource status in real time, thereby refining and contextualizing the established basic strategies. The system performs strategy optimization calculations, with inputs including call type identifiers and terminal device type distribution data. The calculation logic involves examining and adapting current resource capabilities based on the call intent.

[0073] In one embodiment of this application, the strategy optimization calculation includes: calculating the proportion of terminals supporting video communication in the target nurse terminal group; generating adjustment parameters to increase voice terminals when the call type identifier is a video call and the proportion of video terminals is <60%; and generating adjustment parameters to disable the video transmission channel when the call type identifier is a voice call.

[0074] Specifically, when the call type is a video call, but calculations show that the proportion of video-enabled terminals in the target terminal group is below a certain success rate threshold (e.g., 60%), the optimization algorithm generates broadcast policy adjustment parameters to indicate that "while prioritizing video terminals, some voice terminals should also be included in the alternative notification range" as a fallback communication guarantee if the video call is unreachable. Conversely, if the call type is a "voice call," the optimization algorithm may generate parameters to "disable the video transmission channel" to save network bandwidth and terminal processing resources. This scheme fine-tunes the standard policy based on the actual resource situation at the current moment, aiming to maximize the probability of successful call request acceptance and optimize the overall resource utilization efficiency.

[0075] Step S405: Based on the basic broadcast strategy template and broadcast strategy adjustment parameters, generate a nurse terminal broadcast strategy instruction that includes the target terminal group range, broadcast priority, and retry mechanism.

[0076] Specifically, the target terminal group range field integrates the initial notification range determined by the base template and any possible additions made by optimization parameters, clarifying the final list of devices that need to be notified. The broadcast priority field directly inherits the priority corresponding to the care level mapped by the base template, determining the order of the call in the system's internal processing queue. The retry mechanism field specifies how the system should behave when the first broadcast does not receive a response or communication fails (such as the number of retries, the retry interval, and whether to eventually switch to a backup communication link).

[0077] In this embodiment, the nurse terminal broadcast strategy command is sent to the system's broadcast execution module. This module will strictly follow the scope, priority, and retry rules in the command to send notifications to the designated nurse terminals. In this way, abstract nursing levels, call intentions, and real-time resource status are ultimately transformed into precise control signals that drive the coordinated operation of the entire multi-terminal linkage system.

[0078] The above implementation achieves core intelligent scheduling of the entire process from a patient call being issued to being accurately, efficiently, and reliably notified to the corresponding nursing staff. This technical solution begins with the precise decoding of the inherent business attributes of the call, using this as a basis to match predetermined nursing response standards and form a strategy benchmark. Next, it introduces real-time perception of the capabilities of terminal devices within the target notification range, incorporating the current resource status into decision-making. Based on this, an optimized calculation model based on call intent and resource matching degree dynamically and flexibly adjusts the benchmark strategy to address complex situations such as resource imbalances in actual deployment. Finally, all decision-making elements are integrated to generate a detailed instruction that can be executed immediately.

[0079] In practical applications, this technical solution transforms broadcast notifications from simple one-to-many group messaging into a rule-based, environment-aware, and dynamically optimized intelligent resource scheduling system. This ensures that, while adhering to strict nursing standards, each call can reach the target nurse in the most likely way and with the most efficient use of existing terminal resources. This improves the first-response rate and overall communication reliability at the system level, making it a key intelligent decision-making hub supporting the intelligent call system in achieving "multi-terminal linkage and efficient response."

[0080] Reference Figure 5 As one implementation of step S106, the step of generating a nurse response priority sequence based on nursing level and response status data using a dynamic weighting model includes: Step S501: Query the weight configuration database according to the nursing level to obtain the basic weight coefficients of distance factor, task load factor and skill matching factor in the pre-configured dynamic weight model; The system does not apply a fixed evaluation standard to all calls. Instead, it first queries the weight configuration database based on the care level identifier (such as P0 critical, P1 urgent, P2 routine) carried by the call itself.

[0081] Specifically, the weight configuration database pre-stores "basic weight coefficient" configuration schemes for the three core evaluation dimensions—"distance factor," "task load factor," and "skill matching factor"—for different nursing level scenarios. For example, for a "P0 emergency" call, the core challenge is resolving highly complex professional medical problems; therefore, the weight of the "skill matching factor" will be set to the highest (e.g., 0.7), while the weight of the "distance factor" will be relatively lower (e.g., 0.2). This means that the system will prioritize finding the nurse with the best skill match rather than the nurse who is closest in distance. Conversely, for a "P2 routine" call, the main requirement is to quickly respond to daily requests; therefore, the weight of the "distance factor" will be set to the highest.

[0082] Understandably, this step makes subsequent quantitative calculations no longer static and blind, but rather gives them a clear and variable business orientation, ensuring that scheduling strategies can intelligently adapt to clinical scenarios with different levels of urgency, and implementing the principle of precision nursing based on hierarchical classification from the source of decision-making.

[0083] In some embodiments, the configuration rules for the basic weight coefficients include: when the care level is P0 (critical): distance factor weight 0.2, task load weight 0.1, skill matching weight 0.7; when the care level is P1 (urgent): distance factor weight 0.5, task load weight 0.3, skill matching weight 0.2; when the care level is P2 (routine): distance factor weight 0.7, task load weight 0.2, skill matching weight 0.1.

[0084] Step S502: Calculate the spatial distance parameter based on the terminal position coordinates of each nurse's terminal and the coordinate data of the patient's terminal; After obtaining the real-time location coordinates of the nurse's terminal and the patient's terminal, the system does not simply calculate the straight-line Euclidean distance between them. This is because in complex ward environments (with walls, corridor corners, and work area isolation), the straight-line distance often cannot reflect the actual walking path.

[0085] In this embodiment, the shortest feasible path distance from the nurse's current location to the patient's bed can be found by calculating spatial distance parameters. A path planning algorithm based on a "3D spatial topology map of the ward" (i.e., a digital map containing information such as passageways, doors, and obstacles) is employed, such as Dijkstra's algorithm. Subsequently, to eliminate the influence of differences in absolute distance scales between different wards and floors on the model, the path distance value is normalized, for example, by converting it into a relative value between 0 and 1, to provide a fair and realistic measure of physical accessibility. The smaller the value, the fewer spatial obstacles the nurse needs to overcome to reach the patient, providing an objective spatial dimension input for subsequent comprehensive evaluation.

[0086] Step S503: Analyze the current task count in the nurse terminal response status data and generate task load quantization parameters; Specifically, the current task counter value (a continuous integer) reflecting a nurse's workload is mapped to a standardized load parameter that characterizes their immediate ability and willingness to handle new calls. Directly using the raw task counter value is unfair because the impact of different values ​​on nurse workload is non-linear. For example, the perceived increase in workload is significant when going from 0 to 1 tasks; however, the marginal increase in perceived workload may be relatively small when going from 5 to 6 tasks.

[0087] Therefore, the system can generate task load quantification parameters by parsing the count values ​​in the response status data and based on a preset mapping relationship. This mapping relationship is usually a piecewise function or lookup table, which maps continuous ranges of count values ​​to discrete standard parameter values ​​representing different load levels (e.g., 0.1 represents light load, 0.5 represents medium load, and 0.9 represents heavy load). This step achieves discretization and normalization of workload assessment, allowing the differences in workload between nurses due to varying task quantities to be incorporated into the overall assessment model as comparable parameters under a unified dimension. This avoids unreasonably low overall scores due to individual nurses already undertaking too many tasks, reflecting a balanced allocation of human resources.

[0088] Step S504: Extract the nursing skill matching degree from the nurse terminal response status data; Among them, the nursing skills matching degree is a pre-calculated or real-time value that quantifies the degree of matching between a specific nurse's professional skills (such as intravenous puncture, electrocardiogram monitoring, and postoperative care) and the nursing needs implied by the current patient based on their condition, stage of treatment, and call type. This matching degree is derived from the automatic comparison between the nurse's skills database and the patient's care plan.

[0089] In this step, the system directly extracts from the response status data, thus introducing the crucial professional dimension of nursing quality and safety. A nurse with highly matched skills can handle specific calls more accurately and efficiently, especially for technically demanding needs. This directly relates to nursing outcomes and patient safety, elevating intelligent dispatching to a level of professional adaptation.

[0090] Step S505: Input the spatial distance parameters, task load quantification parameters and nursing skill matching degree into the dynamic weight model, and generate the comprehensive priority score of each nurse terminal through weighted calculation based on the basic weight coefficients. Specifically, three standardized parameters (spatial distance parameter, task load quantification parameter, and nursing skill matching coefficient) representing "spatial accessibility", "immediate availability" and "professional fit" respectively, along with three basic weight coefficients representing decision-making orientation in the current scenario, are input into the dynamic weight model.

[0091] In this embodiment, the core of the model is a weighted summation function. The essence of weighted calculation is to comprehensively weigh considerations across different dimensions: each parameter is multiplied by its corresponding weight coefficient, and then the results are summed to obtain a comprehensive priority score. The weight coefficient determines the weight of each dimension in the final decision, and this weight is dynamically set by the nursing level. Through weighted calculation, a nurse with outstanding skills but located further away in an emergency may receive a higher comprehensive score than a nurse who is closer but has mismatched skills or is extremely busy. This step, through a mathematical model, integrates multi-dimensional and potentially conflicting optimization objectives into a single, comparable scalar value, providing a precise basis for the final ranking and realizing the transformation from multi-dimensional information to a unified decision instruction.

[0092] Step S506: Perform a descending sorting operation on the comprehensive priority scores of all nurse terminals to generate a nurse response priority sequence.

[0093] Among them, the nurse terminal with the highest score represents the most suitable responder after weighing distance, load, and skills under the current dynamic weight model evaluation, and is therefore ranked first in the sequence.

[0094] In this application embodiment, the nurse response priority sequence clarifies the precise order in which the system subsequently attempts to establish a connection and assign tasks. This sequence is dynamic and personalized, and is generated specifically for this particular call.

[0095] In the above implementation, based on the nursing level, the decision weights of three core factors—distance, workload, and skills—are dynamically configured. Physical distance is precisely quantified based on actual paths, workload is non-linearly standardized and mapped, and the key quality indicator of nursing skill matching is directly introduced, thus completing the refined preparation of multi-dimensional assessment data. Next, a weighted calculation model integrates these discrete dimensions into a unified comprehensive priority score, achieving the optimal trade-off among multiple objectives.

[0096] Finally, by sorting and outputting a clear sequence of actions, the system achieves scientific, intelligent, and precise scheduling of nursing human resources. This enables the system to go beyond simple geographical proximity or polling principles. When a call occurs, the system can comprehensively consider and intelligently adjust the priority of the three factors according to the urgency of the event. This optimizes response efficiency and nursing quality at the system level, ensuring that the most suitable nurse provides the most timely and professional response to the patient at the right time. This fundamentally improves the collaborative efficiency and safety of ward nursing work.

[0097] Reference Figure 6 As a further implementation of the ward intelligent calling method, after establishing a two-way communication session between the patient terminal and the target nurse terminal, the method further includes: Step S601: Collect physiological parameters such as heart rate, blood pressure and blood oxygen saturation fed back by the patient terminal in real time, and output physiological monitoring dataset; Among these measures, smart terminals deployed on the patient's side continuously and synchronously quantify and convert data on three key physiological parameters reflecting the core stability of vital signs. Heart rate can be indirectly measured using photoplethysmography (PPG) technology, which calculates heart rate by utilizing the characteristic that blood's absorption of specific wavelengths of light changes periodically with the pulse. Blood pressure can be estimated using the oscillometric method by detecting the oscillating waves generated by arterial wall vibration during cuff decompression. Blood oxygen saturation is calculated based on the difference in absorption spectra of oxyhemoglobin and deoxyhemoglobin for different wavelengths of red and infrared light through ratio calculation.

[0098] Next, these biophysical signals are captured at a fixed sampling rate (e.g., once per second) and converted into digital signals. After preprocessing such as filtering, the resulting physiological monitoring dataset is no longer an isolated instantaneous reading, but a multidimensional time series containing parameter values, temporal relationships, and trends.

[0099] Understandably, this step links the patient's subjective "call for help" behavior with an objective and continuous stream of vital signs data, providing a crucial objective data source for the system to upgrade from a passive response to an active and intelligent clinical status monitoring and early warning system.

[0100] Step S602: Associate and encapsulate the physiological monitoring dataset with the patient identifier in the structured call request data packet to generate an enhanced call data packet; Specifically, physiological monitoring datasets are forcibly bound to patient identifiers inherent in the data packets to ensure that physiological data, after encapsulation, can still be clearly attributed to a specific patient, avoiding confusion in multi-patient data streams. For example, the data packet will simultaneously record "Patient Zhang San initiated a video call" and "At the moment of the call, his heart rate was 125 beats / min and his blood oxygen saturation was 90%".

[0101] Understandably, this step creates a highly complete clinical event data object, enabling any subsequent processing steps (such as decision analysis and information broadcasting) to be based on a complete information profile that integrates the patient's subjective needs and objective physiological state, thus laying a solid data foundation for precision nursing intervention.

[0102] Step S603: Send a data query request containing the patient identifier to the hospital information system to obtain the patient's allergy history and current medication records from the electronic medical record; The underlying logic of this step is to use a standardized medical information exchange protocol, with the currently calling patient as the index, to proactively and in real time extract key static file information that affects clinical decisions from the hospital's core information system.

[0103] Specifically, the system uses the patient identifier as a unique key to initiate precise queries to the Hospital Information System (HIS) and Electronic Medical Record System (EMR) through interfaces conforming to medical information exchange standards. The acquired allergy history information serves as a personalized medication safety guideline for each patient, identifying specific drugs or substances that must be absolutely avoided. Current medication records reflect the patient's ongoing medication regimen and are an indispensable medication context for assessing any new symptoms or changes in signs (such as sudden increases in blood pressure or abnormal heart rate).

[0104] It should be noted that this step breaks through the limitations of traditional call systems that only focus on instant communication. It intelligently links bedside dynamic monitoring with the patient's long-term, static medical records, enabling the system to supplement medical staff with two crucial background knowledge points, "what the patient is allergic to" and "what medications the patient is currently using," the moment a call occurs. This provides key static data support for correctly understanding and handling the current call event and avoiding medical errors.

[0105] Step S604: Construct an auxiliary decision-making information tree based on allergy history, current medication records, and physiological monitoring dataset; This involves organizing discrete data (real-time vital signs, allergy history, medication records) from different sources and of different natures into a hierarchical information model with internal logical connections, namely, an auxiliary decision-making information tree, in a way that is machine-processable and easy for nurses to understand.

[0106] In this embodiment, the tree is not constructed by randomly piling up data, but by semantic fusion based on clinical thinking and nursing pathways. The root is the patient identifier, from which branches out a "signs layer" representing the current state (such as real-time values ​​of heart rate, blood pressure, and blood oxygen, and comparisons with normal thresholds), a "medication layer" representing the treatment background (such as the name, dosage, and frequency of currently used drugs), and a "risk layer" representing risk warnings (such as allergens and drug contraindications). During the construction process, the system establishes logical connections between different data nodes based on a built-in medical knowledge base or rule engine. For example, associating "warfarin" (an anticoagulant) in the "current medication record" with abnormal "elevated blood pressure" data in the "physiological monitoring dataset" may dynamically generate a "higher risk of bleeding" warning node in the risk layer.

[0107] Understandably, this step transforms fragmented, multi-source data into a clinically relevant "decision support dashboard" that presents the patient's condition, treatment context, and potential risks in a clear structure. The goal is to help nurses quickly grasp the full picture of the patient's condition and key points of treatment within the very short time it takes to answer a call.

[0108] Step S605: When an anomaly is detected in the physiological monitoring dataset, an enhanced call data packet and an auxiliary decision information tree are broadcast to the top N nurse terminals in the nurse response priority sequence. The system continuously acquires physiological monitoring datasets and performs real-time analysis based on pre-set anomaly judgment rules that conform to clinical guidelines (such as a heart rate consistently above 120 beats / min, systolic blood pressure exceeding 180 mmHg, and blood oxygen saturation below 92%). Once an anomaly is detected, a broadcast process is automatically triggered. The broadcast is not indiscriminate but highly precise, targeting the top N nurse terminals in a pre-calculated "nurse response priority sequence" calculated by an independent algorithm. The value of N can be dynamically adjusted according to the level of care or the severity of the anomaly, ensuring that enough and the most appropriate responsible personnel are notified in critical situations.

[0109] Furthermore, the broadcast content includes both enhanced call data packets reflecting real-time events and auxiliary decision information trees providing clinical context and decision support. When a patient may be unable or unwilling to call due to worsening condition, the system can proactively detect the crisis and guide the intervention of the most appropriate nursing resources, while providing sufficient decision support information, thus upgrading the response mode from passive response to proactive monitoring and intelligent early warning.

[0110] Step S606: When the two-way communication session is interrupted or switched, all communication data packets of the current session and the auxiliary decision information tree are synchronously stored in the nursing record database of the hospital information system.

[0111] In the mobile environment of the ward, the two-way communication session between the nurse's handheld terminal (PDA) and the patient's bedside device may be unexpectedly interrupted due to network switching, signal obstruction, or device battery power issues. The system will immediately activate the data synchronization and storage mechanism upon detecting a session interruption or active switching.

[0112] Specifically, its storage objects not only include all audio and video streaming communication data packets generated during the session, but more importantly, it synchronously saves the auxiliary decision-making information tree closely related to that session. This ensures that every doctor-patient interaction, regardless of whether it has completely ended, involves key information (the patient's chief complaint, the nurse's inquiries, and the decision support provided by the system) that is fully recorded in the patient's official electronic nursing record. This provides information context for subsequent continuous diagnosis and treatment, and forms a complete chain of evidence that complies with medical quality management and legal norms, realizing a digital full record of the nursing process.

[0113] In the above implementation, objective physiological data is added to the call event, enabling the system to perceive changes in the patient's inexpressible state. By integrating electronic medical records, the current event is infused with the patient's individualized history and treatment context. Furthermore, by constructing an auxiliary decision-making information tree, scattered data is transformed into structured clinical insights. When physiological abnormalities are detected, the system can automatically trigger precise warnings and efficiently push key information to the optimal nursing staff. Finally, a data synchronization and storage mechanism ensures that all interaction traces are completely preserved.

[0114] In practical applications, this technology not only provides strong information support when patients actively seek help, but also provides proactive warnings when there are subtle changes in the patient's condition, and ensures that all nursing interactions are traceable. This enhances nursing safety, medical quality, and risk management capabilities while improving response speed and accuracy.

[0115] Reference Figure 7 As a further implementation of the ward intelligent calling method, after establishing a two-way communication session between the patient terminal and the target nurse terminal, the method further includes: Step S701: Real-time identification of key nursing instruction text in the two-way communication session to generate structured nursing task items; The system utilizes an automatic speech recognition engine that integrates a medical knowledge graph to continuously monitor conversational audio streams. By accurately identifying and extracting key nursing instructions embedded in everyday conversations—instructions that often include professional drug names, complex operational terminology, and dosage parameters—it can effectively capture and extract these instructions.

[0116] Specifically, the speech recognition engine not only converts speech to text, but more importantly, it performs deep natural language processing and named entity recognition. It needs to identify from the sentences the "operational entities" (such as "injection" and "dressing change") representing nursing behaviors, the "medical object entities" involved (such as drug names and device names), and the specific "parameter entities" (such as dosage "5mg" and frequency "bid").

[0117] Subsequently, the system combines these identified discrete entities according to preset semantic rules and data structures to generate a structured nursing task item. This task item clarifies "what to do" (operation), "what to do" (object), and "how to do it" (parameters), thereby transforming vague and ambiguous verbal instructions into a standardized work order that can be directly scheduled and executed by a computer program.

[0118] Step S702: Query the clinical pathway database in the hospital information system based on the patient identifier and match the standard operating procedure corresponding to the nursing task item; The clinical pathway database consists of standardized treatment and nursing plan templates developed by hospitals for specific diseases or surgeries. Through semantic matching and contextual adaptation, it determines whether a current "structured nursing task" (e.g., injecting insulin into a patient) conforms to the preset set of operations for the patient's current treatment stage (e.g., "perioperative management pathway for diabetes"). The matching process considers the necessity and sequence of the task and automatically associates it with the corresponding standard operating procedure (SOP) within the pathway. The SOP may include a checklist of preparation tasks, key operational steps, precautions, and risk warning points. For example, a task item like "central venous catheter maintenance" would, after matching, generate a SOP that details aseptic procedures, the selection and dosage of flushing and sealing fluid, and key points for catheter assessment.

[0119] Understandably, this step anchors nurses' individual experience-based operations to institutionalized, optimal clinical practice guidelines, ensuring that every nursing instruction issued complies with regulations and automatically includes complete execution standards, effectively reducing safety risks caused by inconsistent operating procedures or omissions of key steps.

[0120] Step S703: Based on the target nurse's terminal identifier and current task count, generate an electronic nursing work order and assign an execution priority; Based on real-time human resources conditions, the most suitable executors are identified for standardized and contextualized nursing tasks, and scientific execution sequences are assigned.

[0121] In this embodiment, generating an electronic nursing work order is not simply a matter of binding tasks to nurses, but rather a multi-objective optimization decision. The system needs to comprehensively evaluate multiple factors, including the nurse's immediate workload (reflected by task counts), location, and the match between professional skill tags and the current task. Based on this, an execution priority is assigned to the work order. This priority is dynamic and must weigh multiple dimensions: the urgency of the nursing task itself (such as a critical task triggered by a patient's physiological parameters), the time window requirements of the task (such as a scheduled medication administration task), and the overall nurse load balancing. For example, a non-urgent routine task may be assigned to a nurse with a lighter current load, while an emergency resuscitation task will be assigned with the highest priority (potentially interrupting lower-priority tasks) to the nearest qualified nurse, regardless of their current load.

[0122] Understandably, this step aims to systematically optimize the allocation of nursing human resources across the entire ward, ensuring the fastest possible response to emergency tasks while also striving to distribute workload evenly and improve overall work efficiency and nurses' work experience.

[0123] Step S704: Synchronously push the electronic nursing work order to the task queue and nursing management database of the target nurse's terminal; The system will simultaneously push the generated electronic nursing work orders to two destinations: one is the "task queue of the target nurse terminal", which is the instruction channel that drives the front-end execution. It usually uses low-latency message queuing technology to ensure that nurses can receive new task reminders on their mobile devices in an instant; the other is the "nursing management database", which is the authoritative data storage for recording, auditing and management.

[0124] Understandably, this "dual-write" operation must ensure transactional consistency, guaranteeing that the data states in both locations are synchronized. Pushing to the terminal is to drive action, adding tasks to nurses' personal to-do lists; while writing to the central database is to create authoritative records, providing a data source for task dashboards across the entire ward, management supervision, and subsequent quality analysis.

[0125] Step S705: When the nurse terminal detects that it has scanned the patient's wristband or the drug barcode, verify the matching degree between the operation and the work order content, and update the task execution status to the nursing record database.

[0126] The system detects when a nurse's terminal scans a patient's wristband or medication barcode, marking the transition of a nursing task from preparation to execution. The scanning action itself triggers a data flow convergence: the physical identification codes obtained from the scan (patient wristband ID, medication NDC code) are transmitted back to the system in real time.

[0127] Next, the system performs a verification operation, precisely matching the captured entity identity information with the instruction information recorded in the electronic nursing work order ("for which patient", "what medication to use"). If the verification passes, it confirms that "the correct nurse is performing the correct operation on the correct patient", and automatically updates the task execution status, recording the completion time, executor, associated patient and medication IDs, and other information in the nursing record database.

[0128] It should be noted that this step transforms the traditional process of relying on manual verification and post-event recording into an automated hard verification and strong recording that occurs instantly at the operation point through the collaboration of barcodes, terminals, and systems. This not only fundamentally prevents common patient identification and medication errors, but also automatically generates a complete, accurate, and timestamped operation traceability chain, realizing a digital and automated closed loop in the nursing execution process.

[0129] The above implementation frees nurses from the traditional model of tedious verbal order translation, relying on memory to find operating procedures, manually coordinating task allocation, and handwritten record of execution results. It ensures that every nursing instruction conforms to best practices from the moment it is generated and is assigned to the most appropriate executor. During the execution process, mandatory barcode verification eliminates human error. After the execution is completed, a structured and traceable electronic record is automatically generated.

[0130] In practical applications, this technical solution systematically improves the safety, standardization, and overall efficiency of nursing work, upgrading nursing management from an experience-driven, manual coordination-based model to a data-driven, system-automated optimization-based intelligent collaborative model. This not only reduces the risk of medical errors and optimizes human resource allocation but also generates digital assets for the nursing process, laying a solid technical foundation for continuous medical quality improvement and refined management.

[0131] Reference Figure 8 As a further implementation of the intelligent calling method for wards, after the two-way communication session switches or ends, it also includes: Step S801: Collect the performance metrics set of the entire call link, including response latency, session duration and number of handovers; The system uses integrated or built-in monitoring probes to collect detailed performance metrics throughout the entire process, from when a patient initiates a call to when the nursing interaction is completely completed.

[0132] Specifically, the "Response Latency Measure" assesses the initial efficiency of system-human collaboration, precisely capturing the time difference between the moment the call request signal leaves the patient's terminal and the moment the first nurse makes a definite response (such as clicking to answer) on their terminal. This metric comprehensively reflects network transmission latency, system processing time, and the nurse's initial reaction time. The "Session Duration Record" tracks the duration of effective audio / video communication or data interaction, distinguishing between pure connection maintenance time and actual communication time, and is used to evaluate the sufficiency and efficiency of the interaction. The "Switchover Count" counts the number of times the session participants are transferred between different nurse terminals during a single session due to network quality, nurse movement, or proactive scheduling. This is a key indicator for measuring communication stability and scheduling smoothness.

[0133] By collecting these indicator sets, a complex nursing communication event involving multiple stages can be deconstructed into several quantifiable and comparable key performance indicators, thus providing an objective data foundation for subsequent evaluation of system effectiveness and identification of performance bottlenecks.

[0134] Step S802: Based on the patient identifier and the call request timestamp, extract the nursing operation execution timeline associated with the patient identifier from the nursing record database, which includes the nursing work order generation time, barcode verification time, and task completion time. Specifically, the system precisely aligns and links the "call" event, representing the start of communication, with the nursing operation execution record, representing the outcome of the nursing action, along the time dimension. The system uses a unique "call request ID" assigned when the call event is generated and a consistent "patient identifier" as dual index keys to actively query the nursing record database. This database typically records various standardized nursing operations (such as medication administration, dressing changes, and monitoring) performed and confirmed by nurses via mobile terminals after a call response; each record is timestamped. Extracting these records and arranging them chronologically forms a nursing operation execution timeline, including the nursing work order generation time, barcode verification time, and task completion time.

[0135] Understandably, by overlaying and comparing the call session period with the nursing operation period through time series analysis, it is possible to analyze the time interval (preparation time) from the end of communication to the start of action, as well as whether the execution time and sequence of each operation are reasonable.

[0136] Step S803: Based on the performance index set and the nursing operation execution time axis, construct a nursing quality assessment vector, including call response speed parameter, operation execution accuracy parameter, and resource utilization parameter; Among them, the nursing quality assessment vector is a comprehensive measurement model that integrates three core value dimensions: response efficiency, execution accuracy, and resource efficiency.

[0137] Specifically, the call response speed dimension is calculated by the ratio of the collected response delay to the preset standard delay based on the level of care, quantifying how quickly the system responds to patient needs. The operation execution accuracy dimension is calculated based on the ratio of the total number of operation items in the nursing operation record to the number of barcode verifications passed, quantifying the quality and safety of nursing actions from the perspective of operational standardization. The resource utilization efficiency dimension is a more macro-level performance indicator, which may comprehensively calculate the percentage of nurses' effective working hours in direct nursing activities and the turnover rate of key medical equipment, aiming to assess whether the system's allocation and use of human and material resources are efficient.

[0138] Understandably, by using a structured mathematical vector, a multi-dimensional performance profile can be created for the output of a single nursing service or over a period of time, transforming the complex service quality into a standardized data object that can be used for machine learning and further analysis.

[0139] Step S804: Input the nursing quality assessment vector into the pre-trained feedback optimization model and output nursing process improvement parameters, including the factor coefficient adjustment amount in the dynamic weight model and the update instructions of the broadcast rule base. The feedback optimization model is essentially an intelligent agent that can learn from historical experience and the current state and make decisions. The generated nursing quality assessment vector is input into the model, and the model's function is to analyze the current quality performance and decide how to adjust the system's own operating parameters to obtain better quality output in the future.

[0140] Specifically, the feedback optimization model can be trained using machine learning techniques (such as reinforcement learning). Its training objective is to maximize the long-term accumulated "quality rewards," which may be based on composite indicators such as patient satisfaction, nurse efficiency, and resource savings. After internal calculations and inferences, the model outputs a set of specific, actionable adjustment instructions for improving nursing processes. For example, the model might analyze and find that the response speed is acceptable but resource efficiency is low, thus suggesting that the output parameters appropriately reduce the weight of "rapid response" and increase consideration for "load balancing."

[0141] In some embodiments, the feedback optimization model may employ a reinforcement learning framework, with the state space being the nursing quality assessment vector, the action space being the weight coefficient adjustment range or the matching rule adjustment strategy, and the reward function being R = patient satisfaction × 0.7 + nurse operational efficiency × 0.3. The strategy network is iteratively optimized using the Q-learning algorithm.

[0142] Step S805: Based on the nursing process improvement parameters, update the factor coefficients of the dynamic weight model and the matching rules of the broadcast rule base.

[0143] The system improvements involve dynamically adjusting key configurations in its core decision-making logic. This mainly involves two aspects: First, updating the factor coefficients in the dynamic weighting model. For example, when calculating nurse response priorities, the weight ratios of factors such as "distance," "nurse workload," and "skill matching" are adjusted to change the tendency of task allocation. Second, updating the broadcast rule base. For example, modifying the range of nurses that trigger broadcasts for events of different nursing levels, the priority judgment threshold, or the message retry strategy.

[0144] It should be noted that the system can ensure a smooth and controllable parameter update process through technologies such as hot loading, version control, and secure rollback. Through data feedback on actual operating results, the system can continuously iterate and optimize its scheduling strategies and response rules, thereby achieving gradual improvement in the quality of nursing services and continuous improvement in system efficiency.

[0145] In the above implementation, the abstract service quality is transformed into measurable performance indicators and operation timelines, and then a multi-dimensional evaluation vector is used to comprehensively quantify and diagnose it. Subsequently, a pre-trained optimization model is used to conduct in-depth analysis of the diagnostic results and generate specific improvement strategies. The system automatically applies these strategies to the core scheduling algorithm and rule base to achieve iterative updates of behavior.

[0146] In practical applications, this technical solution transforms the intelligent call system of this application from a static, rule-fixed tool into a dynamic, continuously learning organism. The system can automatically adjust its internal parameters and rules based on feedback from actual operation, thereby finding and approaching the optimal balance point in the current environment among multiple interrelated and sometimes even mutually restrictive goals such as "call response speed," "accuracy of nursing operations," and "resource utilization efficiency," ultimately achieving autonomous, continuous, and intelligent improvement in the quality of nursing services.

[0147] This application also discloses a ward intelligent call system based on dual-mode communication and multi-terminal linkage.

[0148] A ward intelligent call system based on dual-mode communication and multi-terminal linkage specifically includes: The call request standardization module is used to receive trigger signals sent by the patient terminal and generate a structured call request data packet containing the patient identifier, care level and call type based on the trigger signals. The communication link redundancy switching module is used to obtain the real-time communication quality parameters of the IP network. When the real-time communication quality parameters are detected to be lower than the preset parameter threshold, the PSTN communication link is activated as the main transmission channel. The data routing and target group decision module is used to send structured call request data packets to the call softswitch through the main transmission channel, and the call softswitch queries the nurse scheduling database to obtain the target nurse terminal group identifier list. The broadcast policy intelligent generation module is used to parse the nursing level and call type in the structured call request data packet, and combine it with the target nurse terminal group identifier list to generate nurse terminal broadcast policy instructions; The multi-terminal status synchronization acquisition module is used to broadcast structured call request data packets to the target nurse terminal group according to the broadcast strategy instructions, and receive response status data returned by each nurse terminal, including terminal location coordinates, current task count and nursing skill matching degree. The dynamic priority scheduling module is used to generate a nurse response priority sequence based on nursing level and response status data through a dynamic weight model. The intelligent audio and video session establishment module initiates an audio and video connection request to the target nurse terminal, which is at the top of the nurse response priority sequence, to establish a two-way communication session between the patient terminal and the target nurse terminal. The session switching module is used to monitor the communication signal strength and the target nurse terminal's movement speed in real time during the bidirectional communication session. When the communication signal strength is continuously lower than the preset interruption threshold or the terminal's movement speed continuously exceeds the safety threshold, the bidirectional communication session is switched to the nurse terminal ranked second in the nurse response priority sequence.

[0149] The ward intelligent call system based on dual-mode communication and multi-terminal linkage according to the embodiments of this application can implement any of the above methods, and the specific working process of each module in the system can refer to the corresponding process in the above method embodiments.

[0150] In the several embodiments provided in this application, it should be understood that the provided methods and systems can be implemented in other ways. For example, the system embodiments described above are merely illustrative; for example, the division of a certain module is merely a logical functional division, and in actual implementation there may be other division methods, such as multiple modules can be combined or integrated into another system, or some features can be ignored or not executed.

[0151] This application also discloses a computer-readable storage medium.

[0152] A computer-readable storage medium storing a computer program that can be loaded by a processor and executed as described above in any of the methods of a ward intelligent calling method based on dual-mode communication and multi-terminal linkage.

[0153] The computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in connection with an instruction execution system, apparatus, or device; the program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to wireless, wire, optical fiber, RF, etc., or any suitable combination thereof.

[0154] The above are all preferred embodiments of this application and are not intended to limit the scope of protection of this application. Any feature disclosed in this specification (including the abstract and drawings) may be replaced by other equivalent or similar features unless specifically stated otherwise. That is, unless specifically stated otherwise, each feature is only one example of a series of equivalent or similar features.

Claims

1. A ward intelligent calling method based on dual-mode communication and multi-terminal linkage, characterized in that, The method includes: Receive a trigger signal sent by the patient terminal, and generate a structured call request data packet containing the patient identifier, care level, and call type based on the trigger signal; The system acquires real-time communication quality parameters of the IP network. When the real-time communication quality parameters are detected to be lower than a preset parameter threshold, the PSTN communication link is activated as the main transmission channel. The structured call request data packet is sent to the call softswitch through the main transmission channel, and the call softswitch queries the nurse scheduling database to obtain the target nurse terminal group identifier list. The nursing level and call type in the structured call request data packet are parsed, and combined with the target nurse terminal group identifier list, a nurse terminal broadcast policy instruction is generated; According to the broadcast strategy instructions, the structured call request data packet is broadcast to the target nurse terminal group, and the response status data returned by each nurse terminal is received, including terminal location coordinates, current task count and nursing skill matching degree; Based on the nursing level and response status data, a nurse response priority sequence is generated through a dynamic weighting model; An audio / video connection request is initiated to the target nurse terminal that is first in the nurse response priority sequence, establishing a two-way communication session between the patient terminal and the target nurse terminal; During the duration of the two-way communication session, the communication signal strength and the target nurse terminal's movement speed are monitored in real time. When the communication signal strength is continuously lower than a preset interruption threshold or the terminal's movement speed continuously exceeds a safety threshold, the two-way communication session is switched to the nurse terminal ranked second in the nurse response priority sequence.

2. The ward intelligent calling method based on dual-mode communication and multi-terminal linkage according to claim 1, characterized in that, The steps of receiving a trigger signal sent by a patient terminal and generating a structured call request data packet containing a patient identifier, care level, and call type based on the trigger signal include: Receive the trigger signal sent by the patient terminal, and extract the device type code and operation type code from the trigger signal; The calling device category identifier is determined based on the device type code and the preset device type mapping table; Obtain the current timestamp of the system clock and combine it with the patient's bed number to generate a globally unique call request identifier; Query the patient database of the hospital information system to obtain the patient identifier and nursing level associated with the patient's bed number; Based on the operation type code and the preset call type mapping rule, a corresponding call type identifier is generated; The call request identifier, patient identifier, care level, call device category identifier, and call type identifier are encapsulated into a structured call request data packet.

3. The ward intelligent calling method based on dual-mode communication and multi-terminal linkage according to claim 1, characterized in that, The steps of acquiring real-time communication quality parameters of the IP network and activating the PSTN communication link as the main transmission channel when the real-time communication quality parameters are detected to be lower than a preset parameter threshold include: The system periodically collects real-time communication quality datasets of the IP network through the network interface, including network latency measurements and packet loss statistics. The communication quality dataset is input into a preset network quality assessment model to generate network status assessment results. When the network status assessment result indicates that the communication quality parameters are lower than the preset parameter threshold, a link activation command is sent to the PSTN gateway; Receive the link ready status signal returned by the PSTN gateway and set the current main transmission channel as a PSTN communication link.

4. The ward intelligent calling method based on dual-mode communication and multi-terminal linkage according to claim 2, characterized in that, The steps for parsing the nursing level and call type in the structured call request data packet, and generating nurse terminal broadcast policy instructions in conjunction with the target nurse terminal group identifier list include: Parse the protocol header fields of the structured call request data packet to extract the care level and call type identifier; Based on the nursing level, query the preset broadcast rule base and match the corresponding basic broadcast strategy template; Obtain the distribution data of terminal device types in the target nurse terminal group identifier list; By combining the call type identifier with the terminal device type distribution data, the broadcast policy adjustment parameters are generated through policy optimization calculation. Based on the basic broadcast strategy template and broadcast strategy adjustment parameters, a nurse terminal broadcast strategy instruction is generated, which includes the target terminal group range, broadcast priority, and retry mechanism.

5. The ward intelligent calling method based on dual-mode communication and multi-terminal linkage according to claim 1, characterized in that, Based on the nursing level and response status data, the steps for generating a nurse response priority sequence using a dynamic weighting model include: Based on the nursing level query weight configuration database, obtain the basic weight coefficients of distance factor, task load factor and skill matching factor in the pre-configured dynamic weight model; Based on the coordinates of the nurse's terminal location and the coordinates of the patient's terminal, the spatial distance parameter is calculated. Analyze the current task count in the nurse terminal response status data to generate task load quantization parameters; Extract the nursing skill matching degree from the nurse terminal response status data; The spatial distance parameter, task load quantification parameter, and nursing skill matching degree are input into the dynamic weight model. Based on the basic weight coefficient, a comprehensive priority score for each nurse terminal is generated through weighted calculation. Perform a descending sort operation on the overall priority scores of all nurse terminals to generate a nurse response priority sequence.

6. A ward intelligent calling method based on dual-mode communication and multi-terminal linkage according to any one of claims 1 to 5, characterized in that, Following the step of establishing a two-way communication session between the patient's terminal and the target nurse's terminal, the following steps are also included: Real-time collection of physiological parameters such as heart rate, blood pressure, and blood oxygen saturation from patient terminals, and output of physiological monitoring dataset; The physiological monitoring dataset is associated and encapsulated with the patient identifier in the structured call request data packet to generate an enhanced call data packet; Send a data query request containing the patient identifier to the hospital information system to obtain the patient's allergy history and current medication records from the electronic medical record; Based on the allergy history, current medication records, and physiological monitoring dataset, construct an auxiliary decision-making information tree; When an anomaly is detected in the physiological monitoring dataset, the enhanced call data packet and auxiliary decision information tree are broadcast to the top N nurse terminals in the nurse response priority sequence. When the two-way communication session is interrupted or switched, all communication data packets of the current session and the auxiliary decision information tree are synchronously stored in the nursing record database of the hospital information system.

7. The ward intelligent calling method based on dual-mode communication and multi-terminal linkage according to claim 6, characterized in that, Following the step of establishing a two-way communication session between the patient terminal and the target nurse terminal, the method further includes: Key nursing instruction texts in the two-way communication session are identified in real time, and structured nursing task items are generated. The clinical pathway database in the hospital information system is queried based on the patient identifier to match the standard operating procedure corresponding to the nursing task item; Based on the target nurse's terminal identifier and current task count, an electronic nursing work order is generated and an execution priority is assigned. The electronic nursing work order is simultaneously pushed to the task queue and nursing management database of the target nurse's terminal; When a nurse's terminal detects that it has scanned a patient's wristband or medication barcode, it verifies the match between the operation and the work order content and updates the task execution status to the nursing record database.

8. The ward intelligent calling method based on dual-mode communication and multi-terminal linkage according to claim 7, characterized in that, After the bidirectional communication session switches or ends, the following is also included: Collect a set of performance metrics for the entire call chain, including response latency, session duration, and number of handovers; Based on the patient identifier and call request timestamp, extract the nursing operation execution timeline associated with the patient identifier from the nursing record database, which includes the nursing work order generation time, barcode verification time, and task completion time; Based on the performance index set and the nursing operation execution time axis, a nursing quality assessment vector is constructed, including call response speed parameter, operation execution accuracy parameter, and resource utilization parameter; The nursing quality assessment vector is input into a pre-trained feedback optimization model, which outputs nursing process improvement parameters, including the adjustment amount of factor coefficients in the dynamic weight model and update instructions for the broadcast rule base. Based on the nursing process improvement parameters, update the factor coefficients of the dynamic weight model and the matching rules of the broadcast rule base.

9. A ward intelligent call system based on dual-mode communication and multi-terminal linkage, characterized in that, The system includes: The call request standardization module is used to receive trigger signals sent by the patient terminal and generate a structured call request data packet containing the patient identifier, care level and call type based on the trigger signals. The communication link redundancy switching module is used to obtain the real-time communication quality parameters of the IP network. When the real-time communication quality parameters are detected to be lower than the preset parameter threshold, the PSTN communication link is activated as the main transmission channel. The data routing and target group decision module is used to send the structured call request data packet to the call softswitch through the main transmission channel, and the call softswitch queries the nurse scheduling database to obtain the target nurse terminal group identifier list. The broadcast strategy intelligent generation module is used to parse the nursing level and call type in the structured call request data packet, and generate broadcast strategy instructions for the nurse terminal by combining the target nurse terminal group identifier list; The multi-terminal status synchronization acquisition module is used to broadcast the structured call request data packet to the target nurse terminal group according to the broadcast strategy instruction, and to receive the response status data returned by each nurse terminal, including terminal location coordinates, current task count and nursing skill matching degree. The dynamic priority scheduling module is used to generate a nurse response priority sequence based on the nursing level and response status data through a dynamic weight model. The intelligent audio and video session establishment module initiates an audio and video connection request to the target nurse terminal that is first in the nurse response priority sequence, thereby establishing a two-way communication session between the patient terminal and the target nurse terminal. The session switching module is used to monitor the communication signal strength and the target nurse terminal's movement speed in real time during the duration of the two-way communication session. When the communication signal strength is continuously lower than a preset interruption threshold or the terminal's movement speed continuously exceeds a safety threshold, the two-way communication session is switched to the nurse terminal ranked second in the nurse response priority sequence.

10. A computer-readable storage medium, characterized in that: The computer program is stored that can be loaded by a processor and executed as described in any one of claims 1 to 8.

Images