A method and system for intelligent equipment distribution in first aid kits based on voice recognition

Abstract

This invention discloses a method and system for intelligently dispensing first-aid kits based on voice recognition, belonging to the field of voice recognition technology. The method includes: automatically entering a ready state in response to non-voice-triggered signals; collecting and fusing voice, visual, and item status information for situational judgment; dynamically generating a decision tree for first-aid operations based on the judgment results and outputting voice guidance; forcibly executing biometric or environmental perception verification at key operation nodes and monitoring operation path deviations in real time; adjusting the decision tree path after operator confirmation; and dynamically dispensing first-aid items. This invention solves the technical problems of slow response speed, inaccurate operation guidance, and lack of dynamic adaptability in existing first-aid kit dispensing methods during emergencies, resulting in low dispensing efficiency and safety. It achieves the technical effect of rapidly responding, accurately guiding, and dynamically adjusting through intelligent voice recognition and multimodal information fusion, significantly improving the dispensing efficiency and operational safety of first-aid kits.

Description

Technical Field

[0001] This invention relates to the field of speech recognition technology, and specifically to a method and system for intelligent equipment provision in a first aid kit based on speech recognition. Background Technology

[0002] In modern medical emergency scenarios, rescue operations often require a series of complex and high-risk procedures to be completed within a very short time. However, the preparation of emergency kits and related equipment still relies on manual operation. The rescue process is frequently affected by factors such as operator judgment, environmental changes, and stress, easily leading to delays or decreased accuracy, especially in urgent and high-pressure emergency situations. Although some intelligent devices are used to assist in rescue efforts, most rely on simple voice control or fixed routes, lacking real-time adaptability to the environment and operational flow, and unable to flexibly respond to complex rescue needs. Furthermore, existing systems often lack sufficient safety verification measures, resulting in potential safety hazards associated with the use of high-risk items. Summary of the Invention

[0003] This application provides a method and system for intelligent first aid kit equipment dispensing based on voice recognition, which addresses the technical problems of slow response speed, inaccurate operation guidance, and lack of dynamic adaptability in existing first aid kit equipment dispensing methods during emergencies, resulting in low equipment dispensing efficiency and safety.

[0004] The first aspect of this application provides a method for intelligently equipping a first aid kit based on voice recognition. The method includes: responding to a non-voice environment or operation trigger signal, causing the first aid kit to automatically enter a continuous listening and ready state without a wake-up word; in the ready state, simultaneously collecting and fusing voice command information, visual operation information, and item status change information, comprehensively judging the current rescue situation and user intent, and generating a situation determination result; based on the situation determination result, matching or dynamically generating a decision tree for the rescue operation, and actively outputting voice guidance information, while controlling the first aid kit to provide equipment instructions and assist in execution according to the decision tree; during the execution of the decision tree, at preset key operation nodes, forcibly initiating and executing at least one verification process based on biometrics or environmental perception, until verification is passed before subsequent processes can be executed; during the rescue process, real-time monitoring of the deviation between the actual operation flow and the expected path of the decision tree, and adaptively adjusting the subsequent decision tree path after operator confirmation, to dynamically equip the kit.

[0005] The second aspect of this application provides a voice recognition-based intelligent equipment allocation system for a first aid kit. The system includes: a scenario triggering module, used to respond to non-voice environmental or operational trigger signals, causing the first aid kit to automatically enter a continuous listening and ready state without a wake-up word; a scenario determination module, used in the ready state to simultaneously collect and fuse voice command information, visual operation information, and item status change information, comprehensively judging the current rescue scenario and user intent, and generating a scenario determination result; a dynamic decision-making module, used to match or dynamically generate a decision tree for rescue operations based on the scenario determination result, and actively output voice guidance information, while controlling the first aid kit to provide equipment instructions and assist in execution of specific items according to the decision tree; a mandatory compliance verification module, used during the execution of the decision tree to forcibly initiate and execute at least one verification process based on biometrics or environmental perception at preset key operation nodes, until verification is passed before subsequent processes can be executed; and an adaptive adjustment module, used to monitor the deviation between the actual operation flow and the expected path of the decision tree in real time during the rescue process, and adaptively adjust the subsequent decision tree path after operator confirmation, performing dynamic equipment allocation.

[0006] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0007] This application provides a method and system for intelligent equipment allocation in first aid kits based on speech recognition, which relates to the field of speech recognition technology. It uses non-voice trigger signals to put the first aid kit into a continuous listening state, simultaneously collecting voice, visual, and item status information to generate a situational judgment result. It dynamically generates or matches a decision tree for equipment allocation and operational guidance, and enforces a verification process at key nodes. It monitors deviations in real time and adaptively adjusts the decision path, achieving efficient, accurate, and safe equipment allocation. This solves the technical problems of existing first aid kit equipment allocation methods, such as slow response speed, inaccurate operational guidance, and lack of dynamic adaptability in emergency situations, leading to low equipment allocation efficiency and safety. It achieves rapid response, accurate guidance, and dynamic adjustment through intelligent speech recognition and multimodal information fusion, significantly improving the efficiency and safety of first aid kit equipment allocation. Attached Figure Description

[0008] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0009] Figure 1 A schematic flowchart of a method for intelligently equipping a first-aid kit based on voice recognition, provided in an embodiment of this application;

[0010] Figure 2 A schematic diagram of the intelligent equipment allocation system for a first aid kit based on voice recognition, provided in an embodiment of this application.

[0011] Figure labeling: Scenario triggering module 11, Scenario determination module 12, Dynamic decision-making module 13, Mandatory compliance verification module 14, Adaptive adjustment module 15. Detailed Implementation

[0012] This application provides a method and system for intelligent first aid kit equipment dispensing based on voice recognition, which addresses the technical problems of slow response speed, inaccurate operation guidance, and lack of dynamic adaptability in existing first aid kit equipment dispensing methods during emergencies, resulting in low equipment dispensing efficiency and safety.

[0013] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0014] It should be noted that the terms "first," "second," etc., in the specification and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or modules not explicitly listed or inherent to such processes, methods, products, or devices.

[0015] Example 1, as Figure 1 As shown, this application provides a method for intelligently dispensing first aid kits based on voice recognition, the method comprising:

[0016] P10: In response to non-voice environment or operation trigger signals, the emergency kit automatically enters a continuous listening and ready state without the need for a wake word.

[0017] The non-voice environmental or operational trigger signals include at least one of the following: an acceleration signal indicating that the resuscitation box is being moved, an attitude signal indicating that the box door is opening or closing, or a critical life parameter signal received from communication with external medical equipment.

[0018] Specifically, the initial trigger configuration of the resuscitation kit is set to automatically enter a continuous listening and ready state without a wake-up word upon responding to non-voice environmental or operational trigger signals. This mechanism aims to improve the device's response speed during resuscitation, ensuring that the resuscitation kit is quickly ready for subsequent operations in any emergency without relying on manual wake-up commands. To achieve this goal, the resuscitation kit can recognize and respond to various non-voice trigger signals, including at least one of the following: acceleration signals, door opening / closing posture signals, and critical life parameter signals sent by external medical devices.

[0019] Specifically, the acceleration signal originates from accelerometers inside the first-aid kit. These sensors monitor any minute movement of the kit in real time. When the first-aid kit is moved or shaken, such as during emergency treatment when it is carried, moved, or rocked, the sensors detect changes in acceleration. This change usually indicates that the first-aid kit may be in an emergency situation requiring its use. For example, when the first-aid kit is moved, the accelerometers generate corresponding acceleration signals. These signals are filtered and analyzed by a built-in signal processing unit to remove noise and extract valid motion features. According to preset rules, once the detected acceleration value exceeds a preset threshold, such as an acceleration exceeding 0.5g in a certain direction (where g is the acceleration due to gravity), it is determined that the first-aid kit has been moved, thus triggering the first-aid kit to enter a continuous listening and ready state. At this time, the first-aid kit is ready to operate at any time, such as providing first-aid supplies or outputting voice guidance.

[0020] In addition to an accelerometer, the rescue kit is also equipped with an attitude sensor to monitor the opening and closing status of the door. The attitude sensor generates an attitude signal by detecting changes in the door's angle. Based on a combination of a three-axis gyroscope and an accelerometer, this sensor can perceive the door's opening and closing movements in real time. When the door is opened, the attitude sensor detects the door's angle changing from a closed state (e.g., 0 degrees) to an open state (e.g., greater than 90 degrees) and generates a corresponding attitude signal. This signal is transmitted to the rescue kit's control unit via an internal communication line. Upon receiving the signal, the control unit immediately initiates continuous monitoring and a ready state for the rescue kit. This mechanism ensures that the system can respond rapidly the moment the rescue kit is opened, providing immediate support for subsequent rescue operations.

[0021] Furthermore, the resuscitation kit also features communication capabilities with external medical devices to receive critical life-threatening parameter signals from these devices. For example, the kit supports multiple wireless communication protocols, including Bluetooth, Wi-Fi, and Zigbee, via its built-in communication module, ensuring seamless connectivity with different types of medical equipment. The communication module can receive patient vital sign data transmitted from external medical devices such as electrocardiogram monitors and pulse oximeters. When the received patient vital sign data includes critical life-threatening parameter signals, such as a heart rate below 60 beats per minute or blood oxygen saturation below 90%, the resuscitation kit's control unit will immediately trigger continuous monitoring and a ready state. This triggering method based on signals from external medical devices ensures that the resuscitation kit can respond promptly and be ready to provide necessary resuscitation support when a patient's vital signs are in crisis.

[0022] The acquisition and response to all these trigger signals rely on sensor and wireless communication technologies to ensure accurate identification and rapid response to various trigger signals. Accelerometers and attitude sensors are implemented using high-precision inertial measurement units (IMUs), providing high-accuracy motion and attitude data. The communication module, through a built-in microcontroller, performs data processing and protocol conversion, ensuring accurate data transmission and reception.

[0023] When any non-voice-triggered signal is activated, the first aid kit immediately enters a continuous listening and ready state. In this state, upon receiving the trigger signal, the control unit of the first aid kit immediately initiates a preset initialization program. This program includes calibrating the sensor system, testing the connection of the communication module, and activating the voice recognition module. At this time, the system can proactively guide the operator to the next step according to a pre-set procedure, or automatically provide the necessary first aid items based on the emergency scenario, and assist the rescue process through voice prompts and item guidance.

[0024] This multi-signal fusion triggering method not only improves the response speed of the emergency kit, but also enhances its adaptability and reliability in complex medical environments.

[0025] P20: In the ready state, voice command information, visual operation information and item status change information are collected and integrated simultaneously to comprehensively judge the current rescue situation and user intention, and generate a situation judgment result.

[0026] The visual operation information refers to the operator's gestures, postures, or line of sight information obtained through image recognition; the item status change information refers to the status information of items in the box being taken out or put back, sensed by weight sensors or radio frequency identification technology.

[0027] Optionally, after the rescue kit enters a continuous listening and ready state without a wake word, the system will enter a critical perception and decision-making stage. By simultaneously collecting and integrating multiple information sources, including voice command information, visual operation information, and information on changes in the status of the items, the system will comprehensively judge the current rescue situation and the user's intentions, and generate a situation judgment result to ensure that the system can accurately understand and respond to the user's actual needs.

[0028] First, the acquisition of voice command information relies on the voice recognition module built into the first aid kit. When the first aid kit is in a ready state, it continuously listens for voice commands from the operator, including instructing the first aid kit to provide specific items, perform certain first aid operations, or issue voice guidance. Through the analysis and processing of voice signals, the system can accurately identify the operator's needs and use this as an important reference for subsequent decision-making.

[0029] Meanwhile, visual operational information is acquired through image recognition technology. This includes analyzing the operator's hand movements, such as pointing to an object or making a specific gesture, like clenching a fist to request an item, as well as the operator's body posture, such as whether they are standing, sitting, or approaching the first-aid kit. This helps determine the urgency or need of the operator. Eye direction information, such as whether the operator's gaze is focused on an object or device, may indicate that the operator intends to obtain that object or perform a certain operation. This visual information will be combined with voice command information to further enrich the system's decision-making basis.

[0030] Furthermore, information on changes in the status of items is acquired through weight sensors or radio frequency identification (RFID) technology. When an operator removes or returns items from the box, the weight sensors detect the corresponding weight changes. Simultaneously, each item in the rescue box is equipped with an RFID tag, which can be read in real time by the system's RFID readers. When an item is moved, the RFID reader can detect changes in the tag's position, thus determining whether the item has been retrieved or returned. By combining weight sensors and RFID technology, the system can accurately track changes in the status of items within the box.

[0031] Furthermore, by comprehensively judging the current rescue situation and the user's intention, a situation determination result is generated. Step P20 in this embodiment of the application also includes:

[0032] P21: Real-time acquisition of ambient voice stream via microphone array, performance of speech recognition and semantic understanding, extraction of first key instruction information; P22: Acquisition of operation scene images via visual sensor, performance of human posture and gesture recognition, extraction of second operation intent information; P23: Real-time perception of the removal and return status of medicines or instruments in the box via item status sensor, acquisition of item status information; P24: Spatiotemporal alignment and fusion analysis of the first key instruction information, second operation intent information, and item status information to generate the current rescue stage situation determination result.

[0033] It should be understood that the context determination process can be further refined. First, the built-in microphone array captures the ambient speech stream in real time. The microphone array employs a multi-microphone layout, enabling omnidirectional speech signal capture. The captured speech stream undergoes preprocessing operations such as noise reduction, echo cancellation, and gain control to improve signal quality. Subsequently, the preprocessed speech signal is transmitted to the speech recognition module. This module, based on deep learning algorithms, particularly recurrent neural networks (RNNs) or long short-term memory networks (LSTMs), can recognize keywords and instructions in the speech in real time and extract the first key instruction information. For example, when an operator issues instructions such as "prepare CPR equipment" or "get an adrenaline injection," the speech recognition module can accurately identify and extract the core content of these instructions.

[0034] Simultaneously, images of the operational scene are acquired via a visual sensor. This sensor, employing a high-resolution camera, captures the operator's posture, gestures, and gaze direction in real time. The acquired image data is transmitted to an image processing module, which utilizes computer vision technology, particularly convolutional neural networks (CNNs), to analyze and process the images. For example, the image processing module first performs human posture recognition, detecting the operator's body posture and movements using a pre-trained model. For instance, when the operator leans forward and points to an object, the system can recognize this posture and extract the corresponding operational intent information. Furthermore, gesture recognition is performed to identify the operator's hand gestures. For example, if the operator extends their index finger to point at a medicine, the system can recognize this gesture and extract secondary operational intent information. Simultaneously, through gaze tracking technology, the system can also determine the operator's gaze direction, further enhancing the understanding of the user's intent.

[0035] Furthermore, the system uses item status sensors to monitor the real-time removal and return of medicines or instruments within the box. These sensors include weight sensors and radio frequency identification (RFID) tags. Weight sensors are installed at the bottom of each storage compartment in the emergency kit, monitoring changes in item weight in real time. When an operator removes or returns an item, the weight sensor detects the corresponding weight change and transmits this information to the system. Simultaneously, each item in the emergency kit is equipped with an RFID tag, which can be read in real-time by the system's RFID reader. When an item is moved, the RFID reader detects changes in the tag's position, thus determining whether the item has been retrieved or returned. By combining weight sensors and RFID technology, the system can accurately track changes in the status of items within the box and obtain item status information.

[0036] Next, the extracted first key instruction information, second operational intent information, and item status information will be spatiotemporally aligned. Spatiotemporal alignment refers to synchronizing information from different points in time and locations to ensure the timeliness and consistency of the information. For example, when the voice command requests "prepare the defibrillator," the visual operation information shows that the operator is pointing to the location of the defibrillator, and the item status information shows that the defibrillator has been removed, spatiotemporally aligning this information ensures their consistency in time and space.

[0037] Next, the information fusion stage begins. This stage utilizes a comprehensive decision-making algorithm to fuse and analyze the aligned information. This algorithm, based on multi-source information fusion technology, uses preset weights and logical rules to comprehensively analyze different information sources. For example, the weight of voice commands can be set higher, while the weights of visual operation information and item status information can be lower, serving a supporting role. If the voice command requests "prepare the defibrillator," the visual operation information shows the operator is pointing to the defibrillator's location, and the item status information shows the defibrillator has been removed, then the user's intention is determined to be to use the defibrillator. Similarly, if the voice command requests "get an adrenaline," the visual operation information shows the operator is looking at the adrenaline's location, and the item status information shows the adrenaline has been removed, then the user's intention is determined to be to use adrenaline.

[0038] Ultimately, the generated scenario assessment results will serve as the basis for generating the decision tree for subsequent rescue operations. For example, in certain emergency situations, if the operator fails to issue a clear voice command but demonstrates an urgent need for a certain item through gestures or postures, the system will automatically recognize the scenario and infer the operator's needs, thus proactively providing assistance. Furthermore, if the system determines that the operator may have made a mistake or failed to correctly retrieve the item, it can also issue voice prompts or take other measures to assist the operator in completing the operation.

[0039] This information fusion and situation determination method not only improves the intelligence level of the emergency kit, but also enhances the system's adaptability to complex environments and various operating scenarios, enabling it to provide timely and accurate support in emergency situations.

[0040] P30: Based on the situation determination result, match or dynamically generate a decision tree for the rescue operation, actively output voice guidance information, and control the rescue box to provide instructions and assist in the execution of specific items required according to the decision tree.

[0041] Specifically, based on the aforementioned scenario assessment results, the first aid kit will automatically match or dynamically generate a decision tree for the first aid operation, and proactively guide the operator to perform the next step through voice guidance information. Simultaneously, the first aid kit will provide instructions on the specific items needed based on the generated decision tree and assist in the retrieval process to ensure the smooth execution of the first aid operation.

[0042] First, based on the situation assessment results, the corresponding rescue operation decision tree is matched from a pre-set decision tree library. The decision tree library is a core component of the system, containing decision tree models for various common rescue scenarios. These decision tree models are pre-built based on expert experience and medication dispensing records from historical emergency cases, covering a wide range of rescue operation data, from basic life support such as CPR to advanced life support such as defibrillation and medication administration. Each decision tree model records in detail the steps and sequence of the rescue operation, as well as the items and equipment required in each step.

[0043] If the current situation does not perfectly match the preset decision tree, the system will dynamically generate a new decision tree based on the real-time situation. This process is achieved through a built-in decision algorithm that can flexibly adjust the structure and nodes of the decision tree according to the complexity and variability of the rescue situation. For example, in a cardiopulmonary resuscitation scenario, if the situation assessment indicates that the patient needs immediate defibrillation, the system will match or generate a decision tree that includes defibrillation operations, and specify the order and conditions for the allocation of supplies and equipment for each step in the decision tree. Each node of the decision tree corresponds to a specific rescue operation and its supply requirements.

[0044] Once the decision tree is generated, the system will proactively output voice guidance information, indicating to the operator that the equipment preparation plan is complete. The voice guidance information is output through the speaker built into the emergency kit, and the content is clear and concise, ensuring the operator can quickly understand it. The voice guidance module is based on text-to-speech (TTS) technology, which can convert the step descriptions in the decision tree into voice commands in real time. For example, the system may issue a voice prompt saying "Defibrillator is ready," or provide reminders for specific operations, such as "Medications are stored in zone three; please retrieve the medication."

[0045] Meanwhile, the emergency kit also uses a decision tree to instruct on the allocation of items and provides auxiliary execution functions. Specifically, the system controls the indicator lights in the storage compartments containing medicines or instruments to flash in specific colors to guide the operator to quickly locate the required items. The color and flashing pattern of the indicator lights have clear meanings; for example, a flashing green light indicates that the item is needed in the current operation, while a flashing red light may indicate that the item is a spare or requires special attention. The indicator lights are controlled by microcontrollers within the storage compartments, which communicate with the system's main control unit to ensure that the indicator lights flash accurately according to the instructions of the decision tree.

[0046] Furthermore, the system can use miniature directional motors to push desired items to easily accessible locations. These miniature motors, installed inside the storage compartments, push items to positions easily accessible to the operator based on instructions from a decision tree. For example, when an operator needs to use hemostatic gauze, the system controls the motor to push the gauze to the front of the compartment for quick access. The miniature motors are also controlled by microcontrollers within the storage compartments, which precisely control the motor's movement to ensure items are accurately pushed to their designated locations.

[0047] Throughout the process, the system provides operators with efficient and accurate assistance in preparing supplies through the synergy of voice guidance and equipment control. The automatic generation and real-time execution of the decision tree enable the first-aid kit to respond flexibly to different emergency situations, ensuring that the necessary items are provided in the shortest possible time and assisting the operator in completing subsequent operations, greatly improving the success rate of rescues and the ease of operation for the user.

[0048] P40: During the execution of the decision tree, at the preset key operation nodes, at least one verification process based on biometrics or environmental perception is forcibly started and executed until the verification is passed before the subsequent process can be executed.

[0049] Furthermore, step P40 in this embodiment of the application also includes:

[0050] P41: At least one mandatory compliance verification node is preset in the execution path of the decision tree; P42: When the operation process is detected to have reached any mandatory compliance verification node, the automatic advancement of the decision tree is paused and a verification sub-process is started; P43: The verification sub-process requires verification through at least one biometric feature or operational evidence. Only after the verification is passed can subsequent key operations be performed or high-risk items be retrieved; P44: The verification methods include independent voice verification by two people based on voiceprint recognition or environmental security confirmation based on visual analysis.

[0051] Optionally, during the execution of the decision tree by the first aid kit system, to ensure the compliance and safety of the rescue operations, the system will forcibly initiate and execute verification processes at preset key operation nodes. These key operation nodes are pre-set according to the risk level and compliance requirements of the rescue operations, and are usually located before steps involving high-risk operations or the handling of high-risk items, such as before using a defibrillator or administering high-risk drugs. The system will set up mandatory compliance verification nodes to help ensure the safety of these key operations.

[0052] Specifically, firstly, the emergency kit pre-sets at least one mandatory compliance verification node in its decision tree execution path. These nodes are designed to be automatically triggered at critical operational moments, ensuring that each operation meets specific safety standards. For example, when the system identifies an upcoming high-risk operation (such as medication dispensing or emergency equipment use), it automatically enters a mandatory verification state. These nodes typically occur at moments in the emergency procedure when special attention to compliance and safety is required, such as the dispensing of critical medications or the activation of equipment.

[0053] When the system detects that the operational process has reached any mandatory compliance verification node, the automatic progression of the decision tree will be paused. At this point, the emergency kit will initiate a verification sub-process, which includes verification mechanisms based on biometrics or environmental awareness. Pausing the progression of the decision tree ensures that necessary verification is performed at this critical stage, preventing unexpected situations caused by operational errors or safety issues. For example, if the decision tree instructs the retrieval of a certain medication or the activation of a first-aid device, the system will require verification of the operator's identity through biometrics or other verification methods to ensure that the operation complies with safety requirements.

[0054] The purpose of the verification sub-process is to complete verification through at least one biometric feature or operational evidence to ensure the compliance and security of the operation. The system will request the operator to provide corresponding verification information according to the preset verification method. For example, the system can require the operator to perform a two-person independent voice verification based on voiceprint recognition. In this case, the system will prompt two operators to issue voice commands separately, and use the built-in voiceprint recognition module to recognize and compare the voices of the two operators to ensure that the commands come from different people and that the command content is consistent, thereby reducing the risk of operator error or malicious operation.

[0055] Another verification method is environmental safety verification based on visual analysis. By analyzing environmental images captured by cameras or other visual sensors, the system confirms whether the rescue environment meets safety requirements. The system can check for factors in the environment that may endanger the safety of the operator or patient, such as misconfigured equipment or improperly stored medications. If the environment is determined to be unsafe, the system will prevent subsequent operations and provide voice guidance to remind the operator to make adjustments until the environment meets safety standards.

[0056] Only after any of the above verifications is successful will the emergency kit allow subsequent critical operations or access to high-risk items. This means that the system will only proceed to the next steps in the decision tree after verification confirms that the operation is safe and compliant. If verification fails, the system will interrupt the operation, prompting the operator to make appropriate adjustments or initiating other safety measures to ensure the safety of the operation.

[0057] In this way, the emergency kit can ensure that every step in the decision tree execution process meets safety and compliance requirements. Especially when high-risk operations are involved, the design of mandatory compliance verification nodes can effectively avoid erroneous operations and unsafe behaviors, and help improve the safety and reliability of the entire emergency process.

[0058] Furthermore, in the embodiment of this application, step P44 of the dual-person independent voice verification based on voiceprint recognition also includes:

[0059] P44-1: When a preset high-risk drug needs to be retrieved, the system will prompt that a double verification is required; P44-2: The first operator is required to read the verification statement and verify their first voiceprint, generating a first voiceprint verification result; P44-3: Within a preset time window, the second operator is required to read the verification statement and verify their second voiceprint, generating a second voiceprint verification result, and verifying whether the second voiceprint is the same as the first voiceprint; P44-4: Verification is considered successful only if both the first and second voiceprint verification results are successful, and the second voiceprint is different from the first voiceprint.

[0060] In one possible embodiment of this application, the verification process can be further refined during the execution of a two-person independent voice verification based on voiceprint recognition, ensuring the accuracy and safety of the operation when handling high-risk drugs.

[0061] When the system detects that an operation requires the retrieval of a pre-set high-risk medication, such as adrenaline or other emergency drugs, the system will automatically initiate a two-person independent voice verification process. First, a voice prompt will be issued through the built-in speaker: "The current operation involves the retrieval of a high-risk medication. Please conduct a two-person voice verification." This prompt clearly informs the operator that two-person verification is required to avoid unsafe incidents caused by the negligence or incorrect operation of a single operator.

[0062] The system then prompts the first operator to read a pre-set verification statement, such as, "I confirm I need to take adrenaline." After the first operator reads the statement as prompted, the system collects the voice signal through a microphone array and transmits it to the voiceprint recognition module. This module, based on a deep learning algorithm, extracts and compares voiceprint features from the collected voice signal. The system pre-stores voiceprint samples of authorized operators. During the verification process, the first operator's voiceprint is compared with the stored samples to generate a first voiceprint verification result. If the first operator's voiceprint matches the stored sample, the system records the first voiceprint verification result as passed and proceeds to the next step.

[0063] Next, within a preset time window, such as 30 seconds to 1 minute, the system will ask the second operator to read the same verification statement: "I confirm that I need to take adrenaline." After the second operator reads the statement, the system will also collect the voice signal and perform voiceprint recognition to generate a second voiceprint verification result. At the same time, the system will verify whether the voiceprint of the second operator is the same as that of the first operator. This verification step ensures that the two operators independently confirm the drug information and can confirm whether they are different individuals, thus avoiding the potential risk that the two operators are from the same person.

[0064] Finally, the system will comprehensively evaluate the results of the first and second voiceprint verifications. Verification will only be considered successful if both the first and second voiceprint verifications pass, and the second operator's voiceprint differs from the first operator's. Only then will the system allow the high-risk medication retrieval process to continue. If verification fails, the system will interrupt the operation, prompting the operator to recheck or take other safety measures.

[0065] This process of independent voice verification by two people ensures that the retrieval of high-risk drugs meets the requirements and minimizes the risk of operational errors.

[0066] Furthermore, in the environmental security verification based on visual analysis, step P44 of this application embodiment also includes:

[0067] P44-5: After the defibrillator is charged, the visual sensing module analyzes the distribution of people around the bed to determine whether people have retreated to a safe area; P44-6: If it is detected that people have not retreated to a safe area, a voice warning will be continuously issued and the discharge will be prevented; P44-7: The discharge execution will be prevented only when the visual analysis confirms that the positions of all people are in a safe area, or when a voice confirmation instruction is received from an authorized commander.

[0068] Specifically, in addition to independent voice verification by two people based on voiceprint recognition, it also involves environmental safety confirmation based on visual analysis, especially when using a defibrillator to ensure the safety of the environment around the bed. This safety verification process uses a visual sensing module to monitor and analyze the environment in real time to ensure that no one is in a dangerous area, thereby avoiding accidental discharge or other safety risks.

[0069] When the defibrillator completes charging and is ready to discharge, the system automatically activates the visual sensing module to monitor the area around the bed in real time. The visual sensing module captures images of the operating environment using a high-resolution camera and analyzes the distribution of people in the area around the bed using computer vision technology. For example, the system pre-defines a safe zone, such as an area with a radius of 1.5 meters centered on the bed. The visual analysis module detects in real time whether anyone is present within this zone and determines whether they have moved outside the safe zone. This process is achieved through a pre-trained deep learning model that accurately identifies the location of people and updates the analysis results in real time.

[0070] If the visual analysis module detects that any individuals around the bed have not yet moved to a safe area—for example, if they are still within the defibrillator's discharge path—the system will immediately activate a safety warning mechanism, continuously issuing a voice warning through the built-in speaker: "Please ensure all personnel have moved to a safe area; the discharge will be paused." This ensures that personnel are in a safe position during resuscitation. Simultaneously, the system will prevent the defibrillator from discharging, ensuring no dangerous operations are performed if safety conditions are not met. The system will continuously monitor the positions of personnel around the bed until the visual analysis confirms that all personnel are within the safe area.

[0071] In addition, the system provides an alternative safety confirmation mechanism: receiving voice confirmation commands from authorized personnel. When an authorized person, such as a senior physician, confirms that the scene is safe, they can send a confirmation signal to the system via voice command to avoid equipment malfunction due to environmental recognition errors. Specifically, the system collects the commander's voice commands through a built-in microphone array. The microphone array can capture voice signals from all directions, and a preprocessing module performs noise reduction and gain control to ensure the quality of the voice signal. The collected voice signal is transmitted to a voiceprint recognition module, which can identify and compare the commander's voiceprint sample in real time. If the voiceprint verification is successful, the system will further analyze the content of the voice command to confirm whether the command is a clear safety confirmation command such as "The scene is safe, you can discharge." If the voiceprint verification is successful and the content of the voice command confirms safety, the system will release the discharge execution block, allowing the defibrillator to continue the discharge operation.

[0072] Through the above process, the safety of the operating environment can be ensured by using a dual mechanism of visual analysis and voice confirmation when using a defibrillator, a high-risk device.

[0073] P50: During the rescue process, the deviation between the actual operation flow and the expected path of the decision tree is monitored in real time, and the subsequent decision tree path is adaptively adjusted after the operator confirms it, so as to dynamically allocate materials.

[0074] Furthermore, step P50 in this embodiment of the application also includes:

[0075] P51: During the rescue process, monitor the deviation between the actual operation flow and the decision tree in real time; P52: When a voice or operation command that deviates from the standard flow of the decision tree is detected, initiate a second voice confirmation; P53: After obtaining positive confirmation from the second voice confirmation, dynamically adjust and update the current decision tree; P54: Based on the updated decision tree, execute the equipment allocation for subsequent decision paths; P55: Encrypt and timestamp the entire voice flow, operation events, verification results, and decision tree changes, and generate a structured rescue report accordingly.

[0076] It should be understood that during the rescue process, the system will continuously monitor the deviation between the actual operation process and the expected path of the decision tree to ensure that the rescue operation can flexibly adapt to emergencies or temporary changes in the operator's decision-making.

[0077] First, the system monitors the deviation between the actual operation flow and the decision tree in real time using built-in sensors and monitoring modules. Built-in sensors may include a microphone array to capture voice commands, a visual sensor to monitor operational actions, and an item status sensor to track item retrieval. The system compares and analyzes this real-time data with the expected path of the decision tree to identify any operations that deviate from the standard flow. When a deviation occurs in the actual operation flow, such as when the operator performs an action that does not follow the preset path of the decision tree, the system will immediately recognize this. This deviation may manifest as inconsistencies between voice commands and the predetermined flow, or an incorrect order of item retrieval. The system will quickly record the specific location and circumstances of the deviation and initiate the corresponding adjustment process.

[0078] When the system detects voice or operational commands that deviate from the standard decision tree process, it immediately initiates a secondary voice confirmation. For example, it may issue a voice prompt through the built-in speaker: "Please note that the operation deviates from the standard procedure. Please confirm whether the rescue plan needs to be adjusted." Simultaneously, the system will require the operator to confirm via voice command. The operator needs to clearly answer "yes" or "no," and the system uses its voice recognition module to recognize the operator's voice commands in real time and determine their intent. If the operator confirms that the rescue plan needs adjustment, the system will further inquire about the specific adjustment requirements, such as: "Please explain the specific content that needs to be adjusted." The operator can explain the adjustment requirements in detail via voice command, and the system will record these commands for subsequent decision tree updates.

[0079] After receiving positive confirmation via secondary voice verification, the system dynamically adjusts and updates the current decision tree. For example, based on the operator's confirmation and real-time monitored workflow data, the system reassesses the rescue situation and dynamically adjusts the decision tree's path using a built-in decision algorithm. For instance, if the operator confirms the need to adjust the rescue plan, the system can regenerate a decision tree containing different operational steps and equipment based on the new situation assessment. This update process includes reassessing the tasks and equipment in the decision tree, adjusting their order, or optimizing subsequent operations based on changes in the situation. By dynamically updating the decision tree, the system can adapt to different rescue situations in real time and flexibly respond to changes that may occur during operations.

[0080] Based on the updated decision tree, the emergency kit will execute the item requisition for subsequent decision paths. For example, if the updated decision tree indicates that the operator now needs specific medications or medical devices, the system will guide the operator to retrieve the corresponding items via voice or visual instructions based on the new decision path. Simultaneously, the emergency kit may control indicator lights in the item storage compartments to flash, or control a micro-motor to push the required items to an easily accessible location. This dynamic adjustment of item requisition ensures the accuracy and timeliness of item requisition, enabling the operator to smoothly obtain the appropriate items and equipment to perform emergency operations according to the updated decision tree path.

[0081] In addition, the system records all voice commands, actions, verification results, and changes to the decision tree during the rescue process in real time, adding an encrypted timestamp to each record. These records are stored in the system's secure storage module to ensure data integrity and security. After the rescue is completed, the system generates a structured rescue report based on these records. The report details every step of the rescue process, changes to the decision tree, and operator confirmation information, helping operators review the process and providing a basis for future optimization and review.

[0082] Through this process, the emergency kit can flexibly adjust its operation path based on real-time feedback during actual rescue operations, ensuring that the system can respond flexibly to emergencies and improving the efficiency and accuracy of equipment configuration.

[0083] Furthermore, step P50 in this embodiment of the application also includes:

[0084] P56: In high-intensity rescue mode, the gain of the microphone array is automatically increased and the directional beamforming algorithm is enabled to reduce the response priority of non-critical voice interactions and focus on the recognition of rescue-related commands.

[0085] Optionally, the voice interaction function can be further optimized, especially in high-intensity rescue mode, to ensure accurate recognition and rapid response of voice commands.

[0086] When the system detects that it has entered high-intensity rescue mode, the rescue kit improves its sensitivity to critical voice commands by adjusting the gain of the microphone array. The microphone array collects audio signals from the surrounding environment through multiple microphone sensors; adjusting the gain effectively amplifies sounds far from the microphone, especially in high-noise environments that may occur during rescue. For example, at an emergency scene, there may be background noise, multiple people talking simultaneously, or other environmental noise, which can affect the accuracy of the system in recognizing voice commands. By increasing the gain of the microphone array, the system can more clearly capture the operator's instructions, improving voice recognition even in noisy environments.

[0087] Simultaneously, the system employs a directional beamforming algorithm to more precisely focus sound signals from a specific direction. Beamforming technology enhances sound signals from a particular direction and suppresses noise from other directions by controlling the directionality of the microphone array. For example, a first-aid kit might focus on voice signals from a direction by determining the operator's position in front of or near the microphone array. This technology ensures that the system focuses on recognizing the operator's rescue commands while ignoring sound input from irrelevant areas, further reducing the possibility of misidentification.

[0088] Furthermore, the system lowers the response priority of non-critical voice interactions. In high-intensity resuscitation mode, the system focuses on recognizing resuscitation-related commands, prioritizing voice commands directly related to the resuscitation operation. For example, the system prioritizes recognizing and responding to critical commands such as "prepare defibrillator" and "administer adrenaline," while lowering the response priority of non-critical commands such as "request patient history." This mechanism can be implemented through a preset command priority list, categorizing commands according to the urgency and importance of the resuscitation operation. The system monitors the content of voice commands in real time and schedules responses according to the priority list.

[0089] Through this optimization strategy, the first-aid kit can better adapt to tense environments and changing situations under high-intensity rescue conditions. By enhancing its ability to recognize key voice commands, the system ensures that every instruction given by the operator during a rescue is quickly and accurately understood and executed. Simultaneously, the system effectively suppresses interference from non-critical instructions, concentrating resources on the most important emergency tasks and ensuring the efficiency and accuracy of medicine and equipment configuration.

[0090] In summary, the embodiments of this application have at least the following technical effects:

[0091] This application enables the rescue kit to quickly enter a continuous listening and ready state without a wake-up word by using non-voice environment or operation trigger signals, significantly improving response speed and reducing operation latency. By synchronously collecting voice commands, visual operation information, and information on changes in the status of the items, and combining multimodal data to generate context judgment results, it can accurately understand user intentions and improve operational accuracy and efficiency. Based on the context judgment results, it dynamically generates or matches a rescue operation decision tree and enforces the verification process at key nodes, which can further assist in strengthening operational compliance and security. It can monitor operational deviations in real time and adaptively adjust the decision path to flexibly respond to emergencies and reduce operational risks. By recording the operation flow, verification results, and decision tree changes, it can ensure high transparency and traceability of the rescue process.

[0092] The technology achieves the goal of rapidly responding, accurately guiding, and dynamically adjusting by integrating intelligent voice recognition and multimodal information, thereby significantly improving the efficiency of emergency kit distribution and operational safety.

[0093] Example 2, based on the same inventive concept as the voice recognition-based intelligent equipment provision method for first aid kits in the foregoing examples, such as... Figure 2 As shown, this application provides a voice recognition-based intelligent equipment distribution system for first aid kits. The system and method embodiments in this application are based on the same inventive concept. The system includes:

[0094] The context trigger module 11 is used to respond to non-voice environment or operation trigger signals, so that the emergency kit automatically enters a continuous listening and ready state without the need for a wake word.

[0095] The scenario determination module 12 is used to simultaneously collect and integrate voice command information, visual operation information and item status change information in the ready state, comprehensively determine the current rescue scenario and user intent, and generate a scenario determination result.

[0096] The dynamic decision module 13 is used to match or dynamically generate a decision tree for rescue operations based on the situation determination result, and actively output voice guidance information. At the same time, it controls the rescue box to provide instructions and assist in the execution of specific items required according to the decision tree.

[0097] The mandatory compliance verification module 14 is used to forcibly initiate and execute at least one verification process based on biometrics or environmental perception at preset key operation nodes during the execution of the decision tree, and the subsequent process can only be executed after the verification is passed.

[0098] The adaptive adjustment module 15 is used to monitor the deviation between the actual operation flow and the expected path of the decision tree in real time during the rescue process, and to adaptively adjust the subsequent decision tree path after the operator confirms it, so as to dynamically allocate materials.

[0099] Furthermore, in the scenario triggering module 11:

[0100] The non-voice environmental or operational trigger signals include at least one of the following: acceleration signal indicating that the resuscitation box is being displaced, attitude signal indicating that the box door is opening or closing, or critical life parameter signal received from communication with external medical equipment.

[0101] Furthermore, in the scenario determination module 12:

[0102] The visual operation information refers to the operator's gestures, postures, or line of sight information obtained through image recognition; the item status change information refers to the status information of items in the box being taken out or put back, sensed by weight sensors or radio frequency identification technology.

[0103] Furthermore, the scenario determination module 12 is also used to perform the following steps:

[0104] The system uses a microphone array to collect ambient voice streams in real time, performs speech recognition and semantic understanding, and extracts the first key instruction information. It also uses a visual sensor to collect images of the operation scene, performs human posture and gesture recognition, and extracts the second operation intention information. Furthermore, it uses an item status sensor to perceive the removal and return of medicines or instruments in the box in real time and obtain item status information. Finally, it performs spatiotemporal alignment and fusion analysis of the first key instruction information, the second operation intention information, and the item status information to generate a situational judgment result for the current rescue stage.

[0105] Furthermore, the mandatory compliance verification module 14 is also used to perform the following steps:

[0106] In the execution path of the decision tree, at least one mandatory compliance verification node is preset; when the operation process is detected to have reached any mandatory compliance verification node, the automatic advancement of the decision tree is paused and a verification sub-process is started; the verification sub-process requires verification through at least one biometric feature or operational evidence, and subsequent key operations or access to high-risk items are allowed only after the verification is passed; the verification methods include independent voice verification by two people based on voiceprint recognition or environmental security confirmation based on visual analysis.

[0107] Furthermore, the mandatory compliance verification module 14 is also used to perform the following steps:

[0108] When a preset high-risk drug needs to be retrieved, the system will prompt a voice message that a two-person verification is required. The first operator is required to read the verification statement and verify their first voiceprint, generating a first voiceprint verification result. Within a preset time window, the second operator is required to read the verification statement and verify their second voiceprint, generating a second voiceprint verification result, and verifying whether the second voiceprint is the same as the first voiceprint. Verification is considered successful only if both the first and second voiceprint verification results pass, and the second voiceprint is different from the first voiceprint.

[0109] Furthermore, the mandatory compliance verification module 14 is also used to perform the following steps:

[0110] After the defibrillator is charged, the visual sensing module analyzes the distribution of people around the bed to determine whether people have retreated to a safe area. If it is detected that people have not retreated to a safe area, it will continuously issue a voice warning and prevent the discharge from being performed. The discharge execution will be prevented only when the visual analysis confirms that the positions of all people are in a safe area, or when a voice confirmation instruction is received from an authorized commander.

[0111] Furthermore, the adaptive adjustment module 15 is also configured to perform the following steps:

[0112] During the rescue process, the deviation between the actual operation flow and the decision tree is monitored in real time; when a voice or operation command that deviates from the standard flow of the decision tree is detected, a second voice confirmation is initiated; after the second voice confirmation receives positive confirmation, the current decision tree is dynamically adjusted and updated; based on the updated decision tree, the equipment allocation for subsequent decision paths is executed; the entire voice flow, operation events, verification results, and decision tree changes are encrypted and timestamped and stored, and a structured rescue report is generated accordingly.

[0113] Furthermore, the adaptive adjustment module 15 is also configured to perform the following steps:

[0114] In high-intensity rescue mode, the microphone array gain is automatically increased and a directional beamforming algorithm is enabled to reduce the response priority of non-critical voice interactions and focus on the recognition of rescue-related commands.

[0115] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. Additionally, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0116] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

[0117] This specification and accompanying drawings are merely illustrative examples of this application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Therefore, if such modifications and variations fall within the scope of this application and its equivalents, this application intends to include such modifications and variations.

Claims

1. A method for intelligently dispensing first-aid kits based on voice recognition, characterized in that, The method includes: In response to non-voice environment or operation trigger signals, the emergency kit automatically enters a continuous listening and ready state without the need for a wake word; In the ready state, voice command information, visual operation information and object status change information are collected and integrated simultaneously to comprehensively judge the current rescue situation and user intention, and generate a situation judgment result. Based on the situation determination result, a decision tree for the rescue operation is matched or dynamically generated, and voice guidance information is actively output. At the same time, the rescue box is controlled to provide instructions and assistance in equipping specific items according to the decision tree. During the execution of the decision tree, at the preset key operation nodes, at least one verification process based on biometrics or environmental perception is forcibly started and executed until the verification is passed before the subsequent process can be executed. During the rescue operation, the deviation between the actual operation flow and the expected path of the decision tree is monitored in real time, and the subsequent decision tree path is adaptively adjusted after the operator confirms the deviation, so as to dynamically allocate materials.

2. The method for intelligent equipment provisioning of a first-aid kit based on voice recognition as described in claim 1, characterized in that, The non-voice environmental or operational trigger signals include at least one of the following: acceleration signal indicating that the resuscitation box is being displaced, attitude signal indicating that the box door is opening or closing, or critical life parameter signal received from communication with external medical equipment.

3. The method for intelligent equipment provisioning of a first-aid kit based on voice recognition as described in claim 1, characterized in that, The visual operation information refers to the operator's gestures, postures, or line of sight information obtained through image recognition; The item status change information refers to the status information of items in the box being taken out or put back, which is sensed by weight sensors or radio frequency identification technology.

4. The method for intelligent equipment provisioning of a first-aid kit based on voice recognition as described in claim 1, characterized in that, Based on a comprehensive assessment of the current rescue situation and the user's intent, a situation determination result is generated, including: The system uses a microphone array to collect ambient voice streams in real time, performs speech recognition and semantic understanding, and extracts the first key instruction information. The operation scene images are acquired by a visual sensor, human posture and gesture recognition are performed, and the second operation intention information is extracted. The system uses an item status sensor to detect the removal and return of medicines or instruments in the box in real time, and obtains item status information. The first key instruction information, the second operation intention information, and the item status information are spatiotemporally aligned and fused for analysis to generate the current rescue stage situation determination result.

5. The method for intelligent equipment provisioning of a first-aid kit based on voice recognition as described in claim 1, characterized in that, At preset key operation nodes, at least one verification process based on biometrics or environmental awareness is forcibly initiated and executed, including: In the execution path of the decision tree, at least one mandatory compliance verification node is preset; When the operation process is detected to have reached any mandatory compliance verification node, the automatic advancement of the decision tree is paused and a verification sub-process is started. The verification sub-process requires verification through at least one biometric feature or operational evidence. Only after successful verification can subsequent critical operations be performed or high-risk items be retrieved. Verification methods include independent voice verification by two people based on voiceprint recognition or environmental security confirmation based on visual analysis.

6. The method for intelligently dispensing first-aid kits based on voice recognition as described in claim 5, characterized in that, The dual-person independent voice verification based on voiceprint recognition includes: When a preset high-risk drug needs to be retrieved, the system will prompt you to verify the information with two people. The first operator is required to read out the verification statement and verify their first voiceprint, generating the first voiceprint verification result; Within a preset time window, the second operator is required to read out the verification statement to verify the second voiceprint, generate the second voiceprint verification result, and verify whether the second voiceprint is the same as the first voiceprint. The verification is considered successful only if both the first and second voiceprint verification results are successful, and the second voiceprint is different from the first voiceprint.

7. The method for intelligent equipment provisioning of a first-aid kit based on voice recognition as described in claim 5, characterized in that, The visual analysis-based environmental security verification includes: After the defibrillator is charged, the visual sensing module analyzes the distribution of people around the bed to determine whether people have retreated to a safe area. If it is detected that personnel have not retreated to the safe area, a voice warning will be continuously issued and the discharge process will be prevented. The discharge execution block will only be lifted after visual analysis confirms that all personnel are within a safe area, or after receiving a voice confirmation instruction from an authorized commander.

8. The method for intelligent equipment provisioning of a first-aid kit based on voice recognition as described in claim 1, characterized in that, Real-time monitoring of the deviation between the actual operation flow and the expected path of the decision tree, and adaptive adjustment of subsequent decision tree paths after operator confirmation, including: During the rescue process, the deviation between the actual operation flow and the decision tree is monitored in real time; When a voice or operation command that deviates from the standard decision tree process is detected, a secondary voice confirmation is initiated. After receiving positive confirmation via secondary voice confirmation, the current decision tree is dynamically adjusted and updated. Based on the updated decision tree, the allocation of resources is carried out for subsequent decision paths; The entire voice stream, operation events, verification results, and decision tree changes are encrypted, timestamped, and stored, and a structured rescue report is generated based on this.

9. The method for intelligent equipment provisioning of a first-aid kit based on voice recognition as described in claim 1, characterized in that, The method further includes: In high-intensity rescue mode, the microphone array gain is automatically increased and a directional beamforming algorithm is enabled to reduce the response priority of non-critical voice interactions and focus on the recognition of rescue-related commands.

10. A voice recognition-based intelligent equipment dispensing system for first-aid kits, characterized in that, The system includes: The context triggering module is used to respond to non-voice environment or operation trigger signals, so that the emergency kit automatically enters a continuous listening and ready state without the need for a wake word; The scenario determination module is used to simultaneously collect and integrate voice command information, visual operation information and object status change information in the ready state, comprehensively determine the current rescue scenario and user intent, and generate scenario determination results. The dynamic decision module is used to match or dynamically generate a decision tree for rescue operations based on the situation judgment result, and actively output voice guidance information. At the same time, it controls the rescue box to provide instructions and assist in the execution of specific items required according to the decision tree. The mandatory compliance verification module is used to forcibly initiate and execute at least one verification process based on biometrics or environmental perception at preset key operation nodes during the execution of the decision tree, and the subsequent process can only be executed after the verification is passed. The adaptive adjustment module is used to monitor the deviation between the actual operation flow and the expected path of the decision tree in real time during the rescue process, and to adaptively adjust the subsequent decision tree path after the operator confirms it, so as to dynamically allocate materials.

Images