Bed unit disinfection system

Abstract

The invention discloses a bed unit disinfection system, and particularly relates to the technical field of medical disinfection equipment, the bed unit disinfection system comprises a disinfection central control module, a disinfection start trigger module, a disinfection in-place safety sensing module, a disinfection efficiency sensing module, a bed unit disinfection execution module, a bed unit disinfection stop trigger module and a disinfection man-machine interaction module; the disinfection central control module receives a disinfection starting trigger signal and a feedback signal of each sensing module, generates a control instruction, sends a disinfection execution and stop instruction, and performs data interaction with the human-computer interaction module at the same time; according to the method, preventive in-situ clearance disinfection for the hospitalized bed unit is clearly proposed and realized for the first time, disinfection is converted from an end event to a normalized prevention and control measure which runs through the whole hospitalization process, the key pain point that neurological patients are continuously exposed to the bed unit to pollute the environment during hospitalization is solved, and the prevention and control efficiency is improved.

Description

Technical Field

[0001] This invention relates to the field of medical disinfection equipment technology, specifically a bed unit disinfection system. Background Technology

[0002] In neurology wards, patients often have limb movement disorders, impaired consciousness, and weakened immune function. They are often bedridden for long periods and are prone to vomiting, urinary and fecal incontinence, which makes bed units (beds, bedside tables, bed rails, etc.) easy breeding grounds for bacteria and viruses, becoming an important medium for the spread of hospital-acquired infections. Therefore, disinfection of bed units is a key link in preventing hospital-acquired infections in neurology departments.

[0003] Currently, clinical practice mainly relies on two disinfection modes: manual daily disinfection, in which nursing staff use handheld disinfection instruments (such as ultraviolet lamps and disinfectant sprayers) to wipe or irradiate the bed unit; and automated terminal disinfection, which uses a combination of sensors (such as infrared and gravity sensors) to infer at a lower cost that a person has left, and to carry out a comprehensive and thorough disinfection of the bed unit after the patient is discharged, transferred to another department, or dies.

[0004] While existing technologies can disinfect bacteria and viruses, they still have limitations. For example, manual disinfection can be incomplete or insufficient in frequency, potentially affecting sensitive instruments or patients. Neurological patients often have variable conditions and require unpredictable outings for examinations (such as CT scans, MRIs, and blood tests). Automated terminal disinfection methods cannot automatically trigger disinfection based on patients' schedules, making them unsuitable for the intermittent and random disinfection needs of neurological beds. Therefore, existing technologies focus on "terminal" or "post-treatment" procedures, lacking a preventative disinfection method that can safely, automatically, and promptly interrupt the colonization and transmission chain of pathogens on bed surfaces during the continuous hospitalization of neurological patients. In particular, how to safely implement high-intensity physical / chemical disinfection in a dynamic environment where patients may return at any time remains a key unsolved technical challenge. Summary of the Invention

[0005] In order to overcome the above-mentioned defects of the prior art, embodiments of the present invention provide a bed unit disinfection system to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a bed unit disinfection system, comprising: Central control module for disinfection: Receives disinfection start trigger signal and feedback signals from each sensing module, generates control commands and sends disinfection execution and stop commands, and interacts with the human-machine interface module for data exchange. Disinfection start trigger module: Includes three modes: automatic trigger, linkage trigger, and manual trigger. The disinfection central control module will transmit the disinfection start request generated based on at least one trigger condition to the disinfection on-site safety perception module, and obtain the disinfection start trigger result based on the on-site safety judgment result. Disinfection in-situ safety perception module: including vital sign detection unit, dynamic area security unit and in-situ dual verification unit. Based on the disinfection trigger request signal, it continuously performs dual verification of vital sign judgment and dynamic area safety judgment before and during disinfection, and transmits the verification result to the disinfection central control module. Disinfection efficacy sensing module: During the disinfection process, the actual disinfection dose is monitored in real time. Based on the deviation between the actual disinfection dose and the target disinfection dose, the control quantity is calculated, and the adjustment results of the execution equipment through the control quantity are transmitted to the disinfection central control module. Bed unit disinfection execution module: Based on the adjustment results of the execution equipment, it receives control commands from the disinfection central control module and performs physical disinfection on each surface of the bed unit in a coordinated manner through a multi-degree-of-freedom robotic arm and a composite disinfection end effector; Bed unit disinfection stop trigger module: includes active stop and passive stop. Active stop is triggered based on preset disinfection path and disinfection dosage, while passive stop is triggered based on interrupt command and system abnormality. Disinfection human-computer interaction module: used by medical staff to set disinfection plan parameters and upload the data generated during the disinfection process to the medical staff management terminal for human-computer interaction in real time.

[0007] Preferably, the disinfection start triggering module includes a linkage triggering mechanism: the system establishes real-time data linkage with the hospital information system through the communication interface of the human-computer interaction module, synchronously receives patient outing and return schedule event signals, and constructs a linkage triggering mechanism; the disinfection central control module transmits the linkage triggering request signal to the on-site safety perception module according to the linkage triggering mechanism, triggers the on-site safety judgment mechanism, if the on-site safety judgment passes, the disinfection start execution process is triggered, if it fails, the linkage triggering is canceled or the disinfection process is interrupted, the reason for the triggering failure is recorded, and a linkage triggering cancellation prompt is issued on the operation terminal, and the system returns to standby state.

[0008] Preferably, the vital signs detection unit: illuminates the target bed unit using a bio-radar sensor to acquire reflected signals from objects on the bed unit, filters out high-frequency interference signals using a low-pass filter, and then amplifies the reflected signals using a signal amplifier to obtain pre-processed reflected signals amplified to a processable range; extracts characteristic parameters related to vital signs, including chest wall micro-motion signals and heart rate signals, and calculates a comprehensive vital signs judgment value S. v (t); if S v(t)≥Set life existence threshold S v,th This indicates the presence of vital signs. If S is continuously within the set time period Δt... v (t)<S v,th This indicates that there are no vital signs.

[0009] Preferably, the dynamic area security unit includes: A1: After the target bed unit area is illuminated by a 3D vision sensor, the light is reflected back to the sensor, and the sensor measures the time-of-flight t of the light signal. fi The distance to the target point is obtained by combining the speed of light c. Based on the intrinsic parameter matrix K of the camera, the two-dimensional position (u,v) and corresponding distance value z of each pixel in the depth image are converted into three-dimensional spatial coordinates (x,y,z) to generate point cloud data. All three-dimensional spatial points are filtered and aggregated to form a three-dimensional spatial model representing the perimeter of the target hospital bed. A2: Through the disinfection human-computer interaction module, the scope of the disinfection danger zone is set, with the center point of the patient bed as the origin and the danger zone as the radius R. safe Height range (z) min ~z max A cylindrical region, the warning area is of radius r warn The annular region; A3: Set the depth image stream within the corresponding area of ​​the 3D spatial model around the target hospital bed as the detection frame and tracking frame. Select one frame as the detection frame every preset number of frames. Execute the pre-trained target recognition algorithm on the detection frame to identify the personnel target and output its 2D bounding box, which is then mapped to the 3D coordinate set P of the k-th detection frame. per (k), calculate the shortest distance d from the three-dimensional coordinates of the personnel to the boundary of the danger zone. per (k) Shortest distance d per (k), according to d per (k) Obtain the dynamic area security judgment value S based on the preset dynamic area security judgment rules. a (t); Let there be a tracking frame k+I between detection frame k and the next detection frame k+M, where I = 1, 2, ..., M-1, and M is the number of tracking frames between two consecutive detection frames. Call the lightweight target tracking algorithm to update the bounding box output by the previous detection frame, mapping the updated bounding box to the three-dimensional coordinate set P of the (k+I)th tracking frame. per (k+I), calculate the shortest distance d per (k+I), based on preset dynamic area security judgment rules, the dynamic area security judgment value is updated in real time to S. a1 (t); When the tracking algorithm determines that the target is abnormal, it immediately triggers the target recognition algorithm to re-detect, re-acquire the accurate two-dimensional bounding box of the person, and correct the three-dimensional coordinate set P.per corr (k+I), calibrate the shortest distance d per corr (k+I), the corrected dynamic region security judgment value is S a2 (t) If the two-dimensional bounding box of a valid person is not obtained after a set number of consecutive re-detections, it is determined that there are no people in the area, and the normal detection frame logic is restored.

[0010] Preferably, the in-situ dual verification unit combines the disinfection central control module with the comprehensive judgment value S of vital signs. v (t) and dynamic area security judgment value S a (t) Perform in-situ double verification and construct the double verification judgment value S at time t. per (t), when S per When (t)=1, it indicates that the in-situ safety judgment has passed, and the verification result is transmitted to the disinfection central control module. per When (t)=0, it indicates that the verification failed; at the same time, after disinfection is started, when S a When (t) ≤ weak alarm indicator r2, a "personnel entering the warning area" prompt will be issued on the operating terminal.

[0011] Preferably, the disinfection efficacy sensing module includes: during the disinfection process, combining the intensity value collected in real time by the disinfectant factor intensity sensor with the ambient temperature and humidity, and performing integral calculations to obtain the actual disinfection dose D. act (t), D act (t) and target disinfection dose D tar The comparison is performed, and the comparison result e(t) is used as the deviation signal for PID control. The control quantity u(t) is calculated by the PID control algorithm, and the moving speed v(t) of the mechanical arm of the execution device and the power P(t) of the disinfection unit are adjusted. The adjustment result is then transmitted to the disinfection central control module.

[0012] Preferably, the bed unit disinfection execution module includes: B1: Based on the adjustment results of the execution device, including the target moving speed v(t) of the execution device's robotic arm and the target disinfection unit power P(t); decomposing the moving speed v(t) into angular velocity commands θ1(t) of each joint of the multi-degree-of-freedom robotic arm through the inverse Jacobian matrix solution; driving the robotic arm to move along a preset disinfection path based on the acceleration commands of each joint; the preset disinfection path includes a bed surface disinfection path, a bed rail disinfection path, and a bedside table disinfection path; simultaneously converting the disinfection unit power P(t) into a drive signal for the disinfection unit of the composite disinfection end effector, controlling the output power P of the disinfection factor. out (t).

[0013] Preferably, the bed unit disinfection execution module further includes: B2: During the disinfection process, the actual moving speed v is monitored. act (t) and actual disinfection unit power P act The absolute errors |Δv| and |ΔP| of the comparison results are then compared in real time with the target v(t) and P(t), and the corresponding maximum permissible errors Δv are then compared with the target v(t) and P(t). max and ΔP max The comparison is used for feedback verification. The execution module will feed back any error exceeding the limit to the disinfection efficiency sensing module. The disinfection efficiency sensing module will recalculate the control quantity based on the error data and issue new target speed and target power commands until both errors meet the requirements.

[0014] The technical effects and advantages of this invention are as follows: 1. This invention is the first to clearly propose and realize preventive in-situ intermittent disinfection for inpatient bed units, transforming disinfection from a terminal event into a routine prevention and control measure throughout the entire hospitalization process. It solves the key pain point of neurology patients being continuously exposed to the contaminated environment of the bed unit during hospitalization, improves prevention and control efficiency, and forms a complete electronic disinfection log. 2. This invention creatively integrates vital sign detection and dynamic area security dual security perception, and establishes a mandatory security interlock with "AND" logic, ensuring that disinfection can only occur within an absolutely safe time-space window, thus solving the core security obstacle that existing terminal disinfection technologies cannot be applied to in-situ scenarios. 3. This invention achieves precise and verifiable disinfection effects. By introducing a disinfection efficacy sensing module and a closed-loop control algorithm, the system achieves a leap from timed disinfection to quantitative disinfection. Based on real-time feedback, it ensures that the target biological killing dose is reached, making the disinfection effect measurable and traceable, and improving the scientificity and reliability of preventive disinfection. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of the overall process of the present invention.

[0016] Figure 2 This is a flowchart illustrating the disinfection activation trigger module of the present invention.

[0017] Figure 3 This is a flowchart illustrating the disinfection in-situ safety sensing module of the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Please see Figure 1 As shown, the present invention provides a bed unit disinfection system, including a disinfection central control module, a disinfection start trigger module, a disinfection in-situ safety sensing module, a disinfection efficiency sensing module, a bed unit disinfection execution module, a bed unit disinfection stop trigger module, and a disinfection human-machine interaction module; The disinfection central control module is connected to all other modules. The disinfection start trigger module is connected to the disinfection in-situ safety sensing module and the disinfection efficiency sensing module. The bed unit disinfection stop trigger module is connected to the bed unit disinfection execution module and the disinfection efficiency sensing module. The disinfection human-machine interaction module is connected to all modules.

[0020] Central control module for disinfection: Receives disinfection start trigger signal and feedback signals from each sensing module, generates control commands and sends disinfection execution and stop commands, and interacts with the human-machine interface module for data exchange. Please see Figure 2 As shown, the disinfection start triggering module includes three modes: automatic triggering, linkage triggering, and manual triggering. The disinfection central control module transmits the disinfection start request generated based on at least one triggering condition to the disinfection on-site safety perception module, and obtains the disinfection start triggering result based on the on-site safety judgment result, including: The priority of the disinfection activation triggering method is: manual triggering > linkage triggering > automatic triggering; The automatic triggering is based on a preset disinfection plan: First, nursing staff set preventative disinfection plan parameters through an operating terminal (local touchscreen / nurse station system), including disinfection time periods (e.g., daily 8:00-12:30, 14:00-18:00, weekly, monthly, etc.), single disinfection duration (e.g., 10 minutes), disinfection mode (UVC disinfection, plasma disinfection, combined disinfection, etc.), and target disinfection dose D. tarThe system sets the in-situ safety judgment delay duration t and the cumulative delay count n. After setting, the data is stored in the memory of the disinfection central control module to obtain the preset disinfection plan. The disinfection central control module has a built-in clock module that synchronizes the current time in real time. According to the automatic triggering mechanism, it transmits the automatic trigger request signal to the in-situ safety perception module to trigger the in-situ safety judgment mechanism. If the in-situ safety judgment is successful, the disinfection start execution process is triggered. If it fails, the automatic trigger is canceled, the reason for the trigger failure is recorded (such as "the bed unit has continuous vital signs" or "personnel have stayed in the disinfection area"), and an automatic trigger cancellation prompt is issued on the operation terminal. The system then returns to standby mode. The automatic triggering mechanism automatically generates a disinfection trigger request signal if the current time falls within a preset disinfection period. The signal format is Sig. auto (t)=1, 1 indicates triggering, 0 indicates no triggering. The disinfection central control module will automatically trigger the request signal to be transmitted to the in-situ safety sensing module. The automatic triggering of the in-situ security judgment mechanism: if the in-situ security judgment passes, i.e., S per If (t)=1, the disinfection process is triggered; otherwise, S per If (t)=0, the on-site safety judgment will be performed again after a delay of t minutes (e.g., 5 minutes). If the judgment fails after a cumulative delay of n times (e.g., 3 times), the automatic trigger interruption command during the disinfection process will be cancelled. The automatic trigger mode allows nursing staff to send pause / cancel commands through the operation terminal. The disinfection central control module will immediately terminate the automatic trigger process and the system will return to standby mode. The linkage trigger automatically initiates disinfection during patient absences such as examinations or meals. The system establishes real-time data linkage with the hospital information system through the communication interface (Ethernet / Wi-Fi) of the human-computer interaction module, synchronously receiving patient absence and return event signals. The signal types include four standardized signals: "patient absence for examination," "patient return from examination," "patient absence for meal," and "patient return from meal," thus constructing a linkage trigger mechanism. According to the linkage trigger mechanism, the disinfection central control module transmits the linkage trigger request signal to the on-site safety perception module, triggering the on-site safety judgment mechanism. If the on-site safety judgment passes, the disinfection start execution process is triggered. If it fails, the linkage trigger is canceled or the disinfection process is interrupted, the reason for the trigger failure is recorded (such as abnormal linkage trigger mechanism or abnormal on-site safety judgment mechanism), and a linkage trigger cancellation prompt is issued on the operation terminal, and the system returns to standby state. The linkage triggering mechanism is as follows: When the disinfection central control module receives a "patient departure event signal sent by the HIS system" (using the HL7 FHIR R4 standard to transmit patient departure events; after communication disconnection, TCP reconnection is triggered; if the retry fails after a set number of times (e.g., 3 times), the system switches to a Wi-Fi backup link, such as "patient departure for CT," "patient departure for MRI," or "patient leaving after eating"), the departure duration timer t is started. out (t) and t out (t) ≥ preset minimum disinfection time T (set according to the average historical disinfection time, such as 5 minutes), generate a linkage trigger request signal, the signal format is Sig link (t)=1, where 1 indicates triggering and 0 indicates no triggering. The disinfection central control module will transmit a linkage trigger request signal to the in-situ safety perception module, triggering the in-situ safety judgment mechanism; if t out If (t) < T, then the disinfection process will be paused and the patient's outing status will be continuously monitored until the outing time reaches the target or a "patient returns" event signal is received. At the same time, a silent period at night (such as 22:00-6:00 the next day) can be set through the human-computer interaction module. If the patient's outing time reaches the target and the on-site safety judgment is passed during this period, the system will only record the disinfection requirement, and the nursing staff will manually confirm whether to start it immediately. The linked triggering in-situ security judgment mechanism: if the in-situ security judgment passes, i.e., S per If (t)=1, the disinfection start execution process is triggered. During disinfection, if a HIS patient return signal is received, the system immediately executes the highest priority safety interruption, shuts down the disinfection factor, and retracts the robotic arm to the safe position to ensure the patient's safe return to bed. If the patient does not return after disinfection is completed, the system automatically records the disinfection data (disinfection sensing data and execution data generated during the linkage trigger disinfection process), returns to standby state, and waits for the patient to return without any additional operation. If the HIS system event signal is abnormal, the system automatically disables linkage triggering and switches to responding only to manual triggering to avoid erroneous actions during unplanned periods. This embodiment needs to specifically explain that the system should communicate with the hospital information system in real time to obtain the patient's "out / return" event, but the communication still fails after being reset, no messages are received, or the received messages are in the wrong format, invalid, or repeated and out of order, so that the system cannot determine whether the patient is in bed, out, or has returned. This state is collectively referred to as HIS system event signal abnormality.

[0021] The manual triggering mechanism is initiated by nursing staff based on pre-set emergency disinfection needs (such as immediate disinfection of the bed unit after a patient vomits or becomes incontinent). The central disinfection control module transmits the manual trigger request signal to the on-site safety sensing module. If the on-site safety assessment passes, the disinfection process is initiated, and disinfection plan parameters (disinfection mode, target disinfection dose, etc.) can be set by the nursing staff in real time during triggering. If the assessment fails, the reason for rejection is immediately displayed on the operating terminal, prompting the nursing staff to investigate potential safety hazards and re-trigger. After manual triggering of disinfection, nursing staff can pause or stop the disinfection operation in real time via the operating terminal, flexibly controlling the disinfection process to adapt to the uncertainty of emergency disinfection scenarios. The manual triggering mechanism supports two operation methods: First, nursing staff can send a manual trigger command by clicking the "Disinfection Start" button on a local touchscreen, entering an authorization password (to prevent unauthorized operation); second, nursing staff can remotely send a manual trigger command through the nurse station system. The command is transmitted to the disinfection central control module via a communication interface. Upon receiving the manual trigger command, the disinfection central control module generates a manual trigger request signal in the format Sig. man (t)=1. Simultaneously, if there are currently unexecuted trigger processes, the process is directly interrupted and a manual trigger is executed. If the process is currently in the disinfection phase, upon receiving a manual trigger command, it first determines whether the completed disinfection dose has reached the target disinfection dose D. tar The preset ratio (which can be referenced from clinical data, such as 80%) is set. If the ratio is reached, the process is immediately interrupted and the disinfection is recorded as complete. If the ratio is not reached, the operation terminal will display a message: "Disinfection dosage not met, do you want to force an interruption?" The nursing staff must confirm before the process can be resumed. Please see Figure 3 As shown, the disinfection in-situ safety perception module includes a vital signs detection unit, a dynamic area security unit, and an in-situ dual verification unit. Based on the disinfection trigger request signal, it continuously performs dual verification of vital signs judgment and dynamic area safety judgment before and during disinfection, and transmits the verification result to the disinfection central control module. The vital signs detection unit uses a bio-radar sensor (preferably a 24GHz millimeter-wave radar sensor, installed on the ceiling above the bed at a height of 2.0-2.5m, with preset detection angles of horizontal (e.g., 60°) and vertical (e.g., 40°)) to illuminate the target bed unit, acquiring reflected signals from objects (living or non-living) on ​​the bed unit. A low-pass filter (cutoff frequency 100Hz) filters out high-frequency interference signals (such as electromagnetic interference and light interference in the ward), and a signal amplifier amplifies the reflected signals to a manageable range (the manageable range refers to the range built into the disinfection central control module). The digital-to-digital conversion module supports standard input voltage ranges, such as amplifying the reflected signal from approximately 0.5mV to 5mV to common ranges like 0 to 3.3V or 0 to 5V. It then extracts characteristic parameters related to vital signs, including chest wall micromotion signals (frequency range 0.1-0.3Hz, corresponding to human respiratory rate) and heart rate signals (frequency range 0.5-3Hz, corresponding to human heart rate range). The chest wall micromotion signals are generated by the chest rise and fall caused by human respiratory movements, while the heart rate signals are generated by the slight vibrations of the chest caused by the beating of the heart. These two signals have significantly different characteristics and can be distinguished using Fourier transform. Finally, it calculates the comprehensive vital sign assessment value S. v (t), f h (t), f h,min and f h,max These represent the heart rate signal frequency at time t (in beats per minute), the minimum threshold for the heart rate signal (usually 30 bpm; values ​​below this are considered as no effective heart rate), and the maximum threshold for the heart rate signal (usually 180 bpm; values ​​above this are considered as interference signals). A b (t), A b,min and A b,max These represent the amplitude of the thoracic micromotion signal at time t (in millimeters), the minimum threshold for the thoracic micromotion signal amplitude (usually 0.1 mm; values ​​below this are considered as no effective thoracic micromotion), and the maximum threshold for the thoracic micromotion signal amplitude (e.g., 5 mm based on clinical data; amplitudes exceeding this threshold are considered as environmental interference). a1 and a2 are the corresponding weights, a1 + a2 = 1, a1 > a2, for example, a1 = 0.8 and a2 = 0.2; when f h (t)<f h,min or f h (t)>f h,max The heart rate signal term takes a value of 0 when A b (t)<A b,min Or A b (t)>A b,max The value of the chest wall micromotion signal term is 0; if S v(t)≥Set life existence threshold S v,th (For example, 0.1–0.2, determined through clinical trials and statistical fitting, suitable for the characteristics of weak vital signs in neurological patients), indicating the presence of vital signs. If S is continuously set within a time period Δt (e.g., 30 seconds), v (t)<S v,th This indicates the absence of vital signs. In this embodiment, it should be specifically noted that the target bed unit refers to the independently defined treatment and nursing space area to which the current hospital bed belongs, including: the bed itself (mattress, bed frame, and bedding coverage area), the activity warning zone adjacent to the bed belonging to that bed (bedside sitting / lying area, bedside table, first step area after getting out of bed, and regular activity range beside the bed), and excluding corridors, doorways, toilets, adjacent beds, and other public areas of the ward; the raw signal reflected back from the bed unit by the sensor is very weak and has a very small amplitude, making it impossible to perform algorithmic judgment. After being amplified by a signal amplifier by a factor of 1000, the signal amplitude is increased to a legal voltage / amplitude range that the back-end acquisition and processing circuit can normally recognize, sample, and calculate. This range is called the processable range. In measurement and control and medical electronic circuits, the processable range generally specifically refers to: meeting the requirements of the subsequent analog-to-digital converter (ADC). The voltage amplitude range required by the rated input range of the operational amplifier and digital processing unit (e.g., common ranges of 0-3.3V, 0-5V, or ±2.5V) is specified. The heart rate signal threshold and chest wall signal threshold are based on clinical medical physiological standards as the basic range. The basic range is adaptively modified by taking into account the distribution characteristics of clinical signs such as slow heart rate and shallow breathing in bedridden patients, patients with impaired consciousness, and long-term bedridden patients. Then, by collecting, statistically analyzing, and fitting background noise of empty beds and real signs of multiple neurological patients, the boundary values ​​that can effectively distinguish between "occupied / unoccupied" and between effective physiological signals and environmental interference are determined. At the same time, the sensor hardware needs to match the minimum resolution sensitivity, linear operating range, and distortion-free output range of the micro-motion detection sensor in the in-situ safety sensing module to ensure that the threshold matches the hardware detection capability.

[0022] The dynamic area security unit includes: A1: After the target bed unit area is illuminated by a 3D vision sensor (preferably a TOF depth camera, such as a PMD CamBoard nano), mounted on the side wall of the head of the hospital bed at a height of 1.5-1.8m above the ground, the light is reflected back to the sensor. The sensor measures the time-of-flight t of the light signal. fiThe distance to the target point is obtained by combining the speed of light c. Based on the camera's intrinsic parameter matrix K (including focal length and principal point coordinates), the two-dimensional position (u,v) and corresponding distance value z of each pixel in the depth image are converted into three-dimensional spatial coordinates (x,y,z), generating point cloud data. After filtering all three-dimensional spatial points (such as voxel downsampling, statistical filtering, etc.), they are aggregated to form a three-dimensional spatial model representing the perimeter of the target hospital bed, z=c×t. fi / 2, Pixel coordinate to 3D coordinate transformation: ; This embodiment specifically explains that the three-dimensional spatial model around the target hospital bed is a spatial model composed of point cloud data collected by a TOF camera. It mainly includes: static background elements: three-dimensional point clouds of fixed objects such as the bed body (bed surface, bed rails, bed frame), bedside table, walls, and floor; dynamic target elements: real-time point clouds of dynamic objects such as personnel (patients, medical staff, visitors) and mobile devices (such as IV stands); spatial topology information: the relative position, size, and shape of each element, as well as the spatial boundaries of danger zones and warning zones; the detection range of the bio-radar sensor in the vital signs detection unit completely covers the target bed unit, and the overlap rate between the detection range of the TOF depth camera and the radar detection range is ≥90%, ensuring that there are no blind spots in the detection of the bed unit; at the same time, the installation positions of the two sensors have avoided obstructions in the ward (such as IV stands, curtain rods, etc.).

[0023] A2: Through the disinfection human-computer interaction module, the scope of the disinfection danger zone is set, with the center point of the patient bed as the origin and the danger zone as the radius R. safe (Value range 1.2-1.5m), height range z min ~z max A cylindrical area (0.5-2.0m high, covering the disinfection area of ​​the hospital bed and bedside table) is designated as the warning area, with a radius of r. warn A ring-shaped area (e.g., r = 2.0-2.5m) is defined as the hazardous area Ω. safe ={(x,y,z)|x 2 +y 2 ≤R safe 2} and warning area Ω warn ={(x,y,z)|R safe 2 <x 2 +y 2 ≤r warn 2}; This embodiment specifically illustrates that the YOLOv5 lightweight target recognition algorithm used in this invention is a pre-trained model. The training process is based on publicly available personnel detection datasets (such as COCO and VOC) and images of personnel activities collected in clinical ward scenes. The activity images are scaled to a fixed size (such as 640×640) and normalized. Multi-scale features of the images are extracted through a lightweight backbone network (such as CSPDarknet53-tiny). The bounding box coordinates, confidence scores, and class probabilities of personnel targets are predicted on the feature maps. Finally, redundant bounding boxes are removed using non-maximum suppression (NMS), and the final personnel target detection results are output. The target recognition algorithm is considered to have completed pre-training when the following conditions are met: 1. In the ward scene... On the scene test dataset, the average accuracy of personnel target detection (mAP@0.5, meaning that more than 98% of the personnel detection boxes output by the algorithm have an overlap of ≥50% with real personnel targets, and the overall accuracy is stable at ≥98% during the recall of all real targets) is ≥1 (e.g., 98%), the false recognition rate is ≤1 (e.g., 2%), and the single-frame inference time on the target embedded hardware is ≤ the set time (e.g., 33ms); 2. In various ward scenarios such as day / night, different lighting, and partial occlusion, the model's recognition accuracy for various personnel targets is stable and meets the standard; 3. The model has been lightweight compressed (e.g., size ≤5MB) and fine-tuned for ward scenarios, and after being solidified and deployed to the embedded processing unit, it runs without abnormalities for a set duration (e.g., 72 hours).

[0024] A3: Set the depth image stream within the corresponding area of ​​the 3D spatial model around the target hospital bed as detection frames and tracking frames. Select one frame as a detection frame every preset number of frames (e.g., 3 frames). Execute a pre-trained target recognition algorithm on the detection frame to identify the personnel target and output its 2D bounding box (the smallest bounding rectangle used to define the target's location in the target detection algorithm), and map it to the 3D coordinate set P of the k-th detection frame. per (k)={(x i ,y i ,z i )|i=1,2,...,n}, where n is the number of three-dimensional coordinate points, (x i ,y i ,z i Given the i-th 3D coordinate, calculate the shortest distance d from the personnel's 3D coordinate to the boundary of the danger zone. per (k), The danger zone is a radius R safe According to d per (k) Obtain the dynamic area security judgment value S based on the preset dynamic area security judgment rules. a (t); Preset dynamic region security judgment rule: if d per(k)≤0, dynamic area safety judgment value S a (t)=1, which is the danger zone; if 0<d per (k)≤ΔR1, where ΔR1 is the width of the near segment of the warning area (e.g., less than or equal to 1 / 3 to 1 / 2 of the ΔR distance), and ΔR1<ΔR, where ΔR = radius r of the warning area. warn -R safe S a (t) = strong warning indicator value r1 (0.5 < r1 < 1, such as 0.8), if ΔR1 < d per (k)≤ΔR,S a (t) = weak alarm flag value r2 (0 < r2 ≤ 0.5, such as 0.3), if d per (k)>ΔR,S a (t)=0; Let there be a tracking frame k+I between detection frame k and the next detection frame k+M, where I=1,2,...,M-1, and M is the number of tracking frames between two consecutive detection frames. A lightweight target tracking algorithm (such as a lightweight version of SORT / ByteTrack, which is a known existing technology) is called to update the bounding box output by the previous detection frame, mapping the updated bounding box to the three-dimensional coordinate set P of the (k+I)th tracking frame. per (k+I), calculate the shortest distance d per (k+I), based on preset dynamic area security judgment rules, the dynamic area security judgment value is updated in real time to S. a1 (t), if d per (k+I)≤0, dynamic area safety judgment value S a (t)=1, indicating a danger zone; if 0<d per (k+I)≤ΔR1,S a (t)=r1, if ΔR1<d per (k+I)≤ΔR,S a (t)=r2, if d per (k+I)>ΔR,S a (t)=0; When the tracking algorithm determines that the target is abnormal (target abnormality includes: target loss, sudden drift of the tracking box position (tracking box position drift > 5cm), abnormal tracking box scale, tracking confidence lower than a preset threshold (e.g., 0.3), or mismatch between the target features of the current frame and the target features of the previous detection frame), the target recognition algorithm is immediately triggered to re-detect, re-acquire the accurate two-dimensional bounding box of the person, and correct the three-dimensional coordinate set P. per corr (k+I), calibrate the shortest distance d per corr (k+I), the corrected dynamic region security judgment value is S a2(t); If, within a set interval (e.g., preferably 300ms, 200ms to 500ms, determined based on the TOF camera frame rate (30fps), YOLOv5 model inference time (≤33ms), and the dynamic response requirements of the clinical scenario), a set number of consecutive re-detections (e.g., 3 times, which is "triple verification" and meets the high safety requirements of medical devices) still fail to obtain a valid person's two-dimensional bounding box, it is determined that there are no people in the area, and the normal detection frame logic is restored; The in-situ dual verification unit: the disinfection central control module combines the comprehensive judgment value of vital signs S. v (t) and dynamic area security judgment value S a (t) Perform in-situ double verification and construct the double verification judgment value S at time t. per (t), when S per When (t)=1, it indicates that the in-situ safety judgment has passed, and the verification result is transmitted to the disinfection central control module. per When (t)=0, it indicates that the verification failed; at the same time, after disinfection is started, when S a When (t) ≤ weak alarm indicator value r2, a "personnel entering the warning area" prompt is issued on the operating terminal, and the status of the safe area is continuously monitored. S v,th t1 and Δt are the set life existence thresholds S, respectively. v,th Continuous output S v (t)<S v,th The duration and continuous time period set; Disinfection efficacy sensing module: During the disinfection process, it monitors the actual disinfection dose in real time. Based on the deviation between the actual disinfection dose and the target disinfection dose, it calculates the control quantity and transmits the adjustment results of the execution equipment through the control quantity to the central disinfection control module. During the disinfection process, it combines the intensity value collected in real time by the disinfectant intensity sensor with the ambient temperature and humidity, and performs integral calculation to obtain the actual disinfection dose D. act (t), D act (t) and target disinfection dose D tar (For example, the target disinfection dose is compared with the results of a kill experiment: using a UVC-LED array with an output intensity of 30mW / cm², a disinfection dose of 60mJ / cm² corresponds to an action time of 2s (duration = dose / intensity = 60mJ / cm² ÷ 30mW / cm²), the kill rate of MRSA reaches 99.99%, which meets the industry disinfection standard requirements). The comparison result e(t) is used as the deviation signal for PID control; the control quantity u(t) is calculated through the PID control algorithm, and the moving speed v(t) of the mechanical arm of the execution equipment and the power P(t) of the disinfection unit are adjusted. The adjustment result is transmitted to the disinfection central control module, so that D... act(t) approaches the target disinfection dose D tar The approximation determination rule is D. act (t) and D tar The absolute deviation |e(t)| ≤ preset threshold D th (e.g., 5mJ / cm) 2 (This can be adjusted according to clinical disinfection needs), where u(t) ≥ the preset maximum threshold u max The operating parameters are adjusted to the first preset threshold. When u(t) < 0, the operating parameters are adjusted to the second preset threshold. When 0 ≤ u(t) < u max The operating parameters change functionally with u(t) between the first preset threshold and the second preset threshold; Q(τ) represents the actual intensity of the disinfection factor at time τ (e.g., ultraviolet intensity, unit: μW / cm²; plasma density, unit: cm³). -3 Hydrogen peroxide concentration, unit: mg / m³ The unit is seconds, and after integration, μWs / cm² = mJ / cm 2 Plasma density unit: cm -3 ×s; Hydrogen peroxide concentration unit: mg×s / m³. Standardize dosage units, such as normalizing to mJ / cm³. 2 Plasma disinfection: A mapping relationship between plasma density, exposure time, and equivalent UVC dose was established through clinical trials. For example, a plasma exposure of 10¹² cm⁻³·s is equivalent to a UVC dose of 1 mJ / cm². The plasma's q(T(τ),H(τ))dτ was converted to an equivalent mJ / cm². Hydrogen peroxide disinfection: Similarly, a mapping relationship between concentration, time, and equivalent UVC dose was established to achieve dimensional uniformity. The target disinfection dose D for all disinfection modes... tar Units should be standardized, such as using UVC equivalent dose as the unit, to ensure D act (t) and D tar (with consistent dimensions), q(T(τ),H(τ)) is the calculation function for the environmental temperature and humidity compensation coefficients corresponding to the environmental temperature T(τ) and relative humidity H(τ) at time τ, which are adapted to different disinfection modes; The unit is mJ / cm 2 K p K I and K d These are the proportional coefficient, integral coefficient, and differential coefficient, respectively. The proportional coefficient is dimensionless (preferably 0.8), the integral coefficient is in units of 1 / time t (preferably 0.05), and the differential coefficient is in units of time t (e.g., seconds, preferably 0.2, which can be obtained by fitting a least squares regression model based on historical data); v(t) = v0 × (1 - 0.5 × u(t) / u maxv0 is the reference moving speed of the robotic arm (preferably 50 mm / s), u max To control the maximum value (e.g., 2 mJ / cm) 2 In typical bedside disinfection scenarios, the adjustment range of the robotic arm speed should not exceed 50% of the base speed (i.e., a minimum of 0.5v0 and a maximum of 1.5v0). This ensures that the speed change is gradual and the system response is stable within the typical control variable variation range. Simultaneously, it ensures that when the control variable u(t) reaches the threshold u_max, the robotic arm speed just drops to the preset minimum safe speed of 0.5v0. When u(t) ≥ u max When u(t) < 0, adjust the moving speed to the first preset speed threshold (e.g., v(t) = (0.4~0.6)v0, preferably 0.5v0); when u(t) < 0, adjust the moving speed to the second preset speed threshold (e.g., v(t) = (1.4~1.6)v0, preferably 1.5v0); P(t) = P0 × (1 + u(t) / u max P0 is the reference power of the disinfection unit (e.g., the reference power of a UVC-LED array is 30W), when u(t) ≥ u max When u(t) < 0, the power of the disinfection unit is adjusted to the first preset power threshold (e.g., P(t) = (1.4~1.6)P0, preferably 1.5P0); when u(t) < 0, the power of the disinfection unit is adjusted to the second preset power threshold (e.g., P(t) = (0.7~0.9)P0, preferably 0.8P0). This embodiment specifically explains that the disinfectant intensity sensor is used to measure the actual dose of disinfectant applied to the surface being disinfected online. The appropriate sensor type is selected based on the disinfection mode. For example, UVC disinfection uses an ultraviolet intensity sensor, installed on the side of the composite disinfection end effector; plasma disinfection uses a plasma density sensor (preferably a Langmuir probe), installed at the outlet of the plasma jet generator, etc. For different disinfection modes, different compensation coefficient calculation functions can be used for conventional adaptation based on the temperature and humidity response characteristics of the corresponding disinfectant. For example, using U... Taking UVC disinfection as an example, increased temperature enhances the activity of UVC photons, while high humidity reduces disinfection efficiency due to water vapor scattering of ultraviolet rays. Therefore, the compensation coefficient is: q(T,H)=1+0.015×(T-T0) / 1℃−0.008×(H-H0) / 1%RH, where T0 and H0 are the standard environmental conditions for UVC disinfection (e.g., T0=25℃, H0=50%RH). For ozone disinfection: high humidity reduces ozone stability, and high temperature accelerates ozone decomposition. Its compensation coefficient formula emphasizes the negative correction of both humidity and temperature: q(T,H)=1-0.012×(TT). 01 ) / 1℃−0.009×(HH 01 ) / 1%RH−0.0001×(TT 01) / 1℃×(HH 01 ) / 1%RH, T 01 H 01 Standard environmental conditions for ozone disinfection (such as T) 01 =22℃, H 01 =40%RH).

[0025] Bed unit disinfection execution module: Based on the adjustment results of the execution equipment, it receives control commands from the central disinfection control module and performs physical disinfection on each surface of the bed unit in a coordinated manner through a multi-degree-of-freedom robotic arm and a composite disinfection end effector, including: B1: Based on the adjustment results of the execution equipment, including the target moving speed v(t) of the robotic arm and the power P(t) of the target disinfection unit; the moving speed v(t) is decomposed into the joints of the multi-degree-of-freedom robotic arm through the inverse Jacobian matrix (a 6-DOF cascaded robotic arm is used, the range of motion of the robotic arm can cover the target bed unit, 6 joints, its Jacobian matrix is ​​a 6×6 matrix, the Jacobian matrix J is in mm / rad, and it is calibrated by a standard displacement sensor before leaving the factory to ensure J -1 After multiplying by v(t) (mm / s), the unit of θ1(t) is (rad / s). The angular velocity command θ1(t) = J -1 (θ(t))×v(t), J -1 (θ(t)) is the pseudo-inverse of the Jacobian matrix (used for inverse solution calculation), and θ(t) is the actual joint angle vector of each joint of the multi-degree-of-freedom robotic arm at time t. Based on the acceleration commands of each joint, the robotic arm is driven to move along a preset disinfection path, which is first the bed surface disinfection path, then the bed rail disinfection path, and finally the bedside table disinfection path. At the same time, the disinfection unit power P(t) is converted into the drive signal of the disinfection unit of the composite disinfection end effector to control the output power P of the disinfection factor. out (t), P out (t)=k dr ×P(t), k dr The power conversion coefficient of the drive circuit (obtained through offline calibration at the system factory: multiple sets of known power commands are applied to the disinfection unit using a standard power detection device, the corresponding actual output power is collected, and the mapping slope between the command and the actual power is obtained by linear fitting using the least squares method, which is the power conversion coefficient, for example 1.02). In this embodiment, it should be specifically noted that the preset disinfection path can be personalized according to the size of the hospital bed and the position of the bedside table. The length and width of the hospital bed, the distance between the bedside table and the hospital bed, the disinfection mode, and other relevant parameters can be input through the human-computer interaction module. Meanwhile, the robotic arm adopts a lightweight design, with noise levels ≤50dB during movement to avoid disturbing patients' rest in the ward. The disinfection area is divided into three sub-areas: a bed surface disinfection area, a bed rail disinfection area, and a bedside table disinfection area. For each sub-area, existing "serpentine scanning" or "spiral covering" algorithms are used to generate basic path points. These path points are then optimized by combining the robotic arm's kinematic constraints (joint angle range, obstacle avoidance requirements) for smoothing. The optimized path points are then converted into a three-dimensional coordinate sequence at the end of the robotic arm, enabling the robotic arm to move along the preset disinfection path.

[0026] In this embodiment, it should be specifically noted that the movement speed of the robotic arm is controlled by a joint controller, and the power drive is controlled by a power controller of the disinfection unit. The robotic arm is installed on the ceiling track or side wall at the rear of the bed, and its range of motion can cover the entire bed. A composite disinfection end effector is installed at the end of the multi-degree-of-freedom robotic arm and is detachably connected to it. It integrates at least one physical disinfection unit (preferably integrating a deep ultraviolet UVC-LED array (wavelength 265-275nm) and a low-temperature plasma jet generator, which can be activated individually or simultaneously according to disinfection needs. In the composite disinfection mode, the actual disinfection dose D... act (t) is the weighted sum of the individual effective doses of each disinfectant, i.e., D act (t)=w1×D1 act (t)+w2×D2 act (t), D1 act (t), D2 act (t) represents combined disinfection modes 1 and 2, respectively, where w1 and w2 are weighting coefficients (w1+w2=1), set according to clinical killing experiment results (e.g., UVC proportion 0.6, plasma proportion 0.4), and the target disinfection dose D. tar (Synchronously broken down into sub-target doses of each factor), based on the start disinfection command sent by the disinfection central control module, the actuator starts the corresponding disinfection unit according to the preset disinfection mode.

[0027] B2: During the disinfection process, the actual moving speed v act (t) and actual disinfection unit power P act The absolute errors |Δv| and |ΔP| of the comparison results are then compared in real time with the target v(t) and P(t), and the corresponding maximum permissible errors Δv are then compared with the target v(t) and P(t). max and ΔP maxThe comparison and feedback verification process involves the execution module feeding back any error exceeding the limit to the disinfection efficacy sensing module. The disinfection efficacy sensing module recalculates the control parameters based on the error data and issues new target speed and target power commands until both errors meet the requirements, for example, absolute error in movement speed ≤ 1 mm / s, and power error ≤ 0.5 W (Δv). max The accuracy of motion control of the robotic arm is determined by factors such as encoder resolution and joint response speed, typically within the range of 0.5–2 mm / s and ΔP. max The value is determined by the power stability of the disinfection unit (such as power fluctuations and LED light decay), and is usually taken as 3% to 5% of the rated power. Bed unit disinfection stop triggering module: Includes active stop and passive stop. Active stop is triggered based on preset disinfection paths and disinfection dosages, while passive stop is triggered based on interrupt commands and system anomalies; the passive stop has higher priority than the active stop; Active stop determination: When the robotic arm completes all preset disinfection paths, and the disinfection efficacy sensing module verifies the actual disinfection dosage D in all preset disinfection paths. act (t) all approach the target disinfection dose D tar When the module generates a disinfection completion signal, the central control module for disinfection issues a stop command. Passive stop response: When an interrupt command is received, or when an absolute deviation in the disinfection dose is detected to be greater than the preset threshold within a duration of t3 (e.g., 10-15 seconds; the response delay for adjusting the robotic arm speed is typically 1-2 seconds, and the response delay for adjusting the disinfection unit power is typically 0.5-1 second; a 5-second buffer is sufficient for the system to complete the complete closed loop of "deviation detection → parameter adjustment → dose correction," avoiding abnormal stop due to the system still undergoing autonomous correction), and the control quantity u(t) after PID adjustment reaches u... max If the absolute deviation is less than 0 and still cannot be reduced to within the threshold, the module generates an interrupt signal, and the disinfection central control module issues a passive stop command to ensure system safety. This embodiment requires specific explanation regarding passive shutdown, which also includes hardware failures. Hardware failures include, but are not limited to: robotic arm joint drive failures, position sensor malfunctions, damage to disinfection unit LEDs / high-voltage modules, distance sensor failures, power module abnormalities, and communication link interruptions. An alarm signal is issued through the central disinfection control module to facilitate timely troubleshooting and handling by maintenance personnel, and the completed disinfection dose D is recorded at the time of the passive shutdown due to hardware failure. done This allows for the supplementary disinfection plan based on the completed disinfection dosage after the fault is repaired. For example, if D... done ≥0.8D tar If D is deemed to have 'basically met the disinfection standards', it will not be restarted; done <0.8D tar The system automatically generates a disinfection plan, and after the on-site safety assessment is passed, it proceeds according to the remaining dose (D).tar -D done "Initiate disinfection, and the additional disinfection time shall not exceed 50% of the original plan."

[0028] The disinfection human-computer interaction module is used by medical staff to set disinfection plan parameters and upload data generated during the disinfection process to the medical staff management terminal for human-computer interaction in real time. Disinfection plan parameter settings include disinfection period, single disinfection duration, disinfection mode, and target disinfection dose. Data generated during the disinfection process refers to data generated by other modules, including disinfection hazard zone boundaries, various judgment thresholds, disinfection start / interruption commands, and system operating status (standby, disinfection in progress, fault, emergency interruption), alarm information when system faults occur, etc. Simultaneously, data from the equipment executing the disinfection process (such as dose, speed, power, etc.) is collected at a set frequency (e.g., 1 second / time) and uploaded to the medical staff management terminal in real time. The system's operating status, alarm information, and disinfection trigger records are uploaded and permanently stored in real time. The original disinfection data is stored in the central disinfection control module for a set number of days 1, and in the medical and nursing management terminal for a set number of days 2. The set number of days 2 is much longer than the set number of days 1. At the same time, data can be automatically deleted according to the set number of days 2, and nursing staff can export the data as needed. For example, after the central disinfection control module has stored the data for 90 days, it automatically deletes the earliest 30 days of original data (retaining the last 60 days of data). After the medical and nursing management terminal has stored the data for 365 days, it automatically compresses and archives the expired data to the cloud server, retains the last 90 days of data locally, and retains the archived data for 5 years. Nursing staff can apply to restore and query the data as needed. Secondly: The accompanying drawings of the embodiments disclosed in this invention only involve the structures involved in the embodiments disclosed in this invention. Other structures can refer to the general design. In the absence of conflict, the same embodiment and different embodiments of this invention can be combined with each other. In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A bed unit disinfection system, characterized in that: include: Central control module for disinfection: Receives disinfection start trigger signal and feedback signals from each sensing module, generates control commands and sends disinfection execution and stop commands, and interacts with the human-machine interface module for data exchange. Disinfection start trigger module: Includes three modes: automatic trigger, linkage trigger, and manual trigger. The disinfection central control module will transmit the disinfection start request generated based on at least one trigger condition to the disinfection on-site safety perception module, and obtain the disinfection start trigger result based on the on-site safety judgment result. Disinfection in-situ safety perception module: including vital sign detection unit, dynamic area security unit and in-situ dual verification unit. Based on the disinfection trigger request signal, it continuously performs dual verification of vital sign judgment and dynamic area safety judgment before and during disinfection, and transmits the verification result to the disinfection central control module. Disinfection efficacy sensing module: During the disinfection process, the actual disinfection dose is monitored in real time. Based on the deviation between the actual disinfection dose and the target disinfection dose, the control quantity is calculated, and the adjustment results of the execution equipment through the control quantity are transmitted to the disinfection central control module. Bed unit disinfection execution module: Based on the adjustment results of the execution equipment, it receives control commands from the disinfection central control module and performs physical disinfection on each surface of the bed unit in a coordinated manner through a multi-degree-of-freedom robotic arm and a composite disinfection end effector; Bed unit disinfection stop trigger module: includes active stop and passive stop. Active stop is triggered based on preset disinfection path and disinfection dosage, while passive stop is triggered based on interrupt command and system abnormality. Disinfection human-computer interaction module: used by medical staff to set disinfection plan parameters and upload the data generated during the disinfection process to the medical staff management terminal for human-computer interaction in real time.

2. The bed unit disinfection system according to claim 1, characterized in that: The disinfection activation trigger module includes a linkage trigger mechanism: the system establishes real-time data linkage with the hospital information system through the communication interface of the human-computer interaction module, synchronously receives patient outing and return schedule event signals, and constructs a linkage trigger mechanism. According to the linkage triggering mechanism, the disinfection central control module transmits the linkage triggering request signal to the on-site safety perception module, triggering the on-site safety judgment mechanism. If the on-site safety judgment passes, the disinfection start execution process is triggered. If it fails, the linkage trigger is canceled or the instruction is interrupted during the disinfection process. The reason for the trigger failure is recorded, and a linkage trigger cancellation prompt is issued on the operation terminal. The system then returns to standby state.

3. The bed unit disinfection system according to claim 1, characterized in that: The vital signs detection unit: illuminates the target bed unit using a bio-radar sensor, acquires the reflected signals from objects on the bed unit, filters out high-frequency interference signals using a low-pass filter, and then amplifies the reflected signals using a signal amplifier to obtain a pre-processed reflected signal amplified to a processable range; extracts characteristic parameters related to vital signs, including chest wall micro-motion signals and heart rate signals, and calculates the comprehensive vital signs judgment value S. v (t); if S v (t)≥Set life existence threshold S v,th This indicates the presence of vital signs. If S is continuously within the set time period Δt... v (t)<S v,th This indicates that there are no vital signs.

4. The bed unit disinfection system according to claim 1, characterized in that: The dynamic area security unit includes: A1: After the target bed unit area is illuminated by a 3D vision sensor, the light is reflected back to the sensor, and the sensor measures the time of flight t of the light signal. fi The distance to the target point is obtained by combining the speed of light c. Based on the intrinsic parameter matrix K of the camera, the two-dimensional position (u,v) and corresponding distance value z of each pixel in the depth image are converted into three-dimensional spatial coordinates (x,y,z) to generate point cloud data. All three-dimensional spatial points are filtered and aggregated to form a three-dimensional spatial model representing the perimeter of the target hospital bed. A2: Through the disinfection human-computer interaction module, the scope of the disinfection danger zone is set, with the center point of the patient bed as the origin and the danger zone as the radius R. safe Height range (z) min ~z max A cylindrical region, the warning area is of radius r warn The annular region; A3: Set the depth image stream within the corresponding area of ​​the 3D spatial model around the target hospital bed as the detection frame and tracking frame. Select one frame as the detection frame every preset number of frames. Execute the pre-trained target recognition algorithm on the detection frame to identify the personnel target and output its 2D bounding box, which is then mapped to the 3D coordinate set P of the k-th detection frame. per (k), calculate the shortest distance d from the three-dimensional coordinates of the personnel to the boundary of the danger zone. per (k) Shortest distance d per (k), according to d per (k) Obtain the dynamic area security judgment value S based on the preset dynamic area security judgment rules. a (t); Let there be a tracking frame k+I between detection frame k and the next detection frame k+M, where I = 1, 2, ..., M-1, and M is the number of tracking frames between two consecutive detection frames. Call the lightweight target tracking algorithm to update the bounding box output by the previous detection frame, mapping the updated bounding box to the three-dimensional coordinate set P of the (k+I)th tracking frame. per (k+I), calculate the shortest distance d per (k+I), based on preset dynamic area security judgment rules, the dynamic area security judgment value is updated in real time to S. a1 (t); When the tracking algorithm determines that the target is abnormal, it immediately triggers the target recognition algorithm to re-detect, re-acquire the accurate two-dimensional bounding box of the person, and correct the three-dimensional coordinate set P. per corr (k+I), calibrate the shortest distance d per corr (k+I), the dynamic region security judgment value is corrected to S. a2 (t) If the two-dimensional bounding box of a valid person is not obtained after a set number of consecutive re-detections, it is determined that there are no people in the area, and the normal detection frame logic is restored.

5. A bed unit disinfection system according to claim 1, characterized in that: The in-situ dual verification unit: the disinfection central control module combines the comprehensive judgment value of vital signs S. v (t) and dynamic area security judgment value S a (t) Perform in-situ double verification and construct the double verification judgment value S at time t. per (t), when S per When (t)=1, it indicates that the in-situ safety judgment has passed, and the verification result is transmitted to the disinfection central control module. per When (t)=0, it indicates that the verification failed; at the same time, after disinfection is started, when S a When (t) ≤ weak alarm indicator value r2, a "personnel entering the warning area" prompt will be issued on the operation terminal.

6. A bed unit disinfection system according to claim 1, characterized in that: The disinfection efficacy sensing module includes: during the disinfection process, combining the intensity value collected in real time by the disinfectant factor intensity sensor with the ambient temperature and humidity, and performing integral calculations to obtain the actual disinfection dose D. act (t), D act (t) and target disinfection dose D tar The comparison is performed, and the comparison result e(t) is used as the deviation signal for PID control. The control quantity u(t) is calculated by the PID control algorithm, and the moving speed v(t) of the mechanical arm of the execution device and the power P(t) of the disinfection unit are adjusted. The adjustment result is then transmitted to the disinfection central control module.

7. A bed unit disinfection system according to claim 1, characterized in that: The bed unit disinfection execution module includes: B1: Based on the adjustment results of the execution device, including the target moving speed v(t) of the execution device's robotic arm and the target disinfection unit power P(t); decomposing the moving speed v(t) into angular velocity commands θ1(t) of each joint of the multi-degree-of-freedom robotic arm through the inverse Jacobian matrix solution; driving the robotic arm to move along a preset disinfection path based on the acceleration commands of each joint, the preset disinfection path including the bed surface disinfection path, the bed rail disinfection path, and the bedside table disinfection path; simultaneously converting the disinfection unit power P(t) into the drive signal of the disinfection unit of the composite disinfection end effector, controlling the output power P of the disinfection factor. out (t).

8. A bed unit disinfection system according to claim 7, characterized in that: The bed unit disinfection execution module also includes: B2: During the disinfection process, the actual moving speed v act (t) and actual disinfection unit power P act The absolute errors |Δv| and |ΔP| of the comparison results are then compared in real time with the target v(t) and P(t), and the corresponding maximum permissible errors Δv are then compared with the target v(t) and P(t). max and ΔP max The comparison is used for feedback verification. The execution module will feed back any error exceeding the limit to the disinfection efficiency sensing module. The disinfection efficiency sensing module will recalculate the control quantity based on the error data and issue new target speed and target power commands until both errors meet the requirements.

Images