Intelligent analgesia pump recovery robot control method, robot, system and medium

Abstract

The invention discloses an intelligent analgesia pump recovery robot control method, a robot, a system and a medium, and belongs to the technical field of intelligent robotics.The method comprises the steps that an analgesia pump recovery instruction is received through a user side; path planning is carried out according to the recycling ward position, and a robot moving path is obtained; according to the robot moving path, a moving chassis of the analgesia pump recycling robot is controlled to move to reach the recycling ward position, and analgesia pump image information is collected through a camera deployed on the analgesia pump recycling robot; analyzing the analgesia pump image information through an analgesia pump quality verification model to obtain a quality verification signal; when the quality verification signal is passing, the analgesia pump is put into a storage cabin to be fixed and transported to a recovery processing center to be stored. The technical problems that in the prior art, the recovery process of the analgesia pump depends on manual operation, so that efficiency is low, recognition errors exist, and the cross infection risk is high are solved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent robots, and in particular to an intelligent analgesic pump recovery robot control method, a robot, a system and a medium. Background Art

[0002] With the improvement of medical level and the popularization of analgesia technology, analgesia pumps have been widely used in postoperative analgesia management. They play an important role in alleviating patients' pain and improving medical quality. However, a large number of analgesia pumps need to be recycled and managed in a timely manner to ensure the safety and efficiency of medical equipment recycling, which puts higher requirements on the equipment management capabilities of hospitals.

[0003] At present, the recycling of analgesic pumps mainly relies on manual operation, which is not only inefficient, but also prone to human negligence, resulting in incorrect recording of equipment information or cross-infection risks. In addition, traditional recycling methods make it difficult to achieve accurate statistics and information synchronization of the remaining liquid in the analgesic pump, which brings additional burdens to the supervision of medical equipment and drug processing. Therefore, there are technical problems in the prior art that the analgesic pump recycling process relies on manual operation, resulting in low efficiency, identification errors and high risk of cross-infection. Summary of the invention

[0004] The present application provides an intelligent analgesic pump recovery robot control method, robot, system and medium, aiming to solve the technical problems in the prior art that the analgesic pump recovery process relies on manual operation, resulting in low efficiency, recognition errors and high risk of cross-infection.

[0005] In the first aspect, the present application provides a method for controlling an intelligent analgesic pump recovery robot, the method comprising receiving an analgesic pump recovery instruction through a user terminal, wherein the analgesic pump recovery instruction includes a recovery ward location; performing path planning according to the recovery ward location to obtain a robot movement path; controlling a mobile chassis of the analgesic pump recovery robot to move to the recovery ward location according to the robot movement path, and collecting analgesic pump image information through a camera deployed on the analgesic pump recovery robot; analyzing the analgesic pump image information through an analgesic pump quality verification model to obtain a quality verification signal; when the quality verification signal is passed, reminding the nursing staff to put the analgesic pump into a storage cabin, controlling the storage cabin to fix the analgesic pump, and transporting it to a recycling and processing center for storage.

[0006] On the second aspect, the present application provides an intelligent analgesic pump recovery robot, including: a user end, used to interact with the user to receive analgesic pump recovery instructions; a central control unit, used to implement an intelligent analgesic pump recovery robot control method; a mobile chassis, used to control the movement of the robot; a storage cabin, used to fix the analgesic pump; an electronic scale, installed in the storage cabin for collecting the weight of the analgesic pump; a visual device, used to collect the required images of the analgesic pump recovery robot.

[0007] In the third aspect, the present application provides an intelligent analgesic pump recovery robot control system, the system comprising: a recovery instruction receiving module, used to receive analgesic pump recovery instructions through a user terminal, wherein the analgesic pump recovery instruction includes a recovery ward location; a moving path acquisition module, used to perform path planning according to the recovery ward location and obtain the robot movement path; an image information acquisition module, used to control the mobile chassis of the analgesic pump recovery robot to move to the recovery ward location according to the robot movement path, and collect analgesic pump image information through a camera deployed on the analgesic pump recovery robot; a verification signal acquisition module, used to analyze the analgesic pump image information through an analgesic pump quality verification model to obtain a quality verification signal; a recovery and warehousing processing module, used to remind the nursing staff to put the analgesic pump into the storage cabin when the quality verification signal is passed, and then control the storage cabin to fix the analgesic pump and transport it to the recycling processing center for warehousing.

[0008] In a fourth aspect, the present application provides a computer-readable storage medium storing a computer program, and the computer program is used to execute an intelligent analgesic pump recovery robot control method provided in the present application.

[0009] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0010] The adoption of the intelligent analgesic pump recovery robot technology solution, including receiving recovery instructions through the user end, precise positioning of the robot's mobile chassis, fixing the analgesic pump in the storage compartment, weighing with an electronic scale, and disinfecting the disinfection nozzle, solves the technical problems of the existing technology that the analgesic pump recovery process relies on manual operation, resulting in low efficiency, identification errors, and high risk of cross-infection, and achieves the technical effect of improving recovery efficiency and ensuring the accuracy and sanitation and safety of the recovery equipment. Through automated operation and intelligent control, this solution significantly reduces manual intervention, reduces the risk of infection, and improves the level of intelligence in hospital equipment management.

[0011] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 A flow chart of a control method for an intelligent analgesic pump recovery robot is provided for this application.

[0013] Figure 2 A structural schematic diagram of an intelligent analgesic pump recovery robot control system is provided for this application.

[0014] Description of the reference numerals: recovery instruction receiving module 11 , moving path obtaining module 12 , image information acquisition module 13 , verification signal obtaining module 14 , recovery and storage processing module 15 . DETAILED DESCRIPTION

[0015] The embodiment of the present application provides a control method, robot, system and medium for an intelligent analgesic pump recovery robot. The analgesic pump recovery task is automatically performed by the intelligent analgesic pump recovery robot. The robot includes a user end receiving a recovery instruction, a central control unit controlling the recovery process, a mobile chassis for path planning, a storage cabin fixing device, an electronic scale for measuring the weight of the device, and a disinfection nozzle for disinfection.

[0016] After introducing the basic principles of the present application, various non-limiting implementation methods of the present application will be specifically described below in conjunction with the drawings in the specification.

[0017] Embodiment 1, as Figure 1 As shown, the embodiment of the present application provides a method for controlling an intelligent analgesic pump recovery robot, the method comprising:

[0018] Step S100: receiving an analgesic pump recovery instruction through the user terminal, wherein the analgesic pump recovery instruction includes a recovery ward location.

[0019] Specifically, the analgesic pump recovery instruction refers to the task information sent by the user through the system, which includes the notification of the end of the use of the analgesic pump, as well as the specific ward location and equipment information. The recovery ward location refers to the specific location of the analgesic pump.

[0020] In the hospital information management system (HIS), when the analgesic pump is finished, log in to the user terminal to confirm that the device has stopped using and can be recycled. The system generates an analgesic pump recycling instruction based on the recycling application, which contains relevant information about the analgesic pump (such as model, number, usage status) and the recycling ward location. Subsequently, the instruction is transmitted to the robot control system through the network, and the central control unit parses the content and converts the ward location into a specific navigation path.

[0021] By sending the recovery command through the user end, not only the automatic allocation of the analgesic pump recovery task is realized, but also the information omission or error caused by manual recording of the location is avoided, and the accuracy and efficiency of task delivery are improved. At the same time, after integrating navigation technology, the robot can respond quickly, significantly improving the overall efficiency of hospital equipment management.

[0022] Step S200: Perform path planning according to the recovery ward location to obtain the robot movement path.

[0023] Specifically, the recovery ward location refers to the specific geographical location of the ward where the analgesic pump is located, including information such as the floor and ward number. Path planning refers to the process of calculating the optimal path between the current location and the target location (recovery ward) to determine the driving route to efficiently complete the task. The robot's moving path refers to the specific driving route planned, which is usually based on a digital model of the hospital map, including turning points, stair or elevator use and other operational information.

[0024] First, the internal structure of the hospital is converted into a digital map, including key locations such as corridors, stairs, and elevators, and the access properties of the space (such as width, height restrictions, etc.) are marked. The robot uses algorithms (such as A* algorithm and Dijkstra algorithm) to search for paths. The algorithm combines the hospital map and real-time environmental information (such as dynamic obstacles in the corridor) to calculate a safe and efficient path. During the driving process, the robot detects obstacles through sensors such as lidar or cameras, and dynamically adjusts the path in combination with SLAM (simultaneous localization and mapping) technology.

[0025] Through path planning, the robot can accurately and safely reach the designated ward location without human intervention, greatly improving the recovery efficiency. The dynamic planning function ensures that the robot can successfully complete the task even in complex or changing environments. At the same time, multi-point path optimization reduces driving distance, reduces energy consumption, and improves the intelligence level and reliability of equipment use.

[0026] Step S300: Control the mobile chassis of the analgesic pump recovery robot to move to the recovery ward according to the robot's moving path, and collect analgesic pump image information through a camera deployed on the analgesic pump recovery robot.

[0027] Specifically, the camera is a visual perception device deployed on the robot, which is used to collect image information of the analgesic pump. It is usually a high-resolution camera that can capture appearance features, label information, etc.

[0028] After receiving the robot's moving path output by the path planning module, the central control unit of the analgesic pump recovery robot decomposes the path into multiple executable motion instructions and controls the mobile chassis to complete them in sequence. For example, drive 20 meters, turn right 90 degrees, and take the elevator to a certain floor. The mobile chassis achieves precise movement through a wheeled drive mechanism, and its control module combines sensor data (such as lidar and ultrasonic sensors) to perceive the surrounding environment and avoid collisions with obstacles.

[0029] When the robot arrives at the target ward, the camera deployed on the robot starts working. The camera captures images of the analgesic pump in the ward and transmits its appearance features, labels, serial numbers and other information to the robot recognition module as an important input for subsequent quality verification. Specifically, motion control is combined with the PID control algorithm, and the mobile chassis can accurately adjust the speed and direction to ensure smooth movement. Environmental perception uses lidar combined with SLAM (simultaneous localization and mapping) technology to dynamically avoid obstacles and ensure accurate paths. Visual acquisition uses RGB cameras or RGB-D depth cameras to capture images of analgesic pumps. Image processing uses edge detection, feature matching and other algorithms to extract image information and provide input for the recognition module.

[0030] Through precise motion control of the mobile chassis, the robot can efficiently drive in complex hospital environments and accurately reach the target ward. The camera combined with visual acquisition technology can automatically identify and collect image information of the analgesic pump, reducing errors caused by manual operation and improving the efficiency and reliability of task execution. The robot's adaptability in dynamic environments is enhanced, making it more suitable for practical applications in complex scenarios such as hospitals.

[0031] Step S400: Analyze the analgesic pump image information through the analgesic pump quality verification model to obtain a quality verification signal.

[0032] Specifically, the analgesic pump quality verification model is an algorithm model used to verify the image information of the analgesic pump. By comparing the actual collected analgesic pump image with the standard image, it is determined whether the device meets the recycling conditions. Specifically, the construction of the analgesic pump quality verification model is divided into the following steps: First, a large number of standard images and abnormal images of the analgesic pump are collected as training data, and they are preprocessed (such as cropping, normalization, denoising, etc.). Secondly, a convolutional neural network (CNN) is used as the basic model, and a pre-trained model (such as ResNet or VGG) is loaded through transfer learning technology, and the data is subjected to feature extraction and classification training. Then, the images are divided into two categories: normal and abnormal, and the classification accuracy is improved by optimizing the cross entropy loss function. Next, the model performance is evaluated using the validation set, and the robustness of the model is improved by adjusting hyperparameters (such as learning rate, regularization coefficient, etc.). Finally, the model is deployed to the robot system, the collected image information is analyzed in real time, and compared with the standard database to output a verification signal. The analgesic pump image information refers to the appearance image of the analgesic pump collected by the robot camera, including the serial number, model, label and other features of the pump. The quality verification signal refers to the output result of the model after image analysis, which is used to determine whether the analgesic pump meets the quality requirements.

[0033] The analgesic pump quality verification model first calls the standard image in the database and matches the corresponding reference image according to the model, specification and other information of the analgesic pump. Then, the key features of the actual image and the standard image are extracted through image processing technology (such as edge detection and feature point matching), and the similarity calculation algorithm (such as cosine similarity and Euclidean distance) is used for comparison.

[0034] If the similarity is higher than the preset threshold, the model outputs a "passed" quality verification signal; otherwise, it outputs "failed" and records the cause of the abnormality (such as damaged appearance, unclear label). It is usually combined with machine learning technology (such as convolutional neural network CNN) to automatically learn image features and improve the accuracy of verification.

[0035] By analyzing image information through the analgesic pump quality verification model, the automatic quality inspection of the analgesic pump can be completed efficiently and accurately, replacing manual inspection to avoid missed inspections and misjudgments. The high fault tolerance of the model ensures the recognition rate in complex scenarios, and the abnormal recording function provides a basis for subsequent equipment management. This process improves the safety and standardization of equipment recycling and provides strong support for intelligent management of hospitals.

[0036] Step S500: When the quality verification signal is passed, the analgesic pump is placed in a storage compartment for fixation and transported to a recycling processing center for return to storage.

[0037] Specifically, once the robot confirms that the analgesic pump meets the recycling requirements, the robot will start automatic operation and move the analgesic pump to the storage compartment. The storage compartment is usually equipped with equipment such as robotic arms or conveyor belts to accurately place the recycled analgesic pump in the compartment.

[0038] After entering the storage compartment, the robot automatically activates the fixing device to securely fix the analgesic pump in the compartment. The fixing methods include pneumatic clamps, magnetic adsorption, mechanical buckles, etc. These devices ensure that the analgesic pump will not become loose or tilted during transportation, preventing damage to the equipment or affecting subsequent operations.

[0039] After the analgesic pump is fixed, the robot begins to transport the storage compartment to the recycling center along the predetermined path. This process relies on the mobility of the robot chassis and its navigation system. The robot's mobile chassis is usually equipped with an electric drive system and a wheeled structure, which enables it to travel efficiently and avoid obstacles. The navigation system ensures that the robot can accurately reach the recycling center in the complex environment inside the hospital.

[0040] This is no longer a fully automated operation. The recovery process of the analgesic pump does not require human intervention, which reduces the possibility of operational errors and ensures the efficiency and accuracy of the recovery process. The fixing device in the storage compartment ensures that the recovered analgesic pump will not be damaged or accidentally moved during transportation, thereby avoiding the risk of equipment damage and loss. Since the robot can autonomously complete a series of tasks such as recovery, fixing, and transportation, it reduces the workload of manual labor and improves the operational efficiency of the hospital.

[0041] Furthermore, the mobile chassis of the analgesic pump recovery robot is controlled to move to the recovery ward according to the robot moving path, including: obtaining an analgesic pump recovery robot in an idle state; collecting storage cavity image information of the analgesic pump recovery robot; identifying the analgesic pump according to the storage cavity image information to obtain an analgesic pump identification result; when the number of analgesic pumps in the analgesic pump identification result is equal to 0, the mobile chassis of the analgesic pump recovery robot is controlled to move to the recovery ward according to the robot moving path.

[0042] Specifically, the idle state refers to the state in which the analgesic pump recovery robot is on standby or not performing a task. In this state, the robot has not been assigned to perform any task and is in a state where it can be started and is ready to work. Storage cavity image information refers to the image data captured by the camera deployed on the robot. These images usually show the internal situation of the robot's storage cavity and are used to identify whether there is an analgesic pump and the number and status of the analgesic pump. Analgesic pump identification refers to the robot processing the captured storage cavity image through image recognition technology to determine whether there is an analgesic pump in the storage cavity and its number. This process uses machine learning or deep learning algorithms (such as convolutional neural networks) for image analysis and pattern recognition. The analgesic pump identification result is the result obtained after analyzing the storage cavity image information, usually including information such as the number, model, and status of the analgesic pump. If there is no analgesic pump in the storage cavity, the identification result is 0.

[0043] Before starting the recycling task, the robot is in an idle state, that is, it does not perform any task. At this time, the robot will wait for user instructions or the start of the task. Once the recycling task instruction is received, the robot enters the task execution state. The robot's control system and standby mode can be triggered manually or automatically by the system to ensure that the robot is in a standby state.

[0044] In the idle state, the robot uses the camera installed on it to collect images of the storage cavity. The storage cavity is usually one of the physical parts of the robot and is used to store the recovered pain pumps. The robot captures the image of the storage cavity through the camera to confirm whether there is a pain pump in the current cavity.

[0045] Based on the collected storage cavity image information, the robot uses an image recognition algorithm (usually based on a convolutional neural network (CNN) or other deep learning technology) to analyze the objects in the image to determine whether there are pain pumps and their number. The core technology of this process is image recognition and object detection, which ensures that the robot can accurately identify the location and number of pain pumps.

[0046] If image recognition determines that there is no analgesic pump in the storage chamber (i.e., the recognition result shows that the number of analgesic pumps is 0), the robot will control the mobile chassis according to the pre-planned path, start heading to the recovery ward, and prepare to start the recovery mission.

[0047] In this step, through automated image recognition and path planning, the robot can quickly determine the status of the storage cavity, reduce manual inspection and intervention, and improve the processing efficiency of the recycling task. Using image recognition technology, the robot can accurately determine whether there is an analgesic pump in the storage cavity, thereby avoiding misoperation or repeated movement caused by the equipment not being in place, and reducing unnecessary waste of time. The entire process, from recognition to path planning to execution of movement, is highly automated, reducing the risk of human error and ensuring the efficiency and accuracy of equipment recycling.

[0048] Furthermore, analgesic pump identification is performed based on the storage cavity image information to obtain analgesic pump identification result, including: collecting a storage cavity image sample data set, wherein the storage cavity image sample data set has analgesic pump quantity identification data set; using the analgesic pump quantity identification data set as supervision and the storage cavity image sample data set as input, training a convolutional neural network to obtain analgesic pump identification model; processing the storage cavity image information according to the analgesic pump identification model to obtain analgesic pump quantity identification, and setting it as the analgesic pump identification result.

[0049] Specifically, the storage cavity image sample dataset refers to a collection of images of the inside of the storage cavity collected by the robot camera, which is used to train and verify the analgesic pump recognition model. This dataset contains images of the storage cavity in different situations, such as when the storage cavity is empty, when there are one or more analgesic pumps in the storage cavity, etc. The analgesic pump quantity identification dataset is a label dataset corresponding to the storage cavity image sample dataset, which is used to indicate the number of analgesic pumps in each image. Convolutional neural network (CNN) is a deep learning model that is particularly suitable for image processing tasks. CNN extracts features from images through multiple convolutional layers and can effectively identify and classify objects in images. During the training process, CNN automatically learns to extract useful features from images for the identification and counting of analgesic pumps. The analgesic pump recognition model is a model obtained by training a convolutional neural network, which is used to identify analgesic pumps in storage cavity images in practical applications. Based on the features learned during the training process, the model analyzes the input storage cavity image and outputs the number and status of analgesic pumps. The analgesic pump quantity identification refers to the analysis result of the robot on the storage cavity image after being processed by the analgesic pump recognition model. This identification is usually an integer, indicating the number of analgesic pumps detected in the storage cavity. For example, the recognition result is "0" (indicating no pain pump), "1" (indicating one pain pump), and so on.

[0050] The robot continuously collects image data of the storage cavity through the camera. These images include different scenes and conditions, such as the storage cavity is empty, there is one analgesia pump in the storage cavity, multiple analgesia pumps, etc. Each image needs to be manually or automatically labeled with the number of analgesia pumps, that is, the analgesia pump quantity identification dataset.

[0051] The collected storage cavity image sample dataset and the corresponding analgesic pump quantity identification dataset are used as input and trained using a convolutional neural network (CNN). During the training process, CNN automatically learns the image features (such as the shape, size, and location of the analgesic pump) based on the labels in the input image and identification dataset, and gradually optimizes the network weights so that the model can correctly predict the number of analgesic pumps when seeing a new image.

[0052] After the model training is completed, the robot will input the new storage cavity image into the trained analgesic pump recognition model, and the model will output the number of analgesic pumps detected in the image. This number is used as the analgesic pump recognition result to decide whether to continue the recovery task.

[0053] In this step, through the training of convolutional neural networks (CNN), the robot can efficiently and accurately identify the number of analgesic pumps in the storage chamber, so as to make accurate judgments and decisions. The analgesic pump recognition model trained with CNN enables the robot to autonomously identify analgesic pumps without human intervention. It improves the automation level of equipment recovery and reduces the operating burden of staff. By using a large amount of image sample data, the CNN model can learn the features of various situations and can operate stably in different hospital environments. Regardless of how the number of analgesic pumps in the storage chamber changes, the robot can effectively identify and perform the correct recovery operation.

[0054] Furthermore, the analgesic pump quantity identification dataset is used as supervision and the storage cavity image sample dataset is used as input to train a convolutional neural network to obtain an analgesic pump recognition model, including: using the analgesic pump quantity identification dataset as supervision and the storage cavity image sample dataset as input to train a convolutional neural network to obtain a first analgesic pump recognition model; extracting the first analgesic pump quantity identification dataset and the first storage cavity image sample dataset whose training accuracy of the first analgesic pump recognition model is less than or equal to the training accuracy threshold; when the number of the first analgesic pump quantity identification dataset and the first storage cavity image sample dataset is greater than the convergence number, using the first analgesic pump quantity identification dataset as supervision and the first storage cavity image sample dataset as input to train a convolutional neural network to obtain a second analgesic pump recognition model; until the number of the Mth analgesic pump quantity identification dataset and the Mth storage cavity image sample dataset is less than or equal to the convergence number, mean fully connecting the first analgesic pump recognition model, the second analgesic pump recognition model, and the Mth analgesic pump recognition model to generate the analgesic pump recognition model.

[0055] Specifically, training accuracy refers to the proportion of correct predictions calculated by the convolutional neural network during the training process by comparing with the actual label (number of pain pumps). The higher the accuracy, the better the prediction effect of the model. The training accuracy threshold refers to a set threshold value. When the accuracy of the model is lower than this value during the training process, the training data or model architecture needs to be readjusted until the qualified accuracy is reached. The number of convergences refers to the situation in the training process when the number of training sample data is reduced to a certain threshold, indicating that the model learning has stabilized and the accuracy cannot be significantly improved by continuing to increase the sample data. Mean full connection refers to the fusion of multiple trained models (such as the first and second models) in the final stage of the model to obtain a final pain pump recognition model. Weighted averaging or other technical means are usually used to combine the output results of each model to obtain a more stable and accurate model.

[0056] The convolutional neural network (CNN) is trained using the storage cavity image sample dataset as input data and the analgesic pump number identification dataset as supervision labels. At this stage, the CNN begins to learn the relationship between the storage cavity image and the number of analgesic pumps, and optimizes its internal parameters to predict the number of analgesic pumps as accurately as possible.

[0057] If the model accuracy obtained in the first training phase is lower than the preset training accuracy threshold, it is necessary to filter out those data sets with lower accuracy. These data include the first analgesia pump quantity identification data set and the first storage cavity image sample data set. These low-quality data will affect the accuracy of the model and therefore need to be further processed.

[0058] After extracting the low-accuracy data set, the robot will retrain a new CNN model (i.e., the second analgesic pump recognition model). This new model will continue to train using this data until it reaches a convergence number, that is, the amount of data becomes sufficient and the model's learning results tend to be stable.

[0059] When the second analgesic pump recognition model is trained, the robot will combine the first model with the second model and use the mean full connection technology to fuse the output results of the two to obtain a final analgesic pump recognition model. This step is to combine the prediction results of multiple models through weighted average or other methods to ensure that the prediction results of the final model are more accurate and stable.

[0060] Through multi-stage training and iterative optimization, the robot can accurately identify the number and status of analgesic pumps in complex environments and changing image conditions, reducing misidentification or missed identification. By screening low-quality data and retraining, the robot can ensure that the model only uses high-quality data for learning, improving the efficiency and accuracy of the training process. The mean full connection technology ensures that the advantages of multiple models are fully utilized, making the final analgesic pump recognition model more robust and able to adapt to image changes in different environments.

[0061] Furthermore, when the quality verification signal is passed, the nursing staff is reminded to put the analgesic pump into the storage cabin, control the storage cabin to fix the analgesic pump, and transport it to the recycling and processing center for return to the warehouse, which also includes: obtaining the first measured weight of the analgesic pump through the electronic scale of the storage cabin; obtaining the first initial weight and liquid density of the analgesic pump model associated with the analgesic pump; calculating the remaining liquid medicine based on the first initial weight, the first measured weight and the liquid medicine density to obtain the remaining amount of medicine, and summarizing the information of the analgesic pump.

[0062] Specifically, the first measured weight refers to the current actual weight of the analgesic pump measured by an electronic scale, that is, after the robot puts the analgesic pump into the storage compartment, it measures the weight of the analgesic pump by an electronic scale. The first initial weight of the associated analgesic pump model refers to the standard weight value corresponding to the currently recovered analgesic pump model. This standard weight is obtained during the production stage or previous data, and usually includes the total weight of the medicine and the pump body. This initial weight can help determine whether there is any residual medicine in the analgesic pump. The density of the medicine refers to the density (mass per unit volume) of the medicine inside the analgesic pump. The density of the medicine is usually known and can be obtained by the type and specification of the medicine. The density of the medicine is an important parameter when calculating the remaining amount of medicine.

[0063] When the robot puts the analgesic pump into the storage compartment, the weight of the analgesic pump is measured by the electronic scale in the storage compartment. The electronic scale measures the actual weight of the analgesic pump through a precise sensor and displays the measurement result. This weight value is used for subsequent calculations and analysis.

[0064] For each type of analgesic pump, the system will obtain the initial weight associated with the analgesic pump (including the total weight of the pump body and the liquid medicine) according to the preset model and database. At the same time, according to the type of liquid medicine in the analgesic pump (such as anesthetic drugs, analgesics, etc.), the system will also obtain the density of the liquid medicine.

[0065] Use the formula to calculate the remaining amount of liquid in the analgesia pump based on the known initial weight, actual measured weight, and liquid density. Specifically:

[0066]

[0067] After the remaining liquid medicine is calculated, the robot will generate summary information including the analgesic pump model, the first measured weight, the remaining liquid medicine volume and other relevant information. These data will be stored in the database for subsequent management, analysis and warehousing.

[0068] This step can help hospitals better manage the use and inventory of medicines by accurately calculating the remaining amount of liquid medicine, avoiding waste and drug expiration. The record of the remaining amount of liquid medicine can also be used for subsequent liquid medicine processing and recycling. By recording the weight and remaining amount of liquid medicine for each analgesia pump, a transparent tracking record can be provided for the recycling process. Medical staff can check the status of each analgesia pump at any time to ensure that the recycling operation is compliant and meets standards.

[0069] Furthermore, the analgesic pump image information is analyzed through the analgesic pump quality verification model to obtain a quality verification signal, including: retrieving the analgesic pump standard image according to the analgesic pump model associated with the analgesic pump; comparing the analgesic pump standard image and the analgesic pump image information through the analgesic pump quality verification model to obtain image similarity; when the image similarity is greater than or equal to a similarity threshold, setting the quality verification signal to passed; otherwise, setting the quality verification signal to failed.

[0070] Specifically, the standard image of the pain pump is a pre-taken standard image of the pain pump, which is used as a reference image for quality verification. The standard image usually contains the appearance and shape of the pain pump under ideal conditions, and these images are known to meet the quality requirements. Image similarity is a metric used to measure the similarity between two images. Common similarity calculation methods include mean square error (MSE), structural similarity (SSIM), cosine similarity, etc. The higher the similarity, the stronger the consistency of the visual content of the two images. The similarity threshold is a preset value used to determine whether two images are similar enough. If the calculated image similarity is greater than or equal to this threshold, the image is considered to meet the quality requirements, otherwise it is considered not to meet the requirements.

[0071] First, the system retrieves the corresponding standard images from the database according to the model of the analgesic pump. These standard images are pre-shot and stored, representing the ideal appearance of the analgesic pump of this model. The standard images serve as a benchmark for comparison. The analgesic pump recycling robot uses its equipped camera to capture the actual image of the analgesic pump during the recycling process. The camera can be a high-resolution camera or an industrial camera, which captures the appearance image of the analgesic pump.

[0072] Use the pain pump quality verification model (usually a deep learning model based on convolutional neural networks) to compare the standard image and the actual image. The model extracts features in the image, such as edges, color distribution, shape, etc., and calculates the similarity of the images.

[0073] The similarity value output by the model reflects the degree of visual match between the standard image and the actual image. Commonly used similarity algorithms, such as the structural similarity index (SSIM), calculate a similarity between 0 and 1 by comparing the brightness, contrast, structure and other features of the two images. A similarity value close to 1 indicates that the two images are very similar.

[0074] The calculated image similarity is compared with the preset similarity threshold. If the image similarity is greater than or equal to the set threshold, the quality verification signal is set to "pass", which means that the appearance of the analgesic pump meets the standard; if the image similarity is lower than the threshold, the signal is set to "fail", indicating that there is a problem with the appearance of the analgesic pump.

[0075] Through automated image comparison, it is possible to quickly determine whether the analgesic pump meets the quality standards, reduce the time and effort of manual inspection, and improve the efficiency of recycling work. Through a high-precision quality verification model, it is ensured that the recycled analgesic pump has no appearance defects or damage, thereby ensuring the quality of the recycling process. Based on deep learning and image processing technology, the robot can automatically perform image comparison, avoiding subjective errors in manual operation and ensuring the consistency and accuracy of quality inspection.

[0076] Furthermore, the intelligent analgesic pump recovery robot control method also includes: when the return button of the analgesic pump recovery robot is activated, obtaining verification identity information; when the verification identity information is the same as the preset identity information, controlling the storage cabin to release the analgesic pump, reminding the staff to return the analgesic pump to the warehouse; when the verification identity information is different from the preset identity information, generating a verification error instruction to remind re-verification; when the analgesic pump is released without identity verification, generating an abnormal return to warehouse alarm signal, and recording the facial information around the analgesic pump for storage.

[0077] Specifically, the return button is a button on the analgesic pump recovery robot that is used to start the return operation. When the button is activated, it means that the operator confirms that the robot is ready for the final step of the return, that is, placing the recovered analgesic pump into the storage compartment and performing related identity authentication operations. Verification identity information refers to data used to confirm the identity of the operator. Usually this includes passwords, QR codes, RFID cards, fingerprint recognition, or face recognition. Verification identity information is to ensure that only authorized personnel can operate the return process of the analgesic pump. Preset identity information refers to a set of valid identity information stored in the system, which is used to compare with the verification identity information obtained in real time. The preset identity information can be the operator's user name, password, or other identity authentication information. The system will compare it with the input identity information to decide whether to allow the operation to continue. Verification error instruction When the input identity information does not match the preset identity information, the system generates an instruction requiring the user to re-authenticate. The system prompts the staff member that the identity authentication failed through an error instruction to ensure the safety of the operation. Abnormal return alarm signal refers to an alarm signal that the system will immediately issue when the analgesic pump is released and prepared for return without verification. This is to prevent unauthorized operation, increase security, and avoid theft or misuse of the device. Facial information recording means that when an abnormal event occurs, the system will automatically capture and record the facial image of the person near the pain pump through the camera. This image will serve as evidence to help trace the responsibility for abnormal operation later.

[0078] The operator presses the return button and the robot starts the return process. The return button is one of the interactive interfaces between the robot and the staff, which starts an automated process controlled by the system and indicates that the robot is ready to perform the return task.

[0079] After activating the return button, the system first requires the operator to authenticate. Identity verification can be obtained in a variety of ways. Common authentication methods include fingerprint recognition, RFID cards, QR code scanning, or face recognition. Choose the appropriate verification method based on the needs of the hospital or management system. For example, when using face recognition, the system will call the camera to capture the facial image of the staff member and compare it with the database in the preset identity information.

[0080] The system compares the identity information collected by tools such as cameras, fingerprint scanners, or RFID readers with the preset identity information. Assuming the preset identity information is the operator's username and password, the system will extract the identity information from the operator's entered password or scanned RFID card and compare it with the records in the database.

[0081] If the verified identity information matches the preset identity information, the system will send a signal to control the storage compartment to release the analgesic pump. At this time, the system will remind the staff to return the analgesic pump to the warehouse. If the identity verification fails, the system will generate a verification error instruction and remind the staff to re-verify the identity to prevent unauthorized operations.

[0082] If the operator attempts to release the analgesic pump and return it to the warehouse without identity verification, the system will immediately detect this abnormal behavior and generate an abnormal return alarm signal. The system will not only send out an alarm signal, but also activate the face recognition camera to record the facial images of people around the analgesic pump and save relevant evidence for subsequent tracing and investigation.

[0083] Through the identity verification process, only authorized staff can return the analgesic pump to the warehouse, avoiding unauthorized operation, thereby improving the security of the system and preventing the loss or abuse of equipment. The system uses face recognition technology to record the facial images of relevant personnel when abnormal operations occur, providing evidence for subsequent investigations and tracking and supervising operators.

[0084] In summary, the intelligent analgesic pump recovery robot control method provided in the embodiment of the present application has the following technical effects:

[0085] 1. Through intelligent path planning and quality verification, the robot can accurately complete the task of recycling the analgesic pump. This method improves the automation and efficiency of the recycling process, reduces manual intervention, and ensures the accuracy and safety of the recycling task.

[0086] 2. Use convolutional neural networks for image recognition and data training to optimize the accuracy of pain pump recognition. This technology can automatically extract and judge the status of the device, improve the robot's recognition ability and data processing accuracy, and reduce human intervention.

[0087] 3. The security of the recycling process is improved through the identity authentication mechanism and abnormal alarm function. The system can prevent unauthorized operations, ensure the compliance of equipment management, and monitor and record abnormal events in real time, effectively reducing security risks.

[0088] In the second embodiment, the present application provides an intelligent analgesic pump recovery robot including: a user end, used to interact with the user to receive analgesic pump recovery instructions; a central control unit, used to implement an intelligent analgesic pump recovery robot control method; a mobile chassis, used to control the movement of the robot; a storage cabin, used to fix the analgesic pump; an electronic scale, installed in the storage cabin for collecting the weight of the analgesic pump; a visual device, used to collect the required images of the analgesic pump recovery robot.

[0089] Specifically, the user end refers to the interface or device that interacts with the user (such as medical staff), usually a touch screen, computer, tablet or other terminal, through which the user can input instructions, monitor the recovery task and receive feedback from the robot. The central control unit refers to the core computing unit used to control the overall operation of the intelligent analgesic pump recovery robot. It is responsible for processing the instructions received from the user end, coordinating the actions of the various components of the robot, and executing the algorithms and logic in the recovery process. The mobile chassis is the moving part at the bottom of the robot, usually including components such as drive wheels and obstacle avoidance sensors, which are used for the robot to move autonomously in the environment, reach the designated recovery location and perform tasks. The storage cabin is used to store the cabin for recovering the analgesic pump. There are mechanical devices (such as fixtures or clamps) inside it to ensure that the analgesic pump will not move or be damaged during storage. The electronic scale is a weight measuring device installed in the storage cabin, which is used to accurately measure the weight of the analgesic pump and help the robot complete the calculation of the remaining liquid and other information identification. The visual device usually includes a camera or a high-resolution camera for image acquisition. The device is responsible for collecting visual data required during the recycling process, such as the appearance image, location, and status of the analgesic pump, for quality verification, identification, and path planning.

[0090] The user end is the main interactive interface between the intelligent robot and the user, allowing medical staff to issue recovery instructions and monitor the progress of the task in real time. The user enters the recovery task through a graphical interface, such as selecting the recovery ward location, recovery time, etc. This part of the tool is usually a touch screen or tablet, which is convenient for the operator to issue instructions and view feedback intuitively. After receiving the recovery instructions issued by the user end, the central control unit is responsible for converting these instructions into specific robot action plans and implementing related tasks. It handles tasks such as path planning, analgesic pump image acquisition, and quality verification. The central control unit controls the mobile chassis to perform precise movements according to the predetermined algorithm, and coordinates the visual device to collect images to ensure that the recovery process is correct. The mobile chassis is the basis of the intelligent recovery robot, including drive motors, wheels, sensors, etc. The robot plans the path according to the instructions of the central control unit, automatically avoids obstacles and arrives at the recovery ward location. This process relies on embedded motion control algorithms and sensors (such as lidar, ultrasonic sensors, etc.) to ensure the accuracy of the path. The robot collects the required analgesic pump image data through its equipped visual device (such as a high-definition camera). The image data is transmitted to the central control unit for subsequent identification, quality verification, etc. For example, the robot can use image recognition to determine whether the analgesic pump meets the recycling conditions and whether there is any damage or leakage. When the robot arrives at the recycling site and detects the analgesic pump, the storage cabin will activate the mechanical device to secure the analgesic pump in the cabin. The robot will adjust the device in the cabin according to the size and shape of the analgesic pump to ensure that the analgesic pump is stable and immobile to avoid damage during transportation. After the analgesic pump is fixed in the storage cabin, the electronic scale starts working to accurately measure the weight of the analgesic pump. This weight data will help to calculate the remaining amount of medicine in the analgesic pump, or be used to generate equipment information records for subsequent processing or management.

[0091] Embodiment 3, based on the same inventive concept as the control method of an intelligent analgesic pump recovery robot in Embodiment 1, as Figure 2 As shown, the embodiment of the present application provides a smart analgesic pump recovery robot control system, the system comprising:

[0092] A recycling instruction receiving module 11 is used to receive an analgesic pump recycling instruction through a user terminal, wherein the analgesic pump recycling instruction includes a recycling ward location; a moving path obtaining module 12 is used to perform path planning according to the recycling ward location to obtain a robot moving path; an image information acquisition module 13 is used to control the mobile chassis of the analgesic pump recycling robot to move to the recycling ward location according to the robot moving path, and collect analgesic pump image information through a camera deployed on the analgesic pump recycling robot; a verification signal obtaining module 14 is used to analyze the analgesic pump image information through an analgesic pump quality verification model to obtain a quality verification signal; a recycling and warehousing processing module 15 is used to remind the nursing staff to put the analgesic pump into the storage cabin when the quality verification signal is passed, and then control the storage cabin to fix the analgesic pump and transport it to the recycling processing center for warehousing.

[0093] Furthermore, the recovery instruction receiving module 11 is also used to perform the following steps: obtain an analgesic pump recovery robot in an idle state; collect the storage cavity image information of the analgesic pump recovery robot; identify the analgesic pump according to the storage cavity image information to obtain the analgesic pump identification result; when the number of analgesic pumps in the analgesic pump identification result is equal to 0, control the mobile chassis of the analgesic pump recovery robot to move according to the robot's moving path to reach the recovery ward location.

[0094] Furthermore, the recycling instruction receiving module 11 is also used to perform the following steps: collecting a storage cavity image sample data set, wherein the storage cavity image sample data set has an analgesic pump quantity identification data set; using the analgesic pump quantity identification data set as supervision and the storage cavity image sample data set as input, training a convolutional neural network to obtain an analgesic pump recognition model; processing the storage cavity image information according to the analgesic pump recognition model to obtain an analgesic pump quantity identification, and setting it as the analgesic pump recognition result.

[0095] Furthermore, the recycling instruction receiving module 11 is also used to perform the following steps: using the analgesic pump quantity identification dataset as supervision and the storage cavity image sample dataset as input, training a convolutional neural network to obtain a first analgesic pump identification model; extracting the first analgesic pump quantity identification dataset and the first storage cavity image sample dataset whose training accuracy of the first analgesic pump identification model is less than or equal to the training accuracy threshold; when the number of the first analgesic pump quantity identification dataset and the first storage cavity image sample dataset is greater than the convergence number, using the first analgesic pump quantity identification dataset as supervision and the first storage cavity image sample dataset as input, training a convolutional neural network to obtain a second analgesic pump identification model; until the number of the Mth analgesic pump quantity identification dataset and the Mth storage cavity image sample dataset is less than or equal to the convergence number, mean fully connecting the first analgesic pump identification model, the second analgesic pump identification model, and the Mth analgesic pump identification model to generate the analgesic pump identification model.

[0096] Furthermore, the recycling and warehousing processing module 15 is also used to perform the following steps: obtaining the first measured weight of the analgesic pump through the electronic scale of the storage compartment; obtaining the first initial weight and the density of the medicine liquid of the analgesic pump model associated with the analgesic pump; calculating the remaining medicine liquid according to the first initial weight, the first measured weight and the density of the medicine liquid, obtaining the remaining medicine amount, and summarizing the information of the analgesic pump.

[0097] Furthermore, the verification signal acquisition module 14 is also used to perform the following steps: according to the analgesic pump model associated with the analgesic pump, retrieve the analgesic pump standard image; compare the analgesic pump standard image and the analgesic pump image information through the analgesic pump quality verification model to obtain image similarity; when the image similarity is greater than or equal to the similarity threshold, set the quality verification signal to passed; otherwise, set the quality verification signal to failed.

[0098] Furthermore, the system is also used to perform the following steps: when the return button of the analgesic pump recovery robot is activated, the verification identity information is obtained; when the verification identity information is the same as the preset identity information, the storage cabin is controlled to release the analgesic pump, and the staff is reminded to return the analgesic pump to the warehouse; when the verification identity information is different from the preset identity information, a verification error instruction is generated to remind re-verification; when the analgesic pump is released without identity verification, an abnormal return to warehouse alarm signal is generated, and the facial information around the analgesic pump is recorded for storage.

[0099] Embodiment 4: A computer-readable storage medium can be used to store software programs, computer executable programs and modules, such as program instructions / modules corresponding to a smart analgesic pump recovery robot control method in the embodiment of the present application. The processor executes various functional applications and data processing of the computer device by running the software programs, instructions and modules stored in the memory, that is, realizing the above-mentioned smart analgesic pump recovery robot control method.

[0100] Any step of the method described above can be stored as a computer instruction or program in an unlimited computer memory, and can be called and recognized by an unlimited computer processor to implement any method in the embodiments of the present application, without any unnecessary restrictions.

[0101] Furthermore, the first or second mentioned above not only represents an order relationship, but also represents a specific concept, and / or refers to the selection of multiple elements individually or in whole. Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalent technologies, the present application intends to include these modifications and variations.

Claims

1. A control method for an intelligent analgesic pump recovery robot, characterized in that: Applied to an analgesic pump recovery robot, the analgesic pump recovery robot comprises a mobile chassis and a storage cabin, including: Receiving, via a user terminal, an analgesic pump recovery instruction, wherein the analgesic pump recovery instruction includes a recovery ward location; Perform path planning according to the recovery ward location to obtain a robot movement path; Controlling the mobile chassis of the analgesic pump recovery robot to move according to the robot moving path to reach the recovery ward location, and collecting analgesic pump image information through a camera deployed on the analgesic pump recovery robot; Analyzing the analgesic pump image information through an analgesic pump quality verification model to obtain a quality verification signal; When the quality verification signal is passed, the nursing staff is reminded to put the analgesic pump into the storage compartment, and the storage compartment is controlled to fix the analgesic pump and transport it to the recycling and processing center for return to the warehouse.

2. The method according to claim 1, characterized in that Controlling the mobile chassis of the analgesic pump recovery robot to move according to the robot movement path to reach the recovery ward location includes: Obtained the idle pain pump recovery robot; Collecting image information of the storage cavity of the analgesic pump recovery robot; Perform analgesia pump identification according to the storage cavity image information to obtain analgesia pump identification result; When the number of analgesic pumps in the analgesic pump identification result is equal to 0, the mobile chassis of the analgesic pump recovery robot is controlled to move according to the robot's moving path to reach the recovery ward location.

3. The method according to claim 2, characterized in that Performing analgesic pump identification according to the storage cavity image information to obtain analgesic pump identification result includes: Collecting a storage cavity image sample data set, wherein the storage cavity image sample data set includes an analgesia pump quantity identification data set; Using the analgesic pump quantity identification dataset as supervision and the storage cavity image sample dataset as input, a convolutional neural network is trained to obtain an analgesic pump identification model; The storage cavity image information is processed according to the analgesic pump identification model to obtain an analgesic pump quantity identifier, which is set as the analgesic pump identification result.

4. The method according to claim 3, characterized in that The analgesic pump quantity identification dataset is used as supervision, and the storage cavity image sample dataset is used as input to train a convolutional neural network to obtain an analgesic pump identification model, including: Using the analgesic pump quantity identification dataset as supervision and the storage cavity image sample dataset as input, a convolutional neural network is trained to obtain a first analgesic pump identification model; Extracting a first analgesia pump quantity identification data set and a first storage cavity image sample data set whose training accuracy of the first analgesia pump identification model is less than or equal to a training accuracy threshold; When the number of the first analgesic pump quantity identification data set and the first storage cavity image sample data set is greater than the convergence number, the first analgesic pump quantity identification data set is used as supervision and the first storage cavity image sample data set is used as input to train a convolutional neural network to obtain a second analgesic pump identification model; Until the number of the Mth analgesic pump quantity identification data set and the Mth storage cavity image sample data set is less than or equal to the convergence number, the first analgesic pump identification model, the second analgesic pump identification model and the Mth analgesic pump identification model are mean fully connected to generate the analgesic pump identification model.

5. The method according to claim 1, characterized in that When the quality verification signal is passed, the nursing staff is reminded to put the analgesic pump into the storage compartment, the storage compartment is controlled to fix the analgesic pump, and the analgesic pump is transported to the recycling processing center for storage, and further includes: Obtaining a first measured weight of the analgesic pump by means of an electronic scale in the storage compartment; Obtaining a first initial weight and a drug solution density of an analgesic pump model associated with the analgesic pump; The remaining liquid medicine is calculated based on the first initial weight, the first measured weight and the density of the liquid medicine to obtain the remaining amount of medicine, and the analgesic pump is summarized and labeled.

6. The method according to claim 1, characterized in that The analgesic pump image information is analyzed by the analgesic pump quality verification model to obtain a quality verification signal, including: Retrieving a standard image of analgesia pump according to the associated analgesia pump model of the analgesia pump; Comparing the analgesic pump standard image with the analgesic pump image information through the analgesic pump quality verification model to obtain image similarity; When the image similarity is greater than or equal to a similarity threshold, setting the quality verification signal as passed; Otherwise, the quality verification signal is set to fail.

7. The method according to claim 1, characterized in that include: When the return button of the analgesic pump recovery robot is activated, the verification identity information is obtained; When the verified identity information is the same as the preset identity information, the storage compartment is controlled to release the analgesic pump, and the staff is reminded to return the analgesic pump to the warehouse; When the verification identity information is different from the preset identity information, a verification error instruction is generated to remind re-verification; When the analgesic pump is released without identity verification, an abnormal storage alarm signal is generated, and the facial information around the analgesic pump is recorded and stored.

8. An intelligent analgesic pump recovery robot, characterized in that: include: The user end is used to interact with the user to receive the analgesia pump recovery instruction; A central control unit, used to implement a control method for an intelligent analgesic pump recovery robot as described in any one of claims 1 to 7; Mobile chassis, used to control the movement of the robot; A storage compartment for holding analgesia pumps; An electronic scale installed in the storage cabin for collecting the weight of the analgesic pump; A visual device is used to acquire the required images for the analgesic pump recovery robot.

9. An intelligent analgesic pump recovery robot system, characterized in that: A method for executing an intelligent analgesic pump recovery robot according to any one of claims 1 to 7, the system comprising: A recovery instruction receiving module, used to receive an analgesic pump recovery instruction through a user terminal, wherein the analgesic pump recovery instruction includes a recovery ward location; A moving path acquisition module, used for performing path planning according to the location of the recovery ward to obtain a moving path of the robot; An image information acquisition module, used to control the mobile chassis of the analgesic pump recovery robot to move to the recovery ward according to the robot's moving path, and to collect analgesic pump image information through a camera deployed on the analgesic pump recovery robot; A verification signal acquisition module, used to analyze the analgesic pump image information through an analgesic pump quality verification model to obtain a quality verification signal; The recycling and warehousing processing module is used to remind the nursing staff to put the analgesic pump into the storage cabin when the quality verification signal is passed, control the storage cabin to fix the analgesic pump, and transport it to the recycling processing center for warehousing.

10. A computer-readable storage medium, characterized in that: The storage medium stores a computer program, so the computer program is used to execute the intelligent analgesic pump recovery robot method described in any one of claims 1 to 7.

Images