Acupuncture robot system

Abstract

The invention discloses an acupuncture robot system in the field of acupuncture equipment, and the system comprises an acupoint recognition module which is used for collecting body surface image information and sign data of a patient, and obtaining acupoint position information through a pre-trained deep learning model; the mechanical arm control module is used for controlling the mechanical arm to perform acupuncture operation according to the acupoint position information and collecting force feedback information in real time; the monitoring feedback module comprises a physiological monitoring unit and a real-time feedback unit, and the physiological monitoring unit is used for monitoring the physiological reaction of the patient and obtaining physiological information; the real-time feedback unit is used for collecting and processing the acupoint position information, the force feedback information and the physiological information so as to adjust the acupuncture strength and angle in real time. According to the acupuncture robot system, the image recognition and sensor fusion technology is introduced, the accuracy of acupoint positioning is improved, physiological data of a patient can be collected in real time, the acupuncture applying force and depth are adjusted according to the physiological data, and the treatment effect is further improved.

Description

Technical Field

[0001] The present invention relates to the technical field of acupuncture equipment, and in particular to an acupuncture robot system. Background Art

[0002] Acupuncture is a traditional Chinese medical treatment method. It uses fine needles to puncture specific acupuncture points on the human body to regulate the body's qi and blood, dredge the meridians, regulate nerves, etc., and activate the body's own repair and immune systems to treat a variety of diseases and symptoms, including pain, stress relief, and regulating human metabolism. These specific acupuncture points on the human body are called acupoints. In the early days, Chinese medicine judged the acupuncture points of the human body by bone measurement, finger comparison, body surface marks, etc. The calibration at this time largely depended on the subjective judgment of Chinese medicine. Therefore, this type of method has certain subjectivity and limitations. At the same time, Chinese medicine needs to spend a lot of energy when performing acupuncture, which will inevitably cause deformation of the acupuncture technique and improper control of the strength in the later stage. These factors have a key impact on the treatment effect. Therefore, Chinese medicine needs a reliable, intelligent, and convenient smart device and operating system to help promote Chinese medicine to all levels of society and international platforms.

[0003] With the development of modern technology, mechanical design and automation, image processing, artificial intelligence and other technologies have continued to mature and are widely used in the medical field. With the development of image recognition technology, in terms of human acupoint identification and positioning, image recognition technology is used to identify human features and compare the distribution map of human meridians to calculate and locate according to the difference in relative position; in addition to the problem of acupoint identification and positioning, acupuncture in traditional Chinese medicine also uses a large number of automated equipment for treatment, among which the use of robotic arms is the most widely used. By carrying a variety of sensors and artificial intelligence technology, the robotic arms can autonomously perform acupuncture operations on patients. We collectively refer to the combination of these automated equipment and sensor equipment as acupuncture robots.

[0004] However, the current acupuncture robot uses image recognition technology, which has the problem of inaccurate feature point collection, resulting in insufficient accuracy in human acupoint identification and positioning. Equipment such as ultrasonic positioning technology and 3D human body scanning technology are expensive and difficult to popularize due to their complex operation. At the same time, the current acupuncture robot also lacks a real-time feedback mechanism. This seriously affects and restricts the practical application of acupuncture robots. Summary of the invention

[0005] In view of this, the embodiment of the present invention provides an acupuncture robot system to solve or partially solve the above problems. The organic integration of human acupoint recognition and positioning technology and acupuncture robots makes acupuncture robots more and more mature, further promoting the intelligent development of acupuncture.

[0006] The embodiment of the present invention provides an acupuncture robot system, comprising: Acupoint recognition module, which is used to collect the patient's body surface image information and physical sign data, and obtain acupoint location information through a pre-trained deep learning model; A mechanical arm control module, comprising a mechanical arm, for controlling the mechanical arm to perform acupuncture operations according to the acupoint position information and collecting force feedback information in real time; The monitoring feedback module includes a physiological monitoring unit and a real-time feedback unit. The physiological monitoring unit is used to monitor the patient's physiological response and obtain physiological information; the real-time feedback unit is used to collect and process the acupoint position information, force feedback information and physiological information to adjust the force and angle of acupuncture in real time.

[0007] Based on the above embodiments, the present application can be further improved as follows: In one embodiment of the present application, the acupoint recognition module includes a data acquisition unit, a data annotation unit, a data preprocessing unit, a feature extraction unit, a deep learning model unit and a result output unit. The data acquisition unit acquires the patient's body surface image information and related vital sign data through a depth camera and a sensor device; the data annotation unit is used to preliminarily annotate the acquired body surface image information and vital sign data; the data preprocessing unit is used to perform data cleaning, image denoising and data enhancement on the body surface image information and vital sign data; the feature extraction unit uses a convolutional neural network to extract feature data from the body surface image information and vital sign data; the deep learning model unit is a pre-trained model used to process the feature data in real time and output acupoint location information; the result output unit is used to identify the acupoint location information on the acquired body surface image information to provide intuitive feedback.

[0008] In one of the embodiments of the present application, the robotic arm control module includes: a sensor integration unit, which integrates a force sensor, a position sensor and a visual sensor; a motion control unit, which ensures that the robotic arm maintains a stable, accurate and moderate force during the acupuncture process through speed control, position control and force control; a control algorithm unit, which uses a predictive control model (MPC) and a motion planning algorithm to optimize the motion trajectory of the robotic arm; a control interface unit, which provides a motion control interface and a real-time feedback mechanism; a human-computer interaction unit, which uses the React Native framework to build a cross-platform application, provides mobile terminal control, and realizes real-time two-way communication through WebSocket and RESTful API.

[0009] One or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages: 1. Traditional acupoint positioning technology is not accurate enough and is often affected by external factors, such as skin characteristics and environmental conditions. This invention can optimize acupoint positioning in real time and significantly improve positioning accuracy by introducing image recognition and sensor fusion technology. This increase in accuracy can directly improve acupuncture results and ensure that patients receive more effective treatment.

[0010] 2. Traditional acupuncture operations usually rely on experienced physicians, are complex to operate, and have high requirements for equipment. By adopting an automated robotic arm control system, the present invention simplifies the operation process, reduces the reliance on the operator's professional skills, and thus reduces training costs and consumption of medical resources. This simplified operation method not only improves work efficiency, but also makes acupuncture treatment more popular.

[0011] 3. Existing acupuncture robots usually lack real-time interaction with the patient's physiological state and are not sensitive enough. By combining physiological monitoring technology, the present invention can collect the patient's physiological data in real time and dynamically adjust the acupuncture force and depth according to the patient's real-time response, further improving the treatment effect.

[0012] 4. Through precise mechanical control and feedback systems, the present invention effectively improves the overall therapeutic effect of acupuncture, so that patients can get better therapeutic effects. This intelligent acupuncture system not only represents the modernization of acupuncture treatment, but also provides new directions and possibilities for the future development of medical technology. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following is a brief introduction to the drawings required for the specific embodiments or the description of the prior art. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn according to the actual scale.

[0014] Figure 1 This is a structural block diagram of an acupuncture robot system in an embodiment of the present application. DETAILED DESCRIPTION

[0015] The present invention is further explained below in conjunction with specific implementation methods. It should be understood that these implementation methods are only used to illustrate the present invention and are not used to limit the scope of the present invention. After reading the present invention, various equivalent forms of modifications to the present invention by those skilled in the art all fall within the scope defined by the claims attached to this application.

[0016] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, further definition and explanation thereof is not required in subsequent drawings.

[0017] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "set", "install", "connect", and "connect" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0018] In the description of the present invention, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine different embodiments or examples described in the present invention and the features of different embodiments or examples without contradiction.

[0019] Example: like Figure 1 As shown, an acupuncture robot system includes an acupoint recognition module, a robotic arm control module and a monitoring feedback module.

[0020] The acupoint recognition module includes a data acquisition unit, a data annotation unit, a data preprocessing unit, a feature extraction unit, a deep learning model unit and a result output unit; the data acquisition unit obtains the patient's body surface image information and related vital sign data through a depth camera and sensor equipment; the data annotation unit is used to preliminarily annotate the acquired body surface image information and vital sign data; the data preprocessing unit is used to perform data cleaning, image denoising and data enhancement on the body surface image information and vital sign data; the feature extraction unit uses a convolutional neural network to extract feature data from the body surface image information and vital sign data; the deep learning model unit is a pre-trained model, which is used to process feature data in real time and output acupoint location information; the result output unit is used to mark the acupoint location information on the acquired body surface image information to provide intuitive feedback.

[0021] Specific: Data acquisition unit: The patient's body surface image information and related vital signs data are obtained through depth cameras and sensor devices, and divided into body shape group and skin color group according to the obtained image information. The body shape group is measured by the depth camera through infrared light to measure the distance between the object and the camera, and the depth image or point cloud data is selected to be generated. Based on the three-dimensional structure information provided by these data, the SACSegmentation module in the PCL open source library is used to extract the shape contour of the patient and segment it, and then the ConvexHull and ConcaveHull modules are used to calculate the convex hull and concave hull of the point cloud, so as to calculate the volume and divide it into four groups according to the calculation results: thin, normal, overweight, and obese. Regarding skin color, the skin color is identified by using HSV color space conversion on the RGB image, and it is divided into three categories: yellow, white, and black. Among them, the sensor equipment includes but is not limited to: temperature sensor, pressure sensor, physiological sensor; the vital signs data includes but is not limited to: temperature data, pressure data, bioelectric signals. The functions of the above vital signs data are as follows: Temperature data is associated with acupoint activity: In the theory of Chinese medicine meridians, the functional state of acupoints is usually closely related to local temperature changes. Real-time monitoring of skin temperature changes through sensor devices can help determine whether acupoints are in an active state or whether they are affected by external stimuli or disease states. Pressure data and acupoint location confirmation: Acupoints generally have high sensitivity and responsiveness, especially when stimulated, they show a strong tenderness reaction. By applying a certain amount of pressure through a pressure sensor, the exact location of certain acupoints and the degree of tenderness of acupoints during pathological changes can be determined. This is of great significance for accurately locating acupoints and evaluating their functions in practical applications. Bioelectric signals and acupoint functions: The electrophysiological activity of the meridian system often has special bioelectric reactions at acupoints, such as changes in local potential or differences in current conduction. These bioelectric signals are closely related to the activity of acupoints. Through bioelectric sensors such as ECG (electrocardiogram) and EMG (electromyography), the electrophysiological characteristics of different areas of the body surface can be monitored, further supporting the identification and functional analysis of acupoints.

[0022] Data annotation unit: Use automated annotation tools, such as CVAT, to automatically annotate the body surface images captured by the depth camera to generate preliminary annotations. On this basis, professional acupuncturists can further revise and improve the data to further ensure data accuracy.

[0023] Data preprocessing unit: (1) Data cleaning: Delete images that do not meet the requirements through blur detection, overexposure detection and low quality detection to ensure the quality of the data set, perform annotation checks to ensure the accuracy of acupoint annotations, and perform size adjustment and normalization processing to facilitate image input to the model while improving the stability of model training; (2) Image preprocessing: Filter the input image (such as Gaussian filtering) to remove noise and enhance image quality. Perform edge detection (such as Canny edge detection) to highlight the features around the acupoints; (3) Data enhancement: Generate more samples through data enhancement technology to increase the diversity of the training set.

[0024] Feature extraction unit: Convolutional neural network is used to automatically extract features through multi-layer convolution and pooling operations. Pre-trained models (such as VGG, ResNet) are used for transfer learning to improve the effect of feature extraction.

[0025] Deep learning model unit: Model training is required in advance. The deep learning model uses convolutional neural networks for image data processing and feature extraction. ResNet-101 is used as the model architecture to obtain more accurate data information. On the basis of ResNet, a custom fully connected layer is added to output acupoint categories and coordinate information. At the same time, the cross entropy loss function and Adam optimizer are defined to measure the gap between the model prediction and the true value, guide the learning direction of the model, accelerate the training process and improve the performance of the model. The data set is divided into training set, validation set and test set with a ratio of 70% training, 15% validation and 15% testing. The training set images are input into the ResNet-101 model for forward propagation. The error between the predicted result and the true label is calculated according to the loss function, and the model parameters are updated using the back propagation algorithm to reduce the loss. Set multiple epochs and evaluate the performance of the model on the validation set at the end of each epoch. Then, we tuned and deployed the model: after each epoch, we used the validation set to evaluate the model performance, recorded the loss and accuracy, and set the early stopping condition. When the performance of the validation set no longer improved within a few rounds, we stopped training to avoid overfitting. Then, we used the test set to conduct a final evaluation of the model, calculated the accuracy, recall, F1 score and other indicators, and visualized the test set results of the model to observe the comparison between the identified acupoint locations and the real annotations. Finally, we integrated it into the acupoint recognition module to process new input information in real time.

[0026] Furthermore, the trained deep learning model is deployed as follows: export the trained model to a usable format, TorchScript, so that the model can be deployed in a C++ environment, embed the TorchScript model into a C++ application for easy operation and integration with other systems and frameworks, and finally integrate the model into the application of the acupoint recognition module to design the user interface (UI). When performing acupoint recognition: input the extracted features into the model for prediction. The model will output the location and name of the acupoint; parse the output of the model to obtain the specific coordinates and related information of the acupoint.

[0027] Result output unit: 1. Result visualization: mark the identified acupoint locations on the image, generate visualization results, and provide intuitive feedback for users to view; 2. Feedback mechanism: allow users to confirm or correct the recognition results to improve the accuracy of the system, and provide feedback for subsequent model updates and optimization.

[0028] The deep learning model unit also has real-time monitoring and adjustment functions: real-time monitoring of the operating status of the acupoint recognition module, evaluation of recognition accuracy and response speed; dynamic adjustment of model parameters or recognition strategies based on monitoring data and user feedback to ensure the efficiency and accuracy of the system.

[0029] The robot control module includes: a sensor integration unit, which integrates force sensors, position sensors and visual sensors; a motion control unit, which ensures that the robot arm maintains a stable, accurate and moderate force during the acupuncture process through speed control, position control and force control according to the acupoint position information fed back by the acupoint recognition module and the physiological information fed back by the monitoring feedback module; a control algorithm unit, which uses a predictive control model (MPC) and a motion planning algorithm to optimize the motion trajectory of the robot arm; a control interface unit, which provides a motion control interface and a real-time feedback mechanism; a human-computer interaction unit, which uses the React Native framework to build a cross-platform application, provides mobile terminal control, and realizes real-time two-way communication through WebSocket and RESTful API.

[0030] Specific: Sensor integration unit: force sensor, used to collect acupuncture force data in real time, and transmit the data to the real-time feedback system for real-time monitoring through the communication protocol to ensure the safety and accuracy of acupuncture; position sensor, used to collect the acupuncture posture of the robotic arm, and transmit the data to the real-time feedback system for real-time monitoring through the communication protocol to ensure the precise positioning of acupuncture; visual sensor, used to assist the robotic arm in collecting patient surface image information in real time, and transmit it to the acupoint recognition algorithm model for data processing through the communication protocol, and output the specific coordinates and related information of the acupoints.

[0031] Motion control unit: speed control, adjusts the movement speed of the robotic arm to ensure that the robotic arm remains stable during the acupuncture process; position control, controls the end effector of the robotic arm to reach the target position, and uses the position sensor for feedback; force control, collects data through force sensing, and the real-time feedback system monitors the feedback torque information to ensure that the acupuncture force is moderate.

[0032] Control algorithm unit: Predictive control model (MPC), which predicts future system behavior and optimizes control input to achieve the optimal control target; motion planning algorithm, which uses linear interpolation for trajectory planning to calculate the motion path, avoid obstacles and optimize the motion trajectory.

[0033] Control interface unit: Motion control interface, which controls the movement of the robot arm through software; real-time feedback mechanism, which enables real-time monitoring and feedback, allowing the system to dynamically adjust the movement of the robot arm based on sensor data.

[0034] Human-computer interaction unit: remote control of the robotic arm, using the React Native framework to build a cross-platform application, providing mobile terminal control, using WebSocket to achieve real-time two-way communication, ensuring that the user's control commands can be instantly transmitted to the robotic arm, providing a RESTful API interface, and users can control the various functions of the robotic arm through HTTP requests. By combining WebSocket and RESTful API, an efficient remote control system for acupuncture robotic arms can be realized. RESTful API is used to manage and send control commands, while WebSocket is used for real-time status updates and feedback, which can effectively improve the response speed and accuracy of the system.

[0035] Furthermore, the safety mechanism of the robot control module: 1. Collision detection: Based on the movement trajectory of the robot, predict possible collision risks and take measures in advance. 2. Safe shutdown: In addition to being equipped with an emergency stop button, combined with a collision detection algorithm, when a collision or abnormal situation is detected, the shutdown program is automatically triggered to stop all movements immediately.

[0036] The monitoring and feedback module includes a physiological monitoring unit and a real-time feedback unit. The physiological monitoring unit is used to monitor the patient's physiological responses (such as heart rate, muscle response, etc.), obtain physiological information, and provide feedback to adjust the acupuncture strategy; the real-time feedback unit is used to collect and process acupoint position information, force and position feedback information and physiological information to adjust the strength and angle of acupuncture in real time.

[0037] Specific: The physiological monitoring unit includes: physiological sensors, which monitor the patient's physiological reactions (such as heart rate, muscle reactions, etc.) and provide feedback to adjust the acupuncture strategy. The physiological sensors here can be shared with the aforementioned sensor devices as long as they can obtain the corresponding information. Physiological sensors include but are not limited to: electromyography (EMG) sensors, which are used to monitor muscle activity and evaluate the effects of acupuncture on muscle relaxation and pain relief; electrocardiogram (ECG) sensors, which monitor heart rate changes and evaluate the effects of acupuncture on the cardiovascular system.

[0038] EMG rules and principles: Acupuncture stimulation can trigger muscle stress responses, such as muscle tension or relaxation. Moderate acupuncture intensity usually causes muscles to relax gradually, and EMG signals show muscle relaxation and a decrease in electrical activity. However, too strong (excessive acupuncture intensity) or too fast stimulation may cause excessive muscle contraction, and EMG signals show increased muscle electrical activity. This provides physiological feedback for adjusting the intensity of acupuncture. That is, through EMG monitoring, when muscle electrical activity (especially resting potential) returns to a lower level, it can be inferred that the intensity of acupuncture is appropriate and the expected muscle relaxation effect is achieved. If the electromyographic activity continues to increase, it may be necessary to reduce the intensity of acupuncture or adjust the acupuncture angle.

[0039] ECG rules and principles: Acupuncture can sometimes affect the heart rate by stimulating acupoints and regulating the autonomic nervous system. Moderate acupuncture may cause a slight decrease in heart rate, which is usually a sign of activation of the parasympathetic nervous system, helping to relax the body and relieve tension. Excessive acupuncture force may cause unnecessary stress response, resulting in an increase in heart rate. The purpose of acupuncture is to stabilize the heart rate by regulating the function of the nervous system. During acupuncture, observing the electrocardiogram may show that changes in heart rate reflect changes in acupuncture force. Excessive acupuncture force may cause sympathetic nerve excitement, arrhythmia or increased heart rate. Appropriate acupuncture force can promote a stable heart rate by affecting the parasympathetic nerve. The ECG sensor continuously monitors changes in the ECG signal, especially the morphology of the P wave, QRS complex and T wave, to infer whether the acupuncture force is reasonable. The ideal acupuncture force should promote the activity of the parasympathetic nerve, resulting in a slow and stable heart rate. If the acupuncture force is too strong, it may cause excessive sympathetic nerve excitement, which may be manifested as an increased heart rate or arrhythmia in the electrocardiogram. At this time, the acupuncture force needs to be reduced.

[0040] In traditional Chinese medicine acupuncture, "getting qi" is the key to acupuncture. It is generally believed that appropriate acupuncture strength can cause "qi sensation", such as soreness, swelling, numbness, etc. Through EMG and ECG feedback, combined with the "getting qi" theory of traditional Chinese medicine, it is possible to more accurately judge whether the acupuncture strength is appropriate.

[0041] The real-time feedback unit collects relevant information and then performs data fusion: First, the collected data is passed through the Kalman filter to reduce the noise in the sensor data to provide a more accurate state estimate, and then normalized. The Kalman filter is initialized to set the initial state estimate and covariance matrix, the current state is predicted according to the state transition equation, and the state estimate is updated with the new observation value. Finally, the fused state estimate is output through the Kalman filter. The output data after the Kalman filter is formatted into a time series format, and the CNN-LSTM hybrid model is used to jointly analyze the physiological data and torque data.

[0042] The monitoring feedback module also includes a user interface and interaction unit and a network and communication unit. The user interface and interaction unit uses visualization tools to provide real-time feedback of the analyzed and processed data in the form of icons, helping doctors to intuitively understand the acupuncture situation and adjust the plan in time; the network and communication unit uses an efficient CAN communication protocol to ensure the lowest data transmission delay between the various modules of the system; at the same time, some computing tasks are allocated to edge devices to reduce data transmission delays and improve real-time processing capabilities.

[0043] Furthermore, the monitoring and feedback module also includes a safety mechanism: real-time monitoring of the system's operating status to ensure that when an abnormality occurs (the patient's physiological data is not within the normal data range or the robotic arm module fails), an alarm and adjustment can be made in a timely manner through the real-time feedback system.

[0044] In the above-mentioned acupuncture robot system, data information of each process and unit is uploaded to the server for traceability and as historical data for big data analysis and processing for optimization and iteration.

[0045] The technical solution in the above embodiment of the present application, the acupuncture robot system has at least the following technical effects or advantages: 1. Physiological monitoring feedback function: Through the joint acquisition of multiple sensors, the Kalman filter is used to provide a more accurate state estimation in advance, and then the deep learning model is combined to further analyze and process the relevant data. The analysis and processing results are visualized through the big data center, and the data is transmitted to each module through the CAN protocol to ensure the real-time transmission of the state data, thereby ensuring the real-time nature of physiological monitoring and the timeliness of state feedback.

[0046] 2. Intelligent acupuncture treatment method: Provide an overall intelligent operation method process including acupoint identification and positioning, robotic arm acupuncture, patient treatment effect feedback and related safety protection mechanisms, providing a reference for subsequent intelligent and unmanned design, which is conducive to the overall intelligent construction of the acupuncture process.

[0047] 3. Joint acupoint identification method: Collect and upload data through sensor equipment and depth camera. The sensor equipment can perceive the changes in physiological characteristics of various parts of the body surface in real time, providing an important basis for acupoint positioning. For example, temperature sensors: local temperature changes can be used as an important indicator to evaluate the activity of acupoints. Temperature sensors can detect temperature increases or decreases in the acupoint area. These changes are related to the flow of qi and blood in the human body, the patency of meridians, and the activity of acupoints. By combining with the three-dimensional body surface model acquired by the depth camera, the location of acupoints can be determined more accurately. Pressure sensors: can be used to detect tender points and sensitive areas in the acupoint area. Especially during acupuncture or massage, the acupoint area may be more sensitive to external pressure than the surrounding area. By applying a certain amount of pressure, the sensor can help confirm the location of these sensitive points. Combined with the 3D images acquired by the depth camera, the accuracy of the acupoint location can be further verified. Physiological sensors (such as EMG, ECG, etc.): can monitor the electrical activity of muscles (such as relaxation or contraction), and can reversely infer the activity of acupoints or the response to acupuncture, thereby helping to locate and confirm the functional status of the target acupoints. During the treatment process, electrophysiological responses can be used to assist in determining whether the acupoints are correctly identified and stimulated. Since the data transmitted by the sensor device is real-time, the sensor data can not only help locate acupoints before acupuncture, but also provide dynamic adaptability throughout the treatment process.

[0048] Then, CVAT and manual methods are used to annotate the data collected and uploaded by the sensor equipment and the depth camera, and provide high-precision data information for training in the deep learning model. The accuracy of acupoint identification and positioning is improved through continuous data comparison. At the same time, visualization can be used to facilitate doctors' observation and a feedback mechanism can be designed to facilitate future model updates and optimization. 4. Standardized acupuncture with robotic arms: Through position sensors and force sensors, the acupuncture techniques of traditional Chinese medicine are simulated to achieve standardized acupuncture operations. At the same time, combined with the real-time feedback system, force and position data are continuously collected and sent, so as to keep the robotic arms standardized and standardized for treatment at all times, effectively reducing the mental loss of traditional Chinese medicine. 5. Human-computer interaction mechanism: Through the human-computer interaction mechanism, acupuncture can be remotely controlled. By receiving acupoint positioning information, the acupuncture robot arm can be remotely controlled to move, insert and remove needles. At the same time, it can also feedback patient status information and terminate or change the treatment plan in time.

[0049] Although the embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and are not to be construed as limitations of the present invention. A person skilled in the art may change, modify, replace and vary the above embodiments within the scope of the present invention.

Claims

1. An acupuncture robot system, characterized in that: include: Acupoint recognition module, which is used to collect the patient's body surface image information and physical sign data, and obtain acupoint location information through a pre-trained deep learning model; A robotic arm control module, used to control the robotic arm to perform acupuncture operations according to the acupoint position information, and to collect force feedback information in real time; A monitoring and feedback module, including a physiological monitoring unit and a real-time feedback unit, wherein the physiological monitoring unit is used to monitor the patient's physiological response and obtain physiological information; The real-time feedback unit is used to collect and process the acupoint position information, force feedback information and physiological information to adjust the force and angle of acupuncture in real time.

2. The acupuncture robot system according to claim 1, characterized in that: The acupoint recognition module includes a data acquisition unit, a data annotation unit, a data preprocessing unit, a feature extraction unit, and a deep learning model unit. The data acquisition unit obtains the patient's body surface image information and related vital sign data through a depth camera and a sensor device; The data annotation unit is used to perform preliminary annotation on the acquired body surface image information and vital sign data; the data preprocessing unit is used to perform data cleaning, image denoising and data enhancement on the body surface image information and vital sign data; The feature extraction unit uses a convolutional neural network to extract feature data from the body surface image information and vital sign data; The deep learning model unit is a pre-trained deep learning model, which is used to process the input feature data in real time and output the acupoint location information.

3. The acupuncture robot system according to claim 2, characterized in that: The acupoint identification module further includes a result output unit, and the result output unit is used to mark the acupoint position information on the collected body surface image information.

4. The acupuncture robot system according to claim 1, characterized in that: The deep learning model pre-trained in the deep learning model unit is deployed as follows: the pre-trained deep learning model is exported to TorchScript format, the TorchScript format model is embedded into a C++ application, and then integrated into the application of the acupoint recognition module, and a user interface is designed.

5. The acupuncture robot system according to claim 1, characterized in that: The robotic arm control module includes: a sensor integration unit that integrates a force sensor, a position sensor and a visual sensor; a motion control unit that controls the acupuncture position and strength of the robotic arm through speed control, position control and force control; a control algorithm unit that uses a predictive control model and a motion planning algorithm to optimize the motion trajectory of the robotic arm; and a control interface unit that provides a motion control interface and a real-time feedback mechanism.

6. The acupuncture robot system according to claim 5, characterized in that: The robotic arm control module also includes a human-computer interaction unit, which uses the React Native framework to build a cross-platform application, provides mobile terminal control, and realizes real-time two-way communication through WebSocket and RESTful API.

7. The acupuncture robot system according to claim 1, characterized in that: The physiological monitoring unit includes an electromyogram sensor for monitoring muscle activity and an electrocardiogram sensor for monitoring heart rate changes.

8. The acupuncture robot system according to claim 1, characterized in that: After the real-time feedback unit collects information, it needs to perform data fusion processing: use Kalman filtering to reduce noise, perform normalization processing, and use the CNN-LSTM hybrid model to jointly analyze the data.

9. The acupuncture robot system according to claim 1, characterized in that: The monitoring feedback module also includes a user interface and an interaction unit, and the user interface and the interaction unit use visualization tools to provide real-time feedback of the analyzed and processed data in the form of icons.

10. The acupuncture robot system according to claim 1, characterized in that: The monitoring feedback module also includes a network and communication unit, which uses the CAN communication protocol to achieve data transmission between various modules in the acupuncture robot system.

Images