Moxibustion manipulation track acquisition and reproduction method

Abstract

The invention discloses a moxibustion technique track acquisition and reproduction method, which relates to the technical field of moxibustion treatment, and comprises the following steps: installing a sensor and a patient feedback device, collecting moxibustion environment data and patient information, and dynamically adjusting initial parameters; according to the initial parameters, collecting a moxibustion technique track of a moxibustion technician, change of a moxibustion application environment and real-time feedback of a patient, performing multi-modal data fusion, generating a high-quality track data set, and training a machine learning model to obtain a track learning result; combining a track learning result with moxibustion application environment data and patient information, and controlling the moxibustion robot to execute manipulation track reproduction; the moxibustion applying effect is evaluated, and a moxibustion applying report is generated; according to the invention, the consistency and safety of the moxibustion effect are greatly improved by integrating a plurality of sensors and combining moxibustion environment data and patient personalized information. Meanwhile, moxibustion applying parameters can be automatically adjusted according to real-time feedback of the patient, the comfort level of the patient is further improved, and the treatment experience is optimized.

Description

Technical Field

[0001] The present invention relates to the technical field of moxibustion therapy, and in particular to a method for acquiring and reproducing a moxibustion manipulation trajectory. Background Art

[0002] With the widespread application of moxibustion therapy, how to achieve standardization and personalization of moxibustion techniques has become a technical problem that needs to be solved urgently. Traditional moxibustion therapy mainly relies on the experience and techniques of the moxibustion therapist. The accuracy of the techniques, the intensity of the moxibustion, and the time of the moxibustion have a direct impact on the treatment effect. However, the instability of the techniques, the interference of environmental factors, and the individual differences of patients often make it impossible to ensure the consistency of the moxibustion effect.

[0003] First, the current moxibustion technique capture technology is relatively simple, usually relying on single sensor data, and cannot fully reflect the multi-dimensional changes during the moxibustion process. Secondly, the existing moxibustion robot reproduction technique usually does not have the ability to dynamically adjust, and cannot be adjusted in time according to real-time environmental changes and patient feedback, resulting in inflexible operation. In addition, the existing machine learning algorithms have limitations in the accuracy of technique trajectory reproduction and the capture of timing characteristics, especially the weak ability to reproduce complex techniques. Summary of the invention

[0004] In view of the above existing problems, the present invention is proposed.

[0005] Therefore, the present invention provides a method for acquiring and reproducing moxibustion manipulation trajectories, which solves the problem of high-precision acquisition and reproduction of moxibustion manipulation trajectories.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] In a first aspect, an embodiment of the present invention provides a method for acquiring and reproducing moxibustion manipulation trajectories, which includes installing sensors and patient feedback devices, collecting moxibustion environment data and patient information, and dynamically adjusting initial parameters;

[0008] According to the initial parameters, moxibustion is performed, the moxibustion technique trajectory of the moxibustion practitioner, the changes in the moxibustion environment and the real-time feedback of the patient are collected, and multimodal data fusion is performed to generate a high-quality trajectory dataset;

[0009] Use high-quality trajectory datasets to train machine learning models and obtain trajectory learning results;

[0010] Combine the trajectory learning results with the moxibustion environment data and patient information to control the moxibustion robot to perform manipulation trajectory reproduction;

[0011] Collect real-time data of the moxibustion robot's manipulation trajectory reproduction process, evaluate the moxibustion effect, and generate a moxibustion report.

[0012] As a preferred solution of the method for acquiring and reproducing the moxibustion manipulation trajectory of the present invention, wherein: the sensor includes an infrared motion capture device, a temperature sensor, a humidity sensor and a pressure sensor;

[0013] The moxibustion environment data includes temperature and humidity;

[0014] The patient information includes age and skin sensitivity;

[0015] The initial parameters include the burning speed of the moxa stick, the moxibustion time, the strength of the manipulation and the distance.

[0016] As a preferred solution of the method for acquiring and reproducing the moxibustion manipulation trajectory of the present invention, the burning time of the moxa stick is dynamically adjusted according to the moxibustion environment data and the patient information, which is expressed as:

[0017]

[0018] Among them, t b Indicates the adjusted burning time of the moxa stick, t d Indicates the basic burning time, T t represents the target moxibustion temperature, T e Indicates the current ambient temperature;

[0019] According to the ambient humidity, the burning speed of the moxa stick is dynamically adjusted, expressed as,

[0020] v b =v d ·exp(-k2·(H e -H o ));

[0021] Among them, v b Indicates the burning speed of the adjusted moxa stick, v d represents the basic burning speed, k2 represents the humidity adjustment coefficient, H e Indicates the current ambient humidity, H o Indicates the optimal humidity range;

[0022] According to the patient's skin sensitivity, the intensity of moxibustion is dynamically adjusted, expressed as,

[0023]

[0024] Among them, F a Indicates the adjusted moxibustion intensity, F b represents the basic moxibustion strength, k3 represents the strength adjustment coefficient, S pt Indicates the patient's skin sensitivity level, S n Indicates standard skin sensitivity;

[0025] According to the patient's age, the moxibustion distance is dynamically adjusted, expressed as,

[0026] d a =d b +k4·ln(1+A pt –A n );

[0027] Among them, d a represents the adjusted moxibustion distance, d b represents the basic moxibustion distance, k4 represents the distance adjustment coefficient, A pt represents the patient's age, A n Indicates standard age.

[0028] As a preferred solution of the method for acquiring and reproducing the moxibustion manipulation trajectory of the present invention, wherein: according to the initial parameters, moxibustion is performed, and the moxibustion manipulation trajectory of the moxibustion practitioner, the changes in the moxibustion environment and the real-time feedback of the patient are collected, including the following steps:

[0029] When the moxibustion practitioner performs moxibustion, the motion trajectory of the moxibustion practitioner's hands and moxa sticks is recorded in three-dimensional coordinates through an infrared motion capture device, and the speed and acceleration of the technique are calculated based on the position changes;

[0030] Collect environmental data through temperature and humidity sensors to monitor changes in skin surface temperature and changes in the burning speed of moxa sticks;

[0031] Through the feedback device installed on the patient, the patient's subjective feelings of temperature and pain are received in real time.

[0032] As a preferred solution of the method for acquiring and reproducing the moxibustion manipulation trajectory of the present invention, the moxibustion manipulation trajectory of the moxibustion practitioner, the changes in the moxibustion environment and the real-time feedback of the patient are subjected to multimodal data fusion to generate a high-quality trajectory data set, including the following steps:

[0033] Use time synchronization mechanism to align data collected by different sensors according to timestamps;

[0034] Through multimodal data fusion technology, the movement trajectory of the moxibustion practitioner's hands and moxa sticks, the speed and acceleration of the calculation technique, the temperature changes on the skin surface, the changes in the burning speed of the moxa sticks, and the patient's feedback on temperature and pain are fused and processed to output a high-quality trajectory data set.

[0035] As a preferred solution of the method for acquiring and reproducing the moxibustion manipulation trajectory of the present invention, the method includes: using a high-quality trajectory data set to train a machine learning model to obtain a trajectory learning result, including the following steps:

[0036] Use standard deviation-based outlier detection to remove outliers from high-quality trajectory datasets;

[0037] Use Kalman filter algorithm to smooth high-quality trajectory data sets;

[0038] Format the processed high-quality trajectory dataset to obtain a multi-dimensional feature vector;

[0039] Select the LSTM network model, input the multi-dimensional feature vector, use the multi-layer LSTM network unit to process the time series data, the output of the LSTM network layer is connected to the fully connected layer to generate the predicted trajectory points, and the output layer provides the predicted value of the future trajectory points;

[0040] Use mean square error as the loss function to calculate the error between the trajectory predicted by the LSTM model and the actual trajectory;

[0041] The Adam optimizer is used to train the LSTM network model, and the learning rate is automatically adjusted to accelerate convergence. The trained LSTM network model is obtained and the trajectory learning results are output.

[0042] As a preferred solution of the method for acquiring and reproducing the moxibustion manipulation trajectory of the present invention, the trajectory learning result is combined with the moxibustion environment data and the patient information, and the moxibustion robot is controlled to perform manipulation trajectory reproduction, including the following steps:

[0043] Loading patient-personalized parameters including skin temperature sensitivity, pain threshold, body part characteristics, and health status parameters;

[0044] Through temperature and humidity sensors, current environmental parameters are collected in real time;

[0045] Combine the patient's personalized parameters and current environmental parameters with the trajectory learning results to generate three-dimensional coordinates and moxibustion intensity;

[0046] Start the moxibustion robot and control the movement of the moxa stick according to the three-dimensional trajectory. Through the pressure sensor, the moxibustion robot senses the force applied to the patient's skin surface in real time and adjusts according to the patient's feedback and skin temperature;

[0047] During the moxibustion process, the temperature sensor monitors the skin temperature and ambient temperature at the moxibustion point, and dynamically adjusts the height of the moxa stick and the moxibustion time according to the temperature changes;

[0048] Monitor the patient's real-time subjective feedback through feedback devices and adjust the manipulation parameters based on the feedback data;

[0049] Monitor the deviation between the real-time execution trajectory of the moxibustion robot and the trajectory predicted by the long short-term memory network model, define the deviation threshold, and detect abnormal situations of the moxibustion robot. When the deviation threshold is exceeded, an alarm will be triggered and the operation will be suspended. According to the real-time environmental parameters, personalized parameters and patient feedback, the three-dimensional coordinates and moxibustion intensity will be regenerated, and the moxibustion robot will re-execute the operation.

[0050] As a preferred solution of the method for acquiring and reproducing the moxibustion manipulation trajectory of the present invention, the real-time data of the moxibustion robot performing the manipulation trajectory reproduction process is collected, the moxibustion effect is evaluated, and a moxibustion report is generated, including the following steps:

[0051] After the moxibustion operation is completed, various sensor data and patient feedback data during the moxibustion process are collected;

[0052] By comparing the difference between skin temperature and target temperature, it is possible to evaluate whether the temperature is kept within the ideal range during moxibustion.

[0053] Compare the duration of moxibustion with the ideal moxibustion time to assess whether the time during moxibustion is kept within the ideal range;

[0054] According to the patients' pain feedback data, the patients' pain perception during moxibustion was evaluated;

[0055] A moxibustion report is generated based on the evaluation results of temperature, time, pain perception and real-time trajectory of the moxibustion robot.

[0056] In a second aspect, an embodiment of the present invention provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program is executed by the processor, any step of the method for acquiring and reproducing the moxibustion manipulation trajectory as described in the first aspect of the present invention is implemented.

[0057] In a third aspect, an embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, any step of the method for acquiring and reproducing the moxibustion manipulation trajectory as described in the first aspect of the present invention is implemented.

[0058] The beneficial effects of the present invention are as follows: by integrating an infrared motion capture device, a temperature sensor, a humidity sensor, and a pressure sensor, it is possible to realize multimodal data collection of moxibustion techniques, covering multi-dimensional factors such as temperature, humidity, and pressure during the moxibustion process. By fusion processing of multimodal data, a high-quality manipulation trajectory data set is generated, and training is performed using a machine learning algorithm, which can accurately capture the timing characteristics and dynamic changes of the technique. In addition, by combining moxibustion environment data, patient personalized information, and real-time feedback, the present invention can dynamically adjust parameters such as moxibustion intensity, time, and trajectory, ensuring personalization and precision during the moxibustion process, and greatly improving the consistency and safety of the moxibustion effect. At the same time, the moxibustion parameters can be automatically adjusted according to the patient's real-time feedback, further improving the patient's comfort and optimizing the treatment experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0060] Figure 1 This is a flow chart of the method for acquiring and reproducing the moxibustion manipulation trajectory in Example 1.

[0061] Figure 2 This is a flow chart of the moxibustion robot reproduction in Example 1. DETAILED DESCRIPTION

[0062] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0063] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0064] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0065] Example 1, reference Figure 1 and Figure 2, which is the first embodiment of the present invention, and provides a method for acquiring and reproducing moxibustion manipulation trajectories, comprising the following steps:

[0066] S1. Install sensors and patient feedback equipment, collect moxibustion environment data and patient information, and dynamically adjust initial parameters, including the following steps:

[0067] S1.1, sensors include infrared motion capture device, temperature sensor, humidity sensor and pressure sensor;

[0068] The infrared motion capture device is installed above the moxibustion operation table to ensure that it can cover the entire range of motion of the moxibustion practitioner's arm. The Mark points of the capture device should be installed on the back of the moxibustion practitioner's hand and part of the moxa stick to monitor the three-dimensional movement of the manipulation trajectory.

[0069] The temperature sensor is installed near the moxibustion operation area, preferably near the patient's skin, so that the temperature changes on the patient's skin surface can be monitored in real time, especially near the moxibustion acupoints. This sensor can help the system determine whether the temperature of moxibustion is appropriate, and the data will be transmitted to the system in real time;

[0070] The humidity sensor is installed in a suitable position in the moxibustion environment to ensure that it can monitor the changes in air humidity in real time. Humidity is an important factor affecting the burning speed of moxa sticks and the effect of moxibustion. The system needs to adjust the burning speed of moxa sticks according to the humidity.

[0071] The pressure sensor is a tactile feedback device installed on the patient's skin surface to monitor the pressure applied by the moxa stick on the patient's skin. The main function of the sensor is to ensure that the force applied during moxibustion is not too great to avoid damaging the patient's skin or causing discomfort.

[0072] The patient feedback device is used to monitor the patient's pain and temperature perception in real time. The device is installed on the patient's finger or other suitable position. The patient can feedback the pain or discomfort by slight pressure. The feedback data will be transmitted to the system in real time to adjust the moxibustion parameters.

[0073] Moxibustion environment data include temperature and humidity;

[0074] Patient information included age and skin sensitivity;

[0075] The initial parameters include the burning speed of the moxa stick, the time of moxibustion, the strength of the manipulation and the distance.

[0076] S1.2. Dynamically adjust the burning time of the moxa stick according to the moxibustion environment data and patient information, expressed as,

[0077]

[0078] Among them, tb Indicates the adjusted burning time of the moxa stick, t d Indicates the basic burning time (such as 10 minutes), T t Indicates the target moxibustion temperature (such as 40°C), T e Indicates the current ambient temperature;

[0079] According to the ambient humidity, the burning speed of the moxa stick is dynamically adjusted, expressed as,

[0080] v b =v d ·exp(-k2·(H e -H o ));

[0081] Among them, v b Indicates the burning speed of the adjusted moxa stick, v d represents the basic burning speed (such as burning 1 gram of moxa stick per minute), k2 represents the humidity adjustment coefficient (constant value, unit: 1 / %), H e Indicates the current ambient humidity, H o Indicates the optimal humidity range (such as 40%);

[0082] According to the patient's skin sensitivity, the intensity of moxibustion is dynamically adjusted, expressed as,

[0083]

[0084] Among them, F a Indicates the adjusted moxibustion intensity, F b represents the basic moxibustion force (such as 1 Newton), k3 represents the force adjustment coefficient (constant value, no unit), S pt Indicates the patient's skin sensitivity level (such as a score of 1-10), S n Indicates standard skin sensitivity (e.g. 5 is the standard value);

[0085] According to the patient's age, the moxibustion distance is dynamically adjusted, expressed as,

[0086] d a =d b +k4·ln(1+A pt –A n );

[0087] Among them, d a represents the adjusted moxibustion distance, d b represents the basic moxibustion distance (such as 3 cm), k4 represents the distance adjustment coefficient (constant value, unit: cm), A pt represents the patient's age, A n Indicates standard age (e.g. 30 years old);

[0088] Furthermore, the moxibustion distance can be dynamically adjusted according to the patient's skin surface temperature, expressed as:

[0089]

[0090] Where k1 represents the distance adjustment coefficient (constant value, unit: cm / ℃), T s Indicates the patient's skin surface temperature.

[0091] It should be noted that by installing sensors and patient feedback devices, collecting moxibustion environment data and patient information, and dynamically adjusting the initial parameters, it is possible to obtain multi-dimensional data during the moxibustion process, including the moxibustion environment (temperature, humidity, etc.) and the patient's personalized information (skin sensitivity, age, etc.), and dynamically adjust the initial parameters (such as moxa stick height, burning speed, and moxibustion intensity) based on these data. The environmental data and patient information collected by the sensors provide personalized, real-time basic data support for subsequent moxibustion, ensuring that the parameter settings during the moxibustion process are more accurate and meet the individual needs of patients. By dynamically adjusting the initial parameters, the system can adapt to the individual differences of different patients and changes in the moxibustion environment, greatly improving the accuracy and safety of moxibustion, and avoiding discomfort or insufficient efficacy caused by improper manual parameter settings.

[0092] S2. According to the initial parameters, moxibustion is performed, the moxibustion technique trajectory of the moxibustion practitioner, the changes in the moxibustion environment and the real-time feedback of the patient are collected, and multimodal data fusion is performed to generate a high-quality trajectory dataset, including the following steps:

[0093] S2.1. When the moxibustion practitioner performs moxibustion, the motion trajectory of the moxibustion practitioner's hands and moxa sticks is recorded in three-dimensional coordinates through an infrared motion capture device, and the speed and acceleration of the manipulation are calculated based on the position changes;

[0094] By calculating the speed and acceleration of the moxibustion practitioner's manipulation, we ensure that the dynamic characteristics of the manipulation can be restored in the machine learning model;

[0095] Furthermore, if the moxibustion technique involves complex rotation or tilt angles, the rotation angle and rotation speed of the moxa stick need to be recorded;

[0096] Environmental data is collected through temperature and humidity sensors to ensure that the impact of environmental changes on trajectory data during moxibustion can be recorded, and the temperature changes on the skin surface and the burning speed of the moxa stick can be monitored;

[0097] The patient's subjective feelings about temperature and pain are received in real time through feedback devices installed on the patient;

[0098] It should be noted that patients can use feedback devices to provide feedback at any time whether they feel the moxibustion is too hot or too cold. The temperature perception data fed back by the patients will be recorded in real time and compared with the actual skin temperature to ensure that the moxibustion technique is within the patient's comfort range. Patients can also provide feedback on the pain during the moxibustion process, and the moxibustion intensity or distance will be adjusted based on the pain feedback.

[0099] S2.2. Use the time synchronization mechanism to align the data collected by different sensors according to the timestamps to ensure the temporal consistency of multimodal data;

[0100] Through multimodal data fusion technology, the movement trajectory of the moxibustion practitioner's hands and moxa sticks, the speed and acceleration of the calculation technique, the temperature changes on the skin surface, the changes in the burning speed of the moxa sticks, and the patient's feedback on temperature and pain are fused and processed to output a high-quality trajectory data set.

[0101] It should be noted that multiple sensors are used to simultaneously collect the moxibustion practitioner's manipulation trajectory, changes in the moxibustion environment (such as temperature and humidity), and real-time feedback from patients (such as pain perception), and a high-quality moxibustion manipulation trajectory dataset is generated through multimodal data fusion technology. Multimodal data fusion technology ensures that data of different dimensions (such as manipulation trajectory, environmental parameters, and patient feedback) can be synchronized in time and fused to form a multidimensional, dynamic trajectory dataset. It can not only capture the key techniques in the moxibustion process, but also take into account the impact of the environment and patient feedback, making the generated trajectory dataset more comprehensive and accurate. This provides a high-quality data foundation for subsequent machine learning model training, ensuring that the model can accurately reproduce moxibustion techniques.

[0102] S3. Use high-quality trajectory data sets to train machine learning models and obtain trajectory learning results, including the following steps:

[0103] Use standard deviation-based outlier detection to remove outliers from high-quality trajectory datasets;

[0104] Use Kalman filtering or sliding average filtering algorithms to smooth high-quality trajectory data sets and reduce noise;

[0105] Format the processed high-quality trajectory dataset to obtain a multi-dimensional feature vector;

[0106] Select the long short-term memory network model, input the multi-dimensional feature vector, use multi-layer long short-term memory network units to process time series data, capture the temporal characteristics of the technique trajectory, and the output of each long short-term memory network unit is expressed as,

[0107] h t =LSTM(X(t),h t-1 );

[0108] Among them, h t represents the hidden state of the current time step, X(t) represents the multidimensional feature vector at time point t, and h t-1 Represents the hidden state of the previous time step, and LSTM represents the LSTM layer in the long short-term memory network model;

[0109] The output of the long short-term memory network layer is connected to the fully connected layer to generate the predicted trajectory points, which are expressed as,

[0110] P pred (t) = W h ·h t +b h ;

[0111] Among them, P pred (t) represents the predicted trajectory point, W h represents the weight matrix of the fully connected layer, b h represents the bias term;

[0112] The output layer provides the predicted value of future trajectory points, expressed as,

[0113] P pred (t) = (x pred (t), y pred (t), z pred (t));

[0114] Among them, x pred (t) represents the x-axis coordinate of the trajectory point predicted by the long short-term memory network model at time point t, and y pred (t) represents the y-axis coordinate of the trajectory point predicted by the long short-term memory network model at time point t, z pred (t) represents the z-axis coordinate of the trajectory point predicted by the long short-term memory network model at time point t;

[0115] Using the mean square error as the loss function, the error between the predicted trajectory and the actual trajectory of the long short-term memory network model is calculated, which is expressed as:

[0116]

[0117] Among them, L represents the loss value, P pred (t) represents the trajectory points predicted by the LSTM network model at time point t, T represents the total number of time steps, P(t) represents the actual trajectory points at time point t, that is, the trajectory points actually collected by the sensor, ||P pred (t)-P(t)|| represents the Euclidean distance between two points, which is used to measure the spatial distance between the trajectory points predicted by the LSTM model and the actual trajectory points;

[0118] The Adam optimizer (adaptive moment estimation optimizer) is used to train the long short-term memory network model, and the learning rate is automatically adjusted to accelerate convergence. The trained long short-term memory network model is obtained and the trajectory learning results are output.

[0119] It should be noted that high-quality data sets generated by multimodal data are used to train machine learning models (such as long short-term memory networks, LSTM) to learn the temporal characteristics of moxibustion manipulation trajectories. After training, the model can predict trajectory points that conform to actual moxibustion techniques. Through the training of machine learning models, the system can learn the complex manipulation trajectories of moxibustion practitioners, including the speed, strength, and change patterns of the techniques. This process ensures that the model can extract the optimal trajectory pattern from a large amount of historical data. It can realize the automatic reproduction of moxibustion techniques, and can dynamically adjust the trajectory to ensure that the ideal moxibustion effect can be achieved under different moxibustion conditions. The machine learning model can also improve the accuracy of trajectory reproduction through continuous training and optimization to ensure the standardization and personalization of moxibustion techniques.

[0120] S4, combining the trajectory learning results with the moxibustion environment data and patient information, and controlling the moxibustion robot to perform manipulation trajectory reproduction, including the following steps:

[0121] Loading patient-personalized parameters including skin temperature sensitivity, pain threshold, body part characteristics, and health status parameters;

[0122] Through temperature and humidity sensors, current environmental parameters are collected in real time;

[0123] Combine the patient's personalized parameters and current environmental parameters with the trajectory learning results to generate three-dimensional coordinates and moxibustion intensity;

[0124] Furthermore, the most suitable type of moxibustion technique can be selected by evaluating environmental parameters and personalized parameters. Common techniques include suspended moxibustion, sparrow pecking moxibustion and tail-swinging moxibustion. Suspended moxibustion is suitable for cases with sensitive skin, low pain threshold or thin skin at the moxibustion site. The trajectory pattern is relatively smooth and the force control is small; sparrow pecking moxibustion is suitable for areas with high pain threshold and thicker skin. The technique is characterized by fast and frequent contact, complex trajectory pattern and high force; tail-swinging moxibustion is suitable for cases with medium skin thickness and high patient tolerance. The technique is characterized by large swinging trajectory and moderate force, and is suitable for moxibustion over a larger area.

[0125] Start the moxibustion robot and control the movement of the moxa stick according to the three-dimensional trajectory. Through the pressure sensor, the moxibustion robot senses the force applied to the patient's skin surface in real time and adjusts according to the patient's feedback and skin temperature;

[0126] During the moxibustion process, the temperature sensor monitors the skin temperature and ambient temperature at the moxibustion point, and dynamically adjusts the height of the moxa stick and the moxibustion time according to the temperature changes;

[0127] The patient's real-time subjective feedback (such as temperature perception and pain perception) is monitored through feedback equipment, and the manipulation parameters are adjusted according to the feedback data. If the patient feedbacks that the temperature is too high, the distance between the moxa stick and the skin will be automatically increased or the moxibustion time will be shortened. If the patient feedbacks that the pain is strong, the moxibustion intensity will be reduced or the moxibustion trajectory will be changed to avoid sensitive areas.

[0128] Monitor the deviation between the real-time execution trajectory of the moxibustion robot and the trajectory predicted by the long short-term memory network model, define the deviation threshold, and detect abnormal situations of the moxibustion robot. When the deviation threshold is exceeded, an alarm will be triggered and the operation will be suspended. According to the real-time environmental parameters, personalized parameters and patient feedback, the three-dimensional coordinates and moxibustion intensity will be regenerated, and the moxibustion robot will re-execute the operation.

[0129] It should be noted that the trajectory learning results are combined with real-time moxibustion environment data and patient personalized information to generate three-dimensional coordinates and moxibustion strength, and the moxibustion robot is used to reproduce the technique trajectory. It is possible to generate personalized three-dimensional trajectories and moxibustion strengths to ensure that the moxibustion robot can dynamically adjust according to the patient's real-time feedback when reproducing the moxibustion technique. It ensures that the moxibustion robot can flexibly respond to changes in the environment and patient status during the moxibustion process, and avoid poor moxibustion effects caused by a single technique. By tracking real-time data, the height, strength and moxibustion time of the moxa stick can be automatically adjusted, greatly improving the personalization and accuracy of moxibustion.

[0130] S5, collecting real-time data of the moxibustion robot performing the manipulation trajectory reproduction process, evaluating the moxibustion effect, and generating a moxibustion report, including the following steps,

[0131] After the moxibustion operation is completed, various sensor data and patient feedback data during the moxibustion process are collected;

[0132] According to the patient's skin temperature sensitivity level, the patient's tolerance to temperature changes is judged, and the target temperature is defined. By comparing the difference between the skin temperature and the target temperature, it is evaluated whether the temperature is kept within the ideal range during the moxibustion process. If the difference is large (for example, greater than the set tolerance threshold, such as ±2°C), it means that the temperature control is not ideal, and the height of the moxa stick or the moxibustion time may need to be adjusted during the next moxibustion.

[0133] According to the patient's personalized parameters, an ideal moxibustion time is set, and the duration of moxibustion is compared with the ideal moxibustion time to evaluate whether the time during the moxibustion process is kept within the ideal range. If the time deviation is too large, the moxibustion strategy is adjusted according to different situations. If the time is too long, the preset time of the next moxibustion will be automatically shortened, or the moxibustion will be ended early according to the real-time temperature during the moxibustion process. If the time is too short, the preset time of the next moxibustion will be extended, or the burning time of the moxa stick will be extended according to the real-time temperature during the moxibustion process.

[0134] According to the patient's pain feedback data, the patient's pain perception during the moxibustion process is evaluated to ensure that the technique and strength are appropriate. The feedback is quantified into a numerical value (such as 1-10). If the pain feedback score is low (such as 1-3), it means that the moxibustion strength is moderate and the patient feels comfortable. If the pain feedback score is high (such as 7-10), it means that the moxibustion strength is too strong or the temperature is too high, which may cause discomfort or pain. According to the patient's pain feedback, the moxibustion strength or technique trajectory will be adjusted in real time;

[0135] A moxibustion report is generated based on the evaluation results of temperature, time, pain perception and real-time trajectory of the moxibustion robot.

[0136] It should be noted that in the process of the moxibustion robot performing the technique reproduction, a variety of sensor data and patient feedback data are collected in real time, and the moxibustion effect is evaluated. Finally, the system generates a detailed moxibustion report. Through real-time data collection, the system can comprehensively monitor and evaluate the temperature, time, moxibustion intensity and patient feedback during the moxibustion process to ensure that the moxibustion effect meets the expectations. In addition, the moxibustion report provides a detailed record of the moxibustion process, which is convenient for subsequent analysis and optimization. It ensures the safety and effectiveness of the moxibustion process and provides a basis for subsequent treatment. Through the evaluation of the moxibustion process, the system can continuously optimize future moxibustion operations and improve the efficacy. At the same time, it also provides patients with personalized treatment suggestions and enhances patients' sense of trust.

[0137] This embodiment also provides a computer device, which is suitable for the method of acquiring and reproducing moxibustion manipulation trajectories, including: a memory and a processor; the memory is used to store computer executable instructions, and the processor is used to execute computer executable instructions to implement the method of acquiring and reproducing moxibustion manipulation trajectories as proposed in the above embodiment.

[0138] The computer device may be a terminal, and the computer device includes a processor, a memory, a communication interface, a display screen and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device may be a touch layer covering the display screen, or a key, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse, etc.

[0139] The present embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, the method for acquiring and reproducing the moxibustion manipulation trajectory proposed in the above embodiment is implemented; the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, disk or optical disk.

[0140] In summary, the present invention can realize multimodal data collection of moxibustion techniques by integrating infrared motion capture devices, temperature sensors, humidity sensors and pressure sensors, covering multi-dimensional factors such as temperature, humidity, and pressure during the moxibustion process. Through the fusion processing of multimodal data, a high-quality manipulation trajectory data set is generated, and the machine learning algorithm is used for training, which can accurately capture the timing characteristics and dynamic changes of the technique. In addition, the present invention can dynamically adjust the parameters such as moxibustion intensity, time, trajectory, etc. by combining moxibustion environment data, patient personalized information and real-time feedback, ensuring personalization and precision during the moxibustion process, and greatly improving the consistency and safety of the moxibustion effect. At the same time, the system can automatically adjust the moxibustion parameters according to the patient's real-time feedback, further improving the patient's comfort and optimizing the treatment experience.

[0141] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A method for acquiring and reproducing moxibustion manipulation trajectories, characterized in that: include, Install sensors and patient feedback devices to collect moxibustion environment data and patient information, and dynamically adjust initial parameters; According to the initial parameters, moxibustion is performed, the moxibustion technique trajectory of the moxibustion practitioner, the changes in the moxibustion environment and the real-time feedback of the patient are collected, and multimodal data fusion is performed to generate a high-quality trajectory dataset; Use high-quality trajectory datasets to train machine learning models and obtain trajectory learning results; Combine the trajectory learning results with the moxibustion environment data and patient information to control the moxibustion robot to perform manipulation trajectory reproduction; Collect real-time data of the moxibustion robot's manipulation trajectory reproduction process, evaluate the moxibustion effect, and generate a moxibustion report.

2. The method for acquiring and reproducing moxibustion manipulation trajectories as claimed in claim 1, characterized in that: The sensors include an infrared motion capture device, a temperature sensor, a humidity sensor and a pressure sensor; The moxibustion environment data includes temperature and humidity; The patient information includes age and skin sensitivity; The initial parameters include the burning speed of the moxa stick, the moxibustion time, the strength of the manipulation and the distance.

3. The method for acquiring and reproducing moxibustion manipulation trajectories as claimed in claim 2, characterized in that: According to the moxibustion environment data and patient information, the burning time of the moxa stick is dynamically adjusted, which is expressed as, Among them, t b Indicates the adjusted burning time of the moxa stick, t d Indicates the basic burning time, T t represents the target moxibustion temperature, T e Indicates the current ambient temperature; According to the ambient humidity, the burning speed of the moxa stick is dynamically adjusted, expressed as, v b =v d ·exp(-k2·(H e -H o )); Among them, v b Indicates the burning speed of the adjusted moxa stick, v d represents the basic burning speed, k2 represents the humidity adjustment coefficient, H e Indicates the current ambient humidity, H o Indicates the optimal humidity range; According to the patient's skin sensitivity, the intensity of moxibustion is dynamically adjusted, expressed as, Among them, F a Indicates the adjusted moxibustion intensity, F b represents the basic moxibustion strength, k3 represents the strength adjustment coefficient, S pt Indicates the patient's skin sensitivity level, S n Indicates standard skin sensitivity; According to the patient's age, the moxibustion distance is dynamically adjusted, expressed as, d a =d b +k4·ln(1+A pt -A n ); Among them, d a represents the adjusted moxibustion distance, d b represents the basic moxibustion distance, k4 represents the distance adjustment coefficient, A pt represents the patient's age, A n Indicates standard age.

4. The method for acquiring and reproducing moxibustion manipulation trajectories as claimed in claim 3, characterized in that: According to the initial parameters, moxibustion is performed to collect the moxibustion technique trajectory of the moxibustion practitioner, the changes in the moxibustion environment and the real-time feedback of the patient, including the following steps: When the moxibustion practitioner performs moxibustion, the motion trajectory of the moxibustion practitioner's hands and moxa sticks is recorded in three-dimensional coordinates through an infrared motion capture device, and the speed and acceleration of the technique are calculated based on the position changes; Collect environmental data through temperature and humidity sensors to monitor changes in skin surface temperature and changes in the burning speed of moxa sticks; Through the feedback device installed on the patient, the patient's subjective feelings of temperature and pain are received in real time.

5. The method for acquiring and reproducing moxibustion manipulation trajectories as claimed in claim 4, characterized in that: The collected moxibustion technique trajectories, changes in the moxibustion environment, and real-time feedback from patients are fused into multimodal data to generate a high-quality trajectory dataset, including the following steps: Use time synchronization mechanism to align data collected by different sensors according to timestamps; Through multimodal data fusion technology, the movement trajectory of the moxibustion practitioner's hands and moxa sticks, the speed and acceleration of the calculation technique, the temperature changes on the skin surface, the changes in the burning speed of the moxa sticks, and the patient's feedback on temperature and pain are fused and processed to output a high-quality trajectory data set.

6. The method for acquiring and reproducing moxibustion manipulation trajectories as claimed in claim 5, characterized in that: Using high-quality trajectory datasets, training machine learning models and obtaining trajectory learning results include the following steps: Use standard deviation-based outlier detection to remove outliers from high-quality trajectory datasets; Use Kalman filter algorithm to smooth high-quality trajectory data sets; Format the processed high-quality trajectory dataset to obtain a multi-dimensional feature vector; Select the LSTM network model, input the multi-dimensional feature vector, use the multi-layer LSTM network unit to process the time series data, the output of the LSTM network layer is connected to the fully connected layer to generate the predicted trajectory points, and the output layer provides the predicted value of the future trajectory points; Use mean square error as the loss function to calculate the error between the trajectory predicted by the LSTM model and the actual trajectory; The Adam optimizer is used to train the LSTM network model, and the learning rate is automatically adjusted to accelerate convergence. The trained LSTM network model is obtained and the trajectory learning results are output.

7. The method for acquiring and reproducing moxibustion manipulation trajectories as claimed in claim 6, characterized in that: The trajectory learning results are combined with the moxibustion environment data and patient information to control the moxibustion robot to perform manipulation trajectory reproduction, including the following steps: Loading patient-personalized parameters including skin temperature sensitivity, pain threshold, body part characteristics, and health status parameters; Through temperature and humidity sensors, current environmental parameters are collected in real time; Combine the patient's personalized parameters and current environmental parameters with the trajectory learning results to generate three-dimensional coordinates and moxibustion intensity; Start the moxibustion robot and control the movement of the moxa stick according to the three-dimensional trajectory. Through the pressure sensor, the moxibustion robot senses the force applied to the patient's skin surface in real time and adjusts according to the patient's feedback and skin temperature; During the moxibustion process, the temperature sensor monitors the skin temperature and ambient temperature at the moxibustion point, and dynamically adjusts the height of the moxa stick and the moxibustion time according to the temperature changes; Monitor the patient's real-time subjective feedback through feedback devices and adjust the manipulation parameters based on the feedback data; Monitor the deviation between the real-time execution trajectory of the moxibustion robot and the trajectory predicted by the long short-term memory network model, define the deviation threshold, and detect abnormal situations of the moxibustion robot. When the deviation threshold is exceeded, an alarm will be triggered and the operation will be suspended. According to the real-time environmental parameters, personalized parameters and patient feedback, the three-dimensional coordinates and moxibustion intensity will be regenerated, and the moxibustion robot will re-execute the operation.

8. The method for acquiring and reproducing moxibustion manipulation trajectories according to claim 7, characterized in that: Collecting real-time data of the moxibustion robot performing manipulation trajectory reproduction, evaluating the moxibustion effect, and generating a moxibustion report includes the following steps: After the moxibustion operation is completed, various sensor data and patient feedback data during the moxibustion process are collected; By comparing the difference between skin temperature and target temperature, it is possible to evaluate whether the temperature is kept within the ideal range during moxibustion. Compare the duration of moxibustion with the ideal moxibustion time to assess whether the time during moxibustion is kept within the ideal range; According to the patients' pain feedback data, the patients' pain perception during moxibustion was evaluated; A moxibustion report is generated based on the evaluation results of temperature, time, pain perception and the real-time trajectory of the moxibustion robot.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method for acquiring and reproducing the moxibustion manipulation trajectory described in any one of claims 1 to 8 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method for acquiring and reproducing the moxibustion manipulation trajectory described in any one of claims 1 to 8 are implemented.

Images