Safety control method, system and equipment for lamina cutting based on multi-mode information fusion

Abstract

The present invention discloses a laminectomy safety control method, system, and device based on multimodal information fusion, relating to the technical field of laminectomy safety control. The method primarily comprises inputting the ultrasonic osteotome electrical impedance, feed force, and motor current values ​​during laminectomy into a trained time series classification network to obtain a classification result at the current moment. Based on the classification result, a control instruction is then determined to determine whether to continue the cutting action. The present invention can maximize the safety of laminectomy surgery.

Description

Technical Field

[0001] The present invention relates to the technical field of lamina cutting safety control, and in particular to a lamina cutting safety control method, system and equipment based on multi-mode information fusion. Background Art

[0002] Laminectomy is a key step in spinal surgery. It carries a high risk of damage to nerve roots, spinal cord, etc., a high incidence of complications, limited operating space in the surgical area, difficulty in accessing and operating surgical tools, and a heavy reliance on the doctor's experience and feeling during blind surgery. These problems lead to a high incidence of sequelae such as insufficient laminectomy and nerve root / dural sac / spinal cord damage.

[0003] There are two main indicators to define whether the laminectomy is successful. One indicator is the effectiveness of the cutting, that is, the thickness of the remaining lamina after the resection. If the remaining lamina is too thick, it will make it difficult to remove the lamina and the cutting effect will be unsatisfactory. Figure 1 As shown in (a); Another indicator is whether the vertebral plate is penetrated during the cutting process. If penetrated, it is very easy to cause damage to the patient's spinal cord and nerve roots. Figure 1 Therefore, the most ideal cutting state should be "almost through but not through", that is, ensuring safety while making it easy to remove the lamina, as shown in (c). Figure 1 As shown in (b).

[0004] Tool penetration identification is crucial for ensuring safe laminectomy. Force signals are a key factor in manual laminectomy. They represent direct feedback from bone tissue on the cutting tool and are strongly correlated with bone quality. However, force signal acquisition suffers from sensor complexity, high noise levels, susceptibility to interference, and interference from calibration thresholds.

[0005] The motor's current is an indirect way to obtain the grinding load, which is measured and calculated by the internal circuit. It is easy to collect and has strong anti-interference and precision. However, the accuracy is not high when using current alone for bone identification. The ultrasonic bone knife is a highly efficient bone tissue cutting tool that breaks bone tissue through mechanical effects, cavitation effects, etc. When cutting, the ultrasonic bone knife has the characteristics of cutting hard but not soft to a certain extent. It allows for a short-distance breakthrough when decompressing the spinal lamina, and is therefore more complete than a traditional grinding drill. The electrical impedance of the ultrasonic bone knife can intuitively reflect the electrical characteristics of the ultrasonic transducer, which plays a vital role in measuring the reliability of the ultrasonic transducer and matching the power supply. It can also be used to reflect the size of the load. However, the accuracy is not high when using electrical impedance alone for bone identification, and it is difficult to meet surgical requirements. Summary of the Invention

[0006] In view of this, the object of the present invention is to provide a laminectomy safety control method, system and equipment based on multi-mode information fusion.

[0007] To achieve the above object, the present invention provides the following solutions:

[0008] In a first aspect, the present invention provides a laminectomy safety control method based on multi-mode information fusion, comprising:

[0009] Acquiring multi-mode information of lamina cutting at the current moment, wherein the multi-mode information includes an ultrasonic osteotome electrical impedance, a feed force value, and a motor current value;

[0010] Inputting the multimodal information of lamina cutting at the current moment into the trained time series classification network to obtain the classification result at the current moment; the classification result includes the result of the ultrasonic bone scalpel penetrating the lamina and the result of the ultrasonic bone scalpel not penetrating the lamina;

[0011] Determine a control instruction according to the classification result at the current moment; the control instruction includes a continue cutting instruction and a stop cutting instruction;

[0012] Among them, the trained time series classification network is obtained by training the time series classification network through a sample data set; the sample data set includes multiple sample data pairs, the sample data pairs include input data and label data, and the timestamp of the input data and the timestamp of the label data in the same sample data pair are the same; the input data includes multimodal information with a timestamp, and the label data is image information of the ultrasonic bone knife penetrating the vertebral plate with a timestamp.

[0013] Optionally, the time series classification network includes a fully convolutional neural network, a long short-term memory network, a fusion network and a classifier;

[0014] The fully convolutional neural network is used to determine the first feature map based on the multimodal information during lamina cutting at the current moment; the long short-term memory network is used to determine the second feature map based on the multimodal information during lamina cutting at the current moment; the fusion network is used to fuse the first feature map and the second feature map to obtain a fused feature map; the classifier is used to determine the classification result at the current moment based on the fused feature map.

[0015] Optionally, the multimodal information of the current moment of laminectomy is input into a trained time series classification network to obtain the classification result at the current moment, specifically including:

[0016] Normalize the ultrasonic bone scalpel electrical impedance, feed force value and motor current value during lamina cutting at the current moment;

[0017] The scale-normalized information is input into the trained time series classification network to obtain the classification result at the current moment.

[0018] In a second aspect, the present invention provides a lamina cutting safety control system based on multi-mode information fusion, comprising: a control module, a perception module, and an execution module;

[0019] The control module is used for the lamina cutting safety control method based on multi-mode information fusion described in the first aspect;

[0020] The continue cutting instruction is used to control the perception module to continue to collect multimodal information during lamina cutting and to control the execution module to continue the lamina cutting operation;

[0021] The stop cutting instruction is used to control the perception module to stop collecting multimodal information during lamina cutting and to control the execution module to stop performing the lamina cutting operation.

[0022] Optionally, the sensing module is mounted on a lamina cutting tool;

[0023] The lamina cutting tool comprises at least a motor, a screw-nut assembly and an ultrasonic bone cutter; the motor is connected to the ultrasonic bone cutter via the screw-nut assembly; wherein the motor is a feeding tool and the ultrasonic bone cutter is a cutting power tool.

[0024] Optionally, the sensing module includes an ultrasonic bone knife electrical impedance sensing unit, a feed force value sensing module and a motor current value sensing module; the motor current value sensing module is connected to the motor; the feed force value sensing module is arranged on the nut in the screw nut assembly; the ultrasonic bone knife electrical impedance sensing unit is connected to the ultrasonic bone knife.

[0025] Optionally, the feed force sensing module includes a first force sensor and a second force sensor respectively arranged at both ends of the nut, and the first force sensor is a sensor close to the output end of the motor; the feed force value is the difference between the first force value and the second force value; the first force value is the force value collected by the first force sensor, and the second force value is the force value collected by the second force sensor.

[0026] Optionally, the ultrasonic osteotome electrical impedance sensing unit is used to:

[0027] Get the total voltage U and total current I of the ultrasonic osteotome at the current moment, as well as the phase angle between the total voltage U and total current I at the current moment

[0028] According to the formula Calculate the current ultrasonic bone scalpel electrical impedance; where Z represents the ultrasonic bone scalpel electrical impedance and j is a complex number:

[0029] Optionally, the motor current value sensing module is an ammeter or a current sensor.

[0030] In a third aspect, the present invention provides an electronic device comprising a memory and a processor, wherein the memory is used to store a computer program, and the processor runs the computer program to enable the electronic device to execute the lamina cutting safety control method based on multi-mode information fusion according to the first aspect.

[0031] According to the specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0032] The present invention uses a motor as a feeding tool and an ultrasonic bone knife as a cutting power tool. By collecting the ultrasonic bone knife electrical impedance, feeding force value and motor current value during lamina cutting, that is, multi-modal information, combined with a trained time series classification network, it can accurately determine whether to continue the cutting action, thereby maximizing the safety of lamina cutting surgery. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0034] Figure 1 A schematic diagram of cutting depth provided by an embodiment of the present invention; Figure 1 (a) is a schematic diagram of insufficient cutting depth; Figure 1 (b) is a schematic diagram of appropriate cutting depth; Figure 1 (c) is a relationship diagram of excessive cutting depth; Td refers to the delay of the penetration recognition algorithm. Td<0 means that the stop is too early and effective cutting is not possible; Td=0 means that the cutting depth is appropriate, ensuring effective and safe cutting. Td>0 means that the stop is too late and safe cutting is not possible; Db=Td*v, which is the penetration distance caused by the delay; v is the feed rate;

[0035] Figure 2 A schematic flow chart of a laminectomy safety control method based on multi-mode information fusion provided by an embodiment of the present invention;

[0036] Figure 3 A schematic structural diagram of a lamina cutting tool provided in an embodiment of the present invention;

[0037] Figure 4 An equivalent circuit diagram of the ultrasonic osteotome provided by an embodiment of the present invention near the resonance frequency;

[0038] Figure 5 A technical roadmap for data acquisition experiments provided by embodiments of the present invention;

[0039] Figure 6 A schematic diagram of the time series annotation results provided by an embodiment of the present invention;

[0040] Figure 7 A network structure diagram of a time series classification network provided by an embodiment of the present invention;

[0041] Figure 8 This is a flow chart of lamina cutting control based on multimodal information fusion provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0042] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0043] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0044] Example 1

[0045] like Figure 2 As shown, an embodiment of the present invention provides a lamina cutting safety control method based on multi-mode information fusion, comprising:

[0046] Step 100: Acquire multi-mode information of lamina cutting at the current moment, wherein the multi-mode information includes the electrical impedance of the ultrasonic osteotome, the feed force value, and the motor current value.

[0047] Step 200: Input the multimodal information of lamina cutting at the current moment into the trained time series classification network to obtain the classification result at the current moment; the classification result includes the result of the ultrasonic bone knife penetrating the lamina and the result of the ultrasonic bone knife not penetrating the lamina.

[0048] Step 300: Determine a control instruction according to the classification result at the current moment; the control instruction includes a continue cutting instruction and a stop cutting instruction.

[0049] Among them, the trained time series classification network is obtained by training the time series classification network through a sample data set; the sample data set includes multiple sample data pairs, the sample data pairs include input data and label data, and the timestamp of the input data and the timestamp of the label data in the same sample data pair are the same; the input data includes multimodal information with a timestamp, and the label data is image information of the ultrasonic bone knife penetrating the vertebral plate with a timestamp.

[0050] Furthermore, the time series classification network includes a fully convolutional neural network, a long short-term memory network, a fusion network and a classifier; the fully convolutional neural network is used to determine a first feature map based on the multimodal information of the lamina cutting at the current moment; the long short-term memory network is used to determine a second feature map based on the multimodal information of the lamina cutting at the current moment; the fusion network is used to fuse the first feature map and the second feature map to obtain a fused feature map; the classifier is used to determine the classification result at the current moment based on the fused feature map.

[0051] The fully convolutional neural network can use the convolution algorithm to collect data feature information, and can obtain feature information hidden in the sequence data except for the time dimension. The first feature map here is the feature set collected by the fully convolutional neural network; the long short-term memory network can obtain the feature information of the sequence data in the time dimension. The second feature map here is the feature set collected by the long short-term memory network.

[0052] Furthermore, step 200 specifically includes: performing scale normalization processing on the ultrasonic bone knife electrical impedance, feed force value and motor current value during lamina cutting at the current moment, and inputting the scale normalized information into the trained time series classification network to obtain the classification result at the current moment.

[0053] Example 2

[0054] An embodiment of the present invention provides a lamina cutting safety control system based on multi-mode information fusion, which includes a control module, a perception module, and an execution module.

[0055] The control module is used to execute the lamina cutting safety control method based on multi-modal information fusion described in Example 1; the continue cutting instruction is used to control the perception module to continue to collect multi-modal information during lamina cutting and control the execution module to continue the lamina cutting operation; the stop cutting instruction is used to control the perception module to stop collecting multi-modal information during lamina cutting and control the execution module to stop the lamina cutting operation.

[0056] The sensing module is installed in Figure 3 The lamina cutting tool shown in the figure comprises at least a motor 1, a screw-nut assembly, and an ultrasonic osteotome. The motor 1 is connected to the ultrasonic osteotome via the screw-nut assembly. The motor 1 serves as a feed tool, and the ultrasonic osteotome serves as a cutting power tool. The ultrasonic osteotome comprises a rear metal cylinder 2, a piezoelectric ceramic 3, a front metal cylinder 4, and an osteotome 5. The osteotome 5 is used to cut a vertebral body 7. The screw-nut assembly comprises a screw 8, a nut 9, and a connector 10. An operator 6 grips the housing that secures the motor 1.

[0057] The sensing module includes an ultrasonic bone knife electrical impedance sensing unit, a feed force value sensing module and a motor current value sensing module; the motor current value sensing module is connected to the motor 1; the feed force value sensing module is arranged on the nut 9 in the screw nut assembly; the ultrasonic bone knife electrical impedance sensing unit is connected to the ultrasonic bone knife

[0058] The feed force sensing module includes a first force sensor 11 and a second force sensor 12 respectively provided at both ends of the nut 9, and the first force sensor 11 is a sensor close to the output end of the motor 1; the feed force value is the difference between the first force value and the second force value; the first force value is the force value collected by the first force sensor, and the second force value is the force value collected by the second force sensor.

[0059] The ultrasonic bone scalpel electrical impedance sensing unit is used to obtain the total voltage U and total current I of the ultrasonic bone scalpel at the current moment, as well as the phase angle between the total voltage U and total current I at the current moment. And according to the formula Calculate the current ultrasonic bone scalpel electrical impedance; where Z represents the ultrasonic bone scalpel electrical impedance and j is a complex number: The motor current value sensing module is an ammeter or a current sensor.

[0060] Example 3

[0061] An embodiment of the present invention provides a lamina cutting safety control method based on multi-mode information fusion, comprising:

[0062] Step 1: Collect feed force and motor current values.

[0063] Step 2: Collect the electrical impedance of the ultrasonic bone scalpel.

[0064] Step 3: Prepare neural network training data.

[0065] Step 4: Build a neural network.

[0066] Step 5: Perform cutting control based on the binary classification results of the neural network.

[0067] Step 1 specifically includes: during the lamina cutting process, the ammeter collects the motor current value I2 in real time and uses it as the input signal 1.

[0068] The values ​​of the first force sensor 11 and the second force sensor 2 at both ends of the nut are collected and recorded as F1 and F2 respectively. The resultant cutting force (i.e., feed force value) F is expressed as: F=F1-F2, and the resultant force F is used as input signal 2.

[0069] Step 2 specifically includes: the ultrasonic bone knife uses a piezoelectric transducer as an acoustic energy generator, the front is connected to a front metal cylinder in contact with the load, and the rear is connected to a rear metal cylinder for fixing, such as Figure 3 As shown in the figure, a piezoelectric transducer converts electrical energy provided by an electronic source into mechanical vibrations. A sinusoidal voltage is applied to the sensor to drive the piezoelectric transducer to or near its resonance. At resonance, the piezoelectric transducer's surface vibration amplitude is maximized, resulting in the highest energy conversion rate. When the physical properties of the electroacoustic system change, some degree of tuning and frequency loss occurs. During spinal lamina decompression, bone tissue acts as a load on the ultrasonic osteotome, altering its operating characteristics.

[0070] The equivalent circuit of ultrasonic bone knife is as follows Figure 4 As shown, R0 is the static resistance, C0 is the static capacitance, R1 is the dynamic resistance, C1 is the dynamic capacitance, and L1 is the dynamic inductance. The characteristic of the Mason model is that the load is simulated as an equivalent impedance element. As the load increases or decreases, the number of impedance elements on the mechanical port increases, but this makes the derivation of the equivalent impedance elements complicated, and it is very difficult to calculate the impedance of the load alone. To simplify the calculation, the piezoelectric transducer impedance, circuit impedance, and equivalent impedance are combined into the total impedance Z, and the total voltage U and total current I of the ultrasonic bone knife are measured, as well as the phase angle between the current and voltage at that moment According to Ohm's law:

[0071]

[0072] Where j is a complex number:

[0073] The calculated impedance Z of the ultrasonic osteotome is used as input signal 3.

[0074] Step three specifically includes: feed force, motor current, and ultrasonic osteotome electrical impedance. These three are essentially one-dimensional data that changes over time, and each can, to a certain extent, reflect the bone density during the cutting process. However, due to the high safety requirements of the surgery, this embodiment combines these three data types to improve the classification accuracy of the neural network. Since the neural network classification method is based on data-driven methods, data processing and neural network training are required.

[0075] After the experimental environment is set up, the operator grasps the lamina cutting tool to grind the lamina. The feed force value, motor current value, and ultrasonic bone knife electrical impedance are collected while grinding. The above three signals are saved while the system clock is saved as a timestamp to facilitate subsequent data analysis. In addition, a high-definition camera is used to shoot the bottom of the lamina to record the penetration process, and the system clock is also saved as a timestamp. Grinding stops after a complete penetration of 3mm. The purpose is to save the complete cutting process from entering the outer cortical bone, breaking through the inner cortical bone, and then penetrating the lamina, to ensure data diversity, so that the gold standard data set takes into account both positive and negative samples. The data acquisition experimental technology roadmap is as follows: Figure 5shown.

[0076] A total of eight sheep vertebrae were planned for this experiment, and 233 sets of cutting data were collected. The process from the ultrasonic osteotome contacting the bone surface to penetrating the inner cortical bone was defined as a group. The experiment was conducted and the experimental data was saved.

[0077] Since the acquisition frequencies of the feed force signal, motor current signal, and ultrasonic bone knife electrical impedance signal are different, we first interpolate the signals through the timestamps corresponding to the three sets of data to unify the acquisition frequencies of the three signals. Then, we record the start and end time of each knife data through the timestamp on the video, and make time-labeled data. Figure 6 shown.

[0078] Step four specifically includes: the penetration identification of the vertebral plate cut by the ultrasonic bone knife is essentially a time series classification problem, and the penetration identification process is divided into two categories, 0 for no penetration and 1 for penetration. Since the hardness, thickness and cutting angle of the vertebral plate cut each time are different, if the traditional feature extraction + threshold method is used, it may lead to poor robustness of penetration identification. The fully convolutional neural network has been proven to have the most advanced performance in time series classification, and when combined with the long short-term memory network, it can significantly improve the performance of the fully convolutional neural network and reduce the size of the data set required for training. Therefore, the time series classification network constructed in this embodiment is capable of classifying time series data of multiple modal information.

[0079] LSTM-FCN (i.e., time series classification network) has two branches. The branch composed of LSTM can effectively extract the data features of force and displacement signals in the time dimension, while the branch composed of full convolution can extract other features hidden in the data. The network structure diagram is as follows Figure 7 shown.

[0080] Among them, Conv1D: one-dimensional convolution; BN: BatchNorm module; ReLU: ReLU activation function module; GlobalPooling: global pooling module; Dimension Shuffle: tensor dimension transformation; LSTM: LSTM (long short-term memory) module; Dropout: Dropout module; Concat: tensor splicing; Softmax: Softmax activation function module.

[0081] In this embodiment, the ultrasonic bone knife electrical impedance, feed force value and motor current value are spliced ​​to obtain a comprehensive signal with a feature dimension of 3, and the signal is used as the input of LSTM-FCN for feature extraction and state classification.

[0082] The scale-normalized data and the time series annotation results are input into the time series classification network and trained until the model converges, that is, the loss of the SoftMax layer meets the expected requirements or the iteration no longer changes it. At this time, the training is completed and the trained time series classification network can be used for lamina cutting test.

[0083] Step 5 specifically includes: To test the accuracy of multimodal data in lamina grinding penetration recognition, a laminectomy system based on ultrasonic bone knife electrical impedance and fully connected long short-term memory network is built. The system control logic is as follows: Figure 8 As the robot clamps the ultrasonic bone scalpel and performs cutting, the sensing module collects the ultrasonic bone scalpel's electrical impedance, motor current, and force sensor combined force information in real time. This information is normalized and then formed into a time series. This information is then input into the control module's time series classification network for binary classification. If the classification result is "not penetrated," the execution module continues cutting, and the motor continues moving. If the classification result is "penetrated," the execution module stops cutting, the motor retracts, and the ultrasonic bone scalpel stops operating.

[0084] This embodiment can improve the safety of surgery, lower the operating threshold for doctors, greatly reduce the cost of use, and shorten the learning curve.

[0085] Example 4

[0086] An embodiment of the present invention provides an electronic device including a memory and a processor, wherein the memory is used to store a computer program, and the processor runs the computer program to enable the electronic device to execute the lamina cutting safety control method based on multi-mode information fusion of embodiment 1.

[0087] Optionally, the above-mentioned electronic device may be a server.

[0088] In addition, an embodiment of the present invention further provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, the lamina cutting safety control method based on multi-mode information fusion of the first embodiment is implemented.

[0089] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. Reference can be made to the common and similar parts between the various embodiments. For the systems disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple, and the relevant parts can be referred to the method description.

[0090] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The above examples are only intended to help understand the method and core concept of the present invention. At the same time, those skilled in the art will find that the specific implementation methods and application scopes may vary based on the concept of the present invention. In summary, the contents of this specification should not be construed as limiting the present invention.

Claims

1. A lamina cutting safety control system based on multi-mode information fusion, characterized in that: include: Control module, perception module and execution module; The control module is used to execute the following method: Acquiring multi-mode information of lamina cutting at the current moment, wherein the multi-mode information includes an ultrasonic osteotome electrical impedance, a feed force value, and a motor current value; Inputting the multimodal information of lamina cutting at the current moment into the trained time series classification network to obtain the classification result at the current moment; the classification result includes the result of the ultrasonic bone scalpel penetrating the lamina and the result of the ultrasonic bone scalpel not penetrating the lamina; Determine a control instruction according to the classification result at the current moment; the control instruction includes a continue cutting instruction and a stop cutting instruction; The trained time series classification network is obtained by training the time series classification network using a sample data set; the sample data set includes multiple sample data pairs, each of which includes input data and label data, and the timestamp of the input data and the timestamp of the label data in the same sample data pair are the same; the input data includes multimodal information with a timestamp, and the label data is image information of the ultrasonic osteotome penetrating the lamina with a timestamp; The time series classification network includes a fully convolutional neural network, a long short-term memory network, a fusion network and a classifier; The fully convolutional neural network is used to determine a first feature map based on the multimodal information of the lamina cutting at the current moment; the long short-term memory network is used to determine a second feature map based on the multimodal information of the lamina cutting at the current moment; The fusion network is used to fuse the first feature map and the second feature map to obtain a fused feature map; the classifier is used to determine the classification result at the current moment according to the fused feature map.

2. A lamina cutting safety control system based on multi-mode information fusion according to claim 1, characterized in that: The multimodal information of the current moment of laminectomy is input into the trained time series classification network to obtain the classification results at the current moment, including: Normalize the ultrasonic bone scalpel electrical impedance, feed force value and motor current value during lamina cutting at the current moment; The scale-normalized information is input into the trained time series classification network to obtain the classification result at the current moment.

3. The lamina cutting safety control system based on multi-mode information fusion according to claim 1 is characterized in that: The sensing module is mounted on the lamina cutting tool; The lamina cutting tool comprises at least a motor, a screw-nut assembly and an ultrasonic bone cutter; the motor is connected to the ultrasonic bone cutter via the screw-nut assembly; wherein the motor is a feeding tool and the ultrasonic bone cutter is a cutting power tool; The sensing module includes an ultrasonic bone scalpel electrical impedance sensing unit, a feed force value sensing module and a motor current value sensing module; the motor current value sensing module is connected to the motor; the feed force value sensing module is arranged on the nut in the screw nut assembly; the ultrasonic bone scalpel electrical impedance sensing unit is connected to the ultrasonic bone scalpel; The feed force sensing module includes a first force sensor and a second force sensor respectively provided at both ends of the nut, and the first force sensor is a sensor close to the output end of the motor; the feed force value is the difference between the first force value and the second force value; the first force value is the force value collected by the first force sensor, and the second force value is the force value collected by the second force sensor; The continue cutting instruction is used to control the perception module to continue to collect multimodal information during lamina cutting and to control the execution module to continue the lamina cutting operation; The stop cutting instruction is used to control the perception module to stop collecting multimodal information during lamina cutting and to control the execution module to stop performing the lamina cutting operation.

4. A lamina cutting safety control system based on multi-mode information fusion according to claim 3, characterized in that: The ultrasonic bone scalpel electrical impedance sensing unit is used to: Get the total voltage of the ultrasonic osteotome at the current moment U and total current I , and the total voltage at the current moment U and total current I The phase angle between φ ; According to the formula Calculate the ultrasonic bone scalpel electrical impedance at the current moment; where, Z Indicates the electrical impedance of the ultrasonic bone scalpel. j For plural: .

5. The lamina cutting safety control system based on multi-mode information fusion according to claim 3 is characterized in that: The motor current value sensing module is an ammeter or a current sensor.

Images