AI model training method and mechanical arm safety detection method and system

Abstract

The invention provides a training method and device of an AI model for safety detection and a safety detection method and system of a mechanical arm of surgical navigation equipment based on a robot technology. When the AI model is trained, each group of training samples comprises the initial speed of the tail end of the mechanical arm during sudden stop triggering, the motor current of the mechanical arm, the time required by sudden stop of the tail end of the mechanical arm and a displacement deviation actual value serving as a sample label; placing each group of training samples in the AI model to obtain a displacement deviation prediction value for each group of training samples; and adjusting parameters of the AI model for safety detection according to a difference value between the actual displacement deviation value and the predicted displacement deviation value. And for the mechanical arm of the to-be-detected surgical navigation equipment, whether the sudden stop displacement of the mechanical arm meets the requirement or not during sudden stop triggering is determined through the trained AI model. According to the method, the safety of sudden stop displacement of the mechanical arm of the surgical navigation equipment based on the robot technology can be more effectively detected.

Description

Technical Field

[0001] One or more embodiments of the present invention relate to medical device technologies, and in particular, to a training method and device for an artificial intelligence (AI) model for safety detection, and a safety detection method and system for a robotic arm of a surgical navigation device based on robotics. Background Art

[0002] See Figure 1 , a surgical navigation equipment based on robotics consists of a computer with relevant software, a tracking device, a robotic arm, etc., and provides real-time feedback through the visualization of a digital model on the computer. The surgical navigation equipment based on robotics performs surgical operations on patients through the coordinated action of the robotic arm and the navigation system, significantly improving the accuracy and safety of surgeries.

[0003] In practical applications, the robotic arm of the surgical navigation equipment based on robotics can execute surgical actions according to a pre-planned path, such as osteotomy or drilling operations, and the operating range is restricted by a tactile sensor. Since the robotic arm has higher stability than a human hand, it is possible to eliminate the tremor of the human hand and avoid potential hazards caused by phenomena such as hand tremors that may occur during surgery for doctors.

[0004] See Figure 1 , during a surgical operation, during the operation of the robotic arm, there will be situations where the robotic arm is required to stop suddenly. For example, sudden power outages or equipment failures that cause the robotic arm to not move along the planned path or other emergencies all require the robotic arm to stop suddenly, that is, to stop moving within a short period of time. Otherwise, it may cause serious situations such as harm to the patient or damage to the robotic arm. It can be seen that whether the robotic arm can stop suddenly according to the specified requirements is an important indicator for evaluating the safety of the robotic arm. Therefore, a more effective method is needed to detect the safety of the robotic arm of the surgical navigation equipment based on robotics. Summary of the Invention

[0005] One or more embodiments of the present invention describe a training method and device for an AI model for safety detection, and a safety detection method and system for a robotic arm of a surgical navigation device based on robotics, which can more effectively detect the safety of the emergency stop displacement of the robotic arm of the surgical navigation device based on robotics.

[0006] According to a first aspect, there is provided a training method for an AI model for safety detection, where the AI model is applied to detect the safety of the emergency stop displacement of the robotic arm of a surgical navigation device; the method includes:

[0007] Obtain multiple groups of training samples;

[0008] Among them, each group of training samples includes: the initial velocity of the end of the robotic arm of the surgical navigation device when an emergency stop is triggered, the braking current of the robotic arm motor of the surgical navigation device when an emergency stop is triggered, the duration required for the end of the robotic arm to make an emergency stop, and the actual value of the displacement deviation as the sample label;

[0009] Input each group of training samples into the AI model to be trained for safety detection, and respectively obtain the predicted value of the displacement deviation for each group of training samples output by the AI model for safety detection;

[0010] According to the difference between the actual value of the displacement deviation corresponding to each group of training samples and the predicted value of the displacement deviation, adjust the parameters of the AI model for safety detection, so that the AI model learns to predict the deviation between the actual emergency stop displacement and the theoretical emergency stop displacement;

[0011] Obtain the trained AI model for safety detection.

[0012] The actual value of the displacement deviation represents: the deviation between the actual emergency stop displacement and the theoretical emergency stop displacement caused by the movement of the robotic arm of the surgical navigation device.

[0013] The formula of the AI model for safety detection is expressed as:

[0014] Δx = w1·v0 + w2·Ⅰ + w3·t + b

[0015] w1, w2, w3: weights (parameters that the model needs to learn)

[0016] b bias term

[0017] Δx: predicted value of the displacement deviation; v0: initial velocity of the end of the robotic arm of the surgical navigation device when an emergency stop is triggered; I: braking current of the robotic arm motor of the surgical navigation device when an emergency stop is triggered; t: duration required for the end of the robotic arm to make an emergency stop.

[0018] According to the second aspect, a method for safety detection of the robotic arm of a surgical navigation device based on robot technology is provided. The method includes:

[0019] Generate an emergency stop command;

[0020] Send the emergency stop command to the first surgical navigation device based on robot technology to be currently detected;

[0021] Trigger the motor of the robotic arm by the first surgical navigation device to brake the robotic arm;

[0022] Obtain the first initial velocity of the end of the robotic arm of the first surgical navigation device when the emergency stop command is sent;

[0023] Obtain the first braking current of the robotic arm motor of the first surgical navigation device when the emergency stop instruction is sent;

[0024] Obtain the first emergency stop duration required from when the emergency stop instruction is sent until the end of the robotic arm of the first surgical navigation device stops moving;

[0025] Calculate the theoretical emergency stop displacement corresponding to the robotic arm of the first surgical navigation device;

[0026] Input the first initial speed, the first braking current, and the first emergency stop duration into a pre-trained AI model for safety detection;

[0027] Obtain the displacement deviation corresponding to the robotic arm of the first surgical navigation device output by the AI model for safety detection;

[0028] Add the theoretical emergency stop displacement corresponding to the robotic arm of the first surgical navigation device to the displacement deviation corresponding to the robotic arm of the first surgical navigation device, and compare the calculated sum value with a pre-set emergency stop displacement threshold;

[0029] If the sum value is less than the pre-set emergency stop displacement threshold, determine that the emergency stop displacement of the robotic arm of the first surgical navigation device based on robotics meets the safety requirements;

[0030] If the sum value is not less than the pre-set emergency stop displacement threshold, determine that the emergency stop displacement of the robotic arm of the first surgical navigation device based on robotics does not meet the safety requirements.

[0031] The generation of the emergency stop instruction includes: simulating the emergency stop instruction through a step signal.

[0032] The obtaining of the initial speed of the end of the robotic arm of the first surgical navigation device when the emergency stop instruction is sent includes: the controller that controls the robotic arm of the first surgical navigation device receives the emergency stop instruction, and takes the current speed of the end of the robotic arm of the first surgical navigation device monitored by the controller when the controller receives the emergency stop instruction as the first initial speed.

[0033] The obtaining of the emergency stop duration required from when the emergency stop instruction is sent until the end of the robotic arm of the first surgical navigation device stops moving includes:

[0034] Record the first time point when the emergency stop instruction is issued, use an encoder or an IMU (Inertial Measurement Unit) to detect the second time point when the speed of the robotic arm drops to zero, and take the duration from the first time point to the second time point as the first emergency stop duration;

[0035] Or,

[0036] Divide the first initial speed by the deceleration of the motor braking, so as to obtain the first emergency stop duration required from the sending of the emergency stop instruction until the end of the robotic arm of the first surgical navigation device stops moving.

[0037] The calculation corresponds to the theoretical emergency stop displacement of the robotic arm of the first surgical navigation device, including:

[0038]

[0039] x theory is the theoretical emergency stop displacement of the robotic arm of the first surgical navigation device; v0 is the first initial speed, t is the first emergency stop duration, and a is the deceleration of the motor braking.

[0040] According to a third aspect, there is provided a training device for an AI model for safety detection, and the AI model is applied to detect the safety of the emergency stop displacement of the robotic arm of a surgical navigation device; the device includes:

[0041] A training sample acquisition module, configured to acquire multiple groups of training samples; wherein, each group of training samples includes: the initial speed of the end of the robotic arm of the surgical navigation device when the emergency stop is triggered, the braking current of the motor of the robotic arm of the surgical navigation device when the emergency stop is triggered, the duration required for the end of the robotic arm to make an emergency stop, and the actual value of the displacement deviation as a sample label;

[0042] A training execution module, configured to input each group of training samples into the AI model to be trained for safety detection, and respectively obtain the displacement deviation prediction value for each group of training samples output by the AI model for safety detection; according to the difference between the actual value of the displacement deviation and the displacement deviation prediction value corresponding to each group of training samples, adjust the parameters of the AI model for safety detection, so that the AI model learns to predict the deviation between the actual emergency stop displacement and the theoretical emergency stop displacement;

[0043] An AI model generation module, configured to obtain the trained AI model for safety detection.

[0044] According to a fourth aspect, there is provided a safety detection system for the robotic arm of a surgical navigation device based on robot technology, and the method includes:

[0045] An emergency stop instruction delivery module, configured to generate an emergency stop instruction; send the emergency stop instruction to the first surgical navigation device based on robot technology to be currently detected, so as to cause the first surgical navigation device to trigger the motor driver of the robotic arm to brake the robotic arm;

[0046] An initial speed acquisition module, configured to obtain the first initial speed of the end of the robotic arm of the first surgical navigation device when the emergency stop instruction is sent;

[0047] A braking current acquisition module, configured to obtain a first braking current of the robotic arm motor of the first surgical navigation device when an emergency stop instruction is sent;

[0048] An emergency stop duration acquisition module, configured to obtain a first emergency stop duration required for the end of the robotic arm of the first surgical navigation device to stop moving after the emergency stop instruction is sent;

[0049] A theoretical emergency stop displacement calculation module, configured to calculate a theoretical emergency stop displacement corresponding to the robotic arm of the first surgical navigation device;

[0050] An AI model processing module, configured to input the first initial velocity, the first braking current, and the first emergency stop duration into a pre-trained AI model for safety detection; wherein, this AI model is trained by the training device of the embodiment of the present invention; and obtain a displacement deviation corresponding to the robotic arm of the first surgical navigation device output by the AI model for safety detection;

[0051] An emergency stop displacement evaluation module, configured to add the theoretical emergency stop displacement corresponding to the robotic arm of the first surgical navigation device to the displacement deviation corresponding to the robotic arm of the first surgical navigation device, and compare the calculated sum value with a pre-set emergency stop displacement threshold; if the sum value is less than the pre-set emergency stop displacement threshold, it is determined that the emergency stop displacement of the robotic arm of the first surgical navigation device based on robotic technology meets the safety requirements; if the sum value is not less than the pre-set emergency stop displacement threshold, it is determined that the emergency stop displacement of the robotic arm of the first surgical navigation device based on robotic technology does not meet the safety requirements.

[0052] Thus, each embodiment of the present invention has at least the following beneficial effects:

[0053] 1. In the embodiment of the present invention, an AI model for safety detection can be trained. In this way, subsequent direct use of the AI model for prediction can determine whether the safety of the emergency stop displacement of the robotic arm of a surgical navigation device to be detected meets the requirements, and the implementation method is simpler.

[0054] 2. In the embodiment of the present invention, when detecting the safety of the emergency stop displacement of the robotic arm of a surgical navigation device to be detected, there is no need to set a reference point at the end of the robotic arm, no need to correspondingly set a laser probe, and no need to use a professional measurement tool, a three-dimensional measuring instrument. Therefore, there is no need for too many professional hardware tools to assist during the detection process, reducing the implementation cost and technical difficulty.

[0055] 3. In the embodiments of the present invention, there is no need to place the surgical navigation device based on robotic technology in the effective working space, that is, there is no need for a specific working environment for testing. Moreover, there is no need to use tools such as professional 3D measuring instruments. The method of the embodiments of the present invention can complete the detection process only through computer software system calculation. For example, remote detection can be achieved. Therefore, the method of the embodiments of the present invention is also suitable for detection in non-laboratory environments. For example, during the use of the surgical navigation device in a hospital after it is sold, it may be necessary to regularly detect whether the emergency stop displacement of the robotic arm meets the requirements in the hospital, so as to continuously verify the safety and further ensure the surgical safety. Another example is that due to difficulties such as long-distance transportation, it is necessary to remotely detect the emergency stop displacement of the robotic arm of the surgical navigation device, that is, detect it without the surgical navigation device on-site. And the method of the embodiments of the present invention can detect these situations, greatly expanding the application and development of the business.

[0056] 4. The method of the embodiments of the present invention does not need to rely on the proficiency of the tester in using the 3D measuring instrument, nor on the measurement accuracy of the 3D measuring instrument, and can more accurately detect whether the robotic arm can stop emergently according to the specified requirements.

[0057] 5. In the embodiments of the present invention, based on the prediction of the AI model and using the prediction result for calculation, it can be determined whether the safety of the emergency stop displacement of the robotic arm of a surgical navigation device to be detected meets the requirements. The entire detection process can be an automatic calculation process of the computer system, without consuming a lot of energy of the detection personnel, which is beneficial to improving the work efficiency of the detection personnel. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the following drawings are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained according to these drawings without creative efforts.

[0059] Figure 1 It is a schematic diagram of the application scenario of the surgical navigation device based on robotic technology.

[0060] Figure 2 It is a schematic diagram of detecting the emergency stop displacement of the robotic arm of the surgical navigation device in the prior art.

[0061] Figure 3 It is a flowchart of the training method of the AI model for safety detection in an embodiment of the present invention.

[0062] Figure 4It is a flowchart of a method for detecting the safety of a robotic arm of a surgical navigation device based on robotic technology in an embodiment of the present invention.

[0063] Figure 5 It is a schematic structural diagram of a training device for an AI model used for safety detection in an embodiment of the present invention.

[0064] Figure 6 It is a schematic structural diagram of a robotic arm safety detection system of a surgical navigation device based on robotic technology in an embodiment of the present invention. Detailed implementation manners

[0065] The solutions provided by the present invention will be described below with reference to the accompanying drawings.

[0066] First of all, it should be noted that the terms used in the embodiments of the present invention are only for the purpose of describing specific embodiments, and are not intended to limit the present invention. The singular forms of "a", "the" and "said" used in the embodiments of the present invention and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise.

[0067] It should be understood that the term " / and" used herein is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist simultaneously, and B exists alone. In addition, the character " / " in this article generally represents an "or" relationship between the associated objects before and after.

[0068] See Figure 2 , in the prior art, in order to detect whether the robotic arm can stop urgently according to the specified requirements, a reference point is usually set at the end of the robotic arm. When the robotic arm is triggered to stop urgently, the displacement of the reference point at the end of the robotic arm is detected, and it is judged whether the robotic arm can stop urgently according to the specified requirements based on whether the displacement of the end reference point is less than a predetermined value. However, the detection method of the prior art has at least the following disadvantages:

[0069] 1. In order to detect the displacement of the end reference point, the prior art needs to correspondingly set a laser probe for the reference point at the end of the robotic arm, and also needs to use professional measurement tools such as a three-dimensional measuring instrument to measure the three-dimensional coordinate values of the reference point at the end of the robotic arm. Therefore, additional professional hardware tools are required to assist in the detection process, increasing the implementation cost and technical difficulty.

[0070] 2. During the detection process of the prior art, the surgical navigation device based on robotic technology needs to be placed in the effective working space, that is to say, a specific working environment is required for testing. Moreover, professional tools such as three-dimensional measuring instruments are also needed. Therefore, it is actually only suitable for detection in the laboratory and not suitable for detection in non-laboratory environments. In actual business applications, detection in non-laboratory environments is often required. For example, when the surgical navigation device is used in a hospital after being sold, it may be necessary to regularly detect whether the emergency stop displacement of the robotic arm meets the requirements in the hospital, so as to continuously verify safety and further ensure surgical safety. Another example is that due to difficulties such as long-distance transportation, it is necessary to remotely detect the emergency stop displacement of the robotic arm of the surgical navigation device, that is, to detect without the surgical navigation device on site. It can be seen that the detection method of the prior art greatly restricts the development of the business.

[0071] 3. During the detection process of the prior art, when using a three-dimensional measuring instrument for measurement, the measurement accuracy highly depends on the use of the three-dimensional measuring instrument. If the tester uses the three-dimensional measuring instrument incorrectly or the measurement accuracy of the three-dimensional measuring instrument is inaccurate, it may lead to inaccurate sampling of a large amount of data during a large number of repeated samplings, resulting in the inability to accurately detect the displacement of the end reference point, and thus the inability to accurately detect whether the robotic arm can stop emergently according to the specified requirements.

[0072] 4. The process of using a three-dimensional measuring instrument in the prior art is time-consuming, resulting in a longer detection process for the emergency stop displacement of the robotic arm of the surgical navigation device, which is not conducive to improving the work efficiency of the detection personnel.

[0073] In view of various problems of the prior art, the present invention proposes a training method and device for an AI model for safety detection, a safety detection method and system for the robotic arm of a surgical navigation device based on robotic technology. The following will be described separately.

[0074] An embodiment of the present invention proposes a training method for an AI model for safety detection, and this AI model is applied to detect the safety of the emergency stop displacement of the robotic arm of a surgical navigation device; see Figure 3 , this method includes:

[0075] Step 301: Obtain multiple groups of training samples.

[0076] Among them, each set of training samples includes: the initial velocity of the end of the robotic arm of the surgical navigation device when an emergency stop is triggered, the braking current of the robotic arm motor of the surgical navigation device when an emergency stop is triggered, the duration required for the end of the robotic arm to make an emergency stop, and the actual value of the displacement deviation as the sample label; where the actual value of the displacement deviation represents the deviation between the actual emergency stop displacement and the theoretical emergency stop displacement caused by the movement of the robotic arm of the surgical navigation device.

[0077] Step 303: Input each set of training samples into the AI model to be trained for safety detection, and respectively obtain the predicted value of the displacement deviation for each set of training samples output by the AI model for safety detection.

[0078] Step 305: Adjust the parameters of the AI model for safety detection according to the difference between the actual value of the displacement deviation and the predicted value of the displacement deviation corresponding to each set of training samples, so that the AI model learns to predict the deviation between the actual emergency stop displacement and the theoretical emergency stop displacement.

[0079] Step 307: Obtain the trained AI model for safety detection.

[0080] According to Figure 3 As can be seen from the process shown, in the embodiment of the present invention, an AI model for safety detection can be trained. In this way, in subsequent direct use of the AI model for prediction, it can be determined whether the safety of the emergency stop displacement of the robotic arm of a surgical navigation device to be detected meets the requirements, and the implementation method is simpler.

[0081] In the embodiment of the present invention, the theoretical emergency stop displacement of the robotic arm can be calculated using the device parameters of the surgical navigation device. However, in actual operations, the displacement of the robotic arm during an emergency stop is likely to be affected by factors such as friction and mechanical vibration, causing the actual emergency stop displacement to deviate from the theoretical value. Therefore, in the embodiment of the present invention, an AI model is considered to learn the deviation between the theoretical emergency stop displacement and the actual emergency stop displacement, so as to compensate for this deviation during the actual detection process for detection.

[0082] Next Figure 3 The process shown will be described in detail.

[0083] First, step 301: Obtain multiple sets of training samples.

[0084] Among them, each set of training samples includes: the initial velocity of the end of the robotic arm of the surgical navigation device when the emergency stop is triggered, the braking current of the robotic arm motor of the surgical navigation device when the emergency stop is triggered, the duration required for the end of the robotic arm to make an emergency stop, and the actual displacement deviation value as the sample label; wherein, the actual displacement deviation value represents: the deviation between the actual emergency stop displacement and the theoretical emergency stop displacement caused by the movement of the robotic arm of the surgical navigation device.

[0085] In an embodiment of the present invention, for the deviation between the theoretical emergency stop displacement and the actual emergency stop displacement, the factors affecting this deviation include:

[0086] 1. The initial velocity of the end of the robotic arm of the surgical navigation device when the emergency stop is triggered (the greater this initial velocity, the greater the deviation between the final theoretical emergency stop displacement and the actual emergency stop displacement. For example, high-speed emergency stops are more difficult to control);

[0087] 2. The braking current of the robotic arm motor of the surgical navigation device when the emergency stop is triggered (this braking current reflects the magnitude of the braking force. The greater this braking current, the greater the braking force. Excessive braking force may cause vibration, resulting in a greater deviation between the final theoretical emergency stop displacement and the actual emergency stop displacement);

[0088] 3. The duration required for the end of the robotic arm to make an emergency stop (the greater this duration, the smaller the deviation between the final theoretical emergency stop displacement and the actual emergency stop displacement, because slower braking may be more stable).

[0089] Therefore, these factors are used to train the AI model so that the AI model can better learn the relationship between these factors and the above-mentioned deviation.

[0090] In an embodiment of the present invention, various sets of training samples can be obtained under different working conditions. For example, 5000 tests under random working conditions are carried out to obtain 5000 sets of training samples. Different working conditions include, for example:

[0091] - Different loads (0 - 5 kg)

[0092] - Different initial velocities (0.1 - 2 m / s)

[0093] - Different braking curves (exponential decay, step, etc.)

[0094] For step 303: Input each set of training samples into the AI model to be trained for safety detection, and respectively obtain the displacement deviation prediction value for each set of training samples output by the AI model for safety detection.

[0095] In an embodiment of the present invention, the AI model for safety detection can use a linear regression model, and the expression formula can be represented as:

[0096] Δx = w1·v0 + w2·Ⅰ + w3·t + b

[0097] w1, w2, w3: Weights (parameters that the model needs to learn)

[0098] b: Bias term

[0099] Δx: Predicted value of displacement deviation; v0: Initial velocity of the end of the robotic arm of the surgical navigation device when the emergency stop is triggered; I: Braking current of the motor driving the robotic arm of the surgical navigation device when the emergency stop is triggered; t: Duration required for the end of the robotic arm to make an emergency stop, i.e., the total duration required for the end of the robotic arm to decelerate from the moving state (initial velocity v0) to the stationary state (v = 0);

[0100] b, w1, w2, and w3 are all parameters that the AI model for safety detection needs to learn during the training process.

[0101] Through the above formula, the AI model is enabled to find the relationship between v0, I, t, and the predicted value of displacement deviation.

[0102] In other embodiments of the present invention, the AI model for safety detection may also use a neural network, such as a fully connected network.

[0103] For step 305: According to the difference between the actual value of displacement deviation and the predicted value of displacement deviation corresponding to each set of training samples, adjust the parameters of the AI model for safety detection so that the AI model learns to predict the deviation between the actual emergency stop displacement and the theoretical emergency stop displacement.

[0104] In the embodiments of the present invention, the gradient descent method may be used to adjust the parameters of the AI model for safety detection, such as the weights in the above formula, to gradually reduce the loss.

[0105] For step 307: Obtain the trained AI model for safety detection.

[0106] Through multiple rounds of training, such as 1000 rounds, when the AI model converges, the trained AI model for safety detection can be obtained.

[0107] An embodiment of the present invention also proposes a method for detecting the safety of the robotic arm of a surgical navigation device based on robotic technology. Refer to Figure 4 , this method includes:

[0108] Step 401: Generate an emergency stop command;

[0109] Step 403: Send the emergency stop command to the first surgical navigation device based on robotic technology that is currently to be detected;

[0110] Step 405: Trigger the motor driver of the robotic arm by the first surgical navigation device to brake the robotic arm;

[0111] Step 407: Obtain the first initial velocity of the end of the robotic arm of the first surgical navigation device when the emergency stop instruction is sent;

[0112] Step 409: Obtain the first braking current of the robotic arm motor of the first surgical navigation device when the emergency stop instruction is sent;

[0113] Step 411: Obtain the first emergency stop duration required from when the emergency stop instruction is sent until the end of the robotic arm of the first surgical navigation device stops moving;

[0114] Step 413: Calculate the theoretical emergency stop displacement corresponding to the robotic arm of the first surgical navigation device;

[0115] Step 415: Input the first initial velocity, the first braking current, and the first emergency stop duration into a pre-trained AI model for safety detection; wherein, this AI model is trained by using the method provided in any embodiment of the present invention;

[0116] Step 417: Obtain the displacement deviation corresponding to the robotic arm of the first surgical navigation device output by the AI model for safety detection;

[0117] Step 419: Add the theoretical emergency stop displacement corresponding to the robotic arm of the first surgical navigation device to the displacement deviation corresponding to the robotic arm of the first surgical navigation device, and compare the calculated sum value with a pre-set emergency stop displacement threshold; if this sum value is less than the pre-set emergency stop displacement threshold, it is determined that the emergency stop displacement of the robotic arm of the first surgical navigation device based on robotic technology meets the safety requirements; if this sum value is not less than the pre-set emergency stop displacement threshold, it is determined that the emergency stop displacement of the robotic arm of the first surgical navigation device based on robotic technology does not meet the safety requirements.

[0118] According to Figure 4 As can be seen from the shown process, a method for detecting the safety of the robotic arm of a surgical navigation device based on robotic technology proposed in an embodiment of the present invention has at least the following beneficial effects:

[0119] 1. In the embodiment of the present invention, when detecting the safety of the emergency stop displacement of the robotic arm of a surgical navigation device to be detected, there is no need to set a reference point at the end of the robotic arm, no need to correspondingly set a laser probe, and no need to use a professional measuring tool, a three-dimensional measuring instrument. Therefore, there is no need for too many professional hardware tools to assist during the detection process, reducing the implementation cost and technical difficulty.

[0120] 2. In the embodiments of the present invention, there is no need to place the surgical navigation device based on robotic technology in the effective working space, that is, there is no need for a specific working environment for testing. Moreover, there is no need to use tools such as professional 3D measuring instruments. The method of the embodiments of the present invention can complete the detection process only by computer software system calculation. For example, remote detection can be achieved. Therefore, the method of the embodiments of the present invention is also suitable for detection in a non-laboratory environment. For example, when the surgical navigation device is in use in a hospital after being sold, it may be necessary to regularly detect whether the emergency stop displacement of the robotic arm meets the requirements in the hospital, so as to continuously verify safety and further ensure surgical safety. Another example is that due to difficulties such as long-distance transportation, it is necessary to remotely detect the emergency stop displacement of the robotic arm of the surgical navigation device, that is, to detect without the surgical navigation device on site. And the method of the embodiments of the present invention can detect these situations, greatly expanding the application and development of the business.

[0121] 3. The method of the embodiments of the present invention does not need to rely on the proficiency of the tester in using the 3D measuring instrument, nor on the measurement accuracy of the 3D measuring instrument, and can more accurately detect whether the robotic arm can stop emergently according to the specified requirements.

[0122] 4. In the embodiments of the present invention, based on the AI model for prediction and using the prediction result for calculation, it can be determined whether the safety of the emergency stop displacement of the robotic arm of a surgical navigation device to be detected meets the requirements. The entire detection process can be an automatic calculation process of the computer system, without consuming a lot of energy of the tester, which is beneficial to improving the work efficiency of the tester.

[0123] The following Figure 4 explains each step shown below.

[0124] For step 401: Generate an emergency stop command.

[0125] In an embodiment of the present invention, the emergency stop command can be simulated by a step signal, similar to a switch changing from closed to fully open instantaneously.

[0126] For step 403: Send the emergency stop command to the first surgical navigation device based on robotic technology to be detected currently.

[0127] In an embodiment of the present invention, it can be that the test end (such as a detection agency) remotely sends the emergency stop command to the first surgical navigation device based on robotic technology to be detected currently.

[0128] For step 405: After the first surgical navigation device receives the emergency stop command, the first surgical navigation device triggers the motor of the robotic arm to brake the robotic arm.

[0129] In this step 405, the step signal forces the torque of the motor to suddenly change from the current value to 0. The magnitude of the braking current I of the motor driving the robotic arm when the emergency stop is triggered can be recorded.

[0130] For step 407: Obtain the initial velocity of the end of the robotic arm of the first surgical navigation device when the emergency stop instruction is sent. For ease of description, it is denoted as the first initial velocity.

[0131] In a surgical navigation device based on robotic technology, the controller will monitor the movement trajectory of the robotic arm and the velocity of the end of the robotic arm in real time. Therefore, in an embodiment of the present invention, the implementation method of this step 407 includes: The controller that controls the robotic arm of the first surgical navigation device receives the emergency stop instruction, and uses the current velocity of the end of the robotic arm of the first surgical navigation device monitored by the controller when the controller receives the emergency stop instruction as the first initial velocity.

[0132] For step 409: Obtain the braking current of the motor driving the robotic arm in the first surgical navigation device when the emergency stop instruction is sent. For ease of description, it is denoted as the first braking current.

[0133] In many applications, the braking current of the motor driving the robotic arm in the first surgical navigation device can be provided by the device manufacturer in the form of inherent performance parameters.

[0134] Step 411: Obtain the emergency stop duration required from when the emergency stop instruction is sent until the end of the robotic arm of the first surgical navigation device stops moving. For ease of description, it is denoted as the first emergency stop duration.

[0135] In an embodiment of the present invention, the implementation method of obtaining the emergency stop duration required from when the emergency stop instruction is sent until the end of the robotic arm of the first surgical navigation device stops moving in this step 411 may include: Record the first time point when the emergency stop instruction is issued, use an encoder or an IMU (Inertial Measurement Unit) to detect the second time point when the velocity of the robotic arm drops to zero, and use the duration from the first time point to the second time point as the first emergency stop duration;

[0136] In another embodiment of the present invention, the implementation method of obtaining the emergency stop duration required from when the emergency stop instruction is sent until the end of the robotic arm of the first surgical navigation device stops moving in this step 411 may include: Divide the first initial velocity by the deceleration a of the motor braking, so as to obtain the first emergency stop duration required from when the emergency stop instruction is sent until the end of the robotic arm of the first surgical navigation device stops moving.

[0137] Step 413: Calculate the theoretical emergency stop displacement corresponding to the robotic arm of the first surgical navigation device.

[0138] In an embodiment of the present invention, calculating the theoretical emergency stop displacement of the robotic arm corresponding to the first surgical navigation device in step 413 includes:

[0139]

[0140] x theory is the theoretical emergency stop displacement of the robotic arm of the first surgical navigation device; v0 is the first initial velocity, t is the first emergency stop duration, and a is the deceleration of the motor braking.

[0141] In an embodiment of the present invention, the deceleration a of the motor braking can be the motor braking performance parameter provided by the manufacturer of the first surgical navigation device and does not need to be measured in real time.

[0142] In another embodiment of the present invention, the deceleration a of the motor braking can also be obtained by calculation, that is, first calculate the product of the braking current I of the motor driving the robotic arm when the emergency stop is triggered and the braking coefficient, and then divide the product by the mass of the robotic arm to obtain the deceleration a of the motor braking. Among them, the braking coefficient is a parameter provided by the equipment manufacturer, indicating the proportional relationship between the braking force and the braking current.

[0143] In an embodiment of the present invention, a training device for an AI model for safety detection is proposed. Refer to Figure 5 , and this AI model is applied to detect the safety of the emergency stop displacement of the robotic arm of the surgical navigation device; the device includes:

[0144] A training sample acquisition module 501, configured to acquire multiple groups of training samples; wherein, each group of training samples includes: the initial velocity of the end of the robotic arm of the surgical navigation device when the emergency stop is triggered, the braking current of the robotic arm motor of the surgical navigation device when the emergency stop is triggered, the duration required for the end of the robotic arm to make an emergency stop, and the actual value of the displacement deviation as the sample label;

[0145] A training execution module 502, configured to input each group of training samples into the AI model for safety detection to be trained, and respectively obtain the displacement deviation prediction value for each group of training samples output by the AI model for safety detection; according to the difference between the actual value of the displacement deviation corresponding to each group of training samples and the displacement deviation prediction value, adjust the parameters of the AI model for safety detection so that the AI model learns to predict the deviation between the actual emergency stop displacement and the theoretical emergency stop displacement;

[0146] An AI model generation module 503, configured to obtain the AI model for safety detection trained by the training execution module 502.

[0147] In one embodiment of the present invention, a safety detection system for the robotic arm of a surgical navigation device based on robotic technology is proposed. Refer to Figure 6 , the system includes:

[0148] An emergency stop command transmission module 601, configured to generate an emergency stop command; send the emergency stop command to the first surgical navigation device based on robotic technology to be currently detected, so that the first surgical navigation device triggers the motor driver of the robotic arm to brake the robotic arm;

[0149] An initial speed acquisition module 602, configured to obtain the first initial speed of the end of the robotic arm of the first surgical navigation device when the emergency stop command is sent;

[0150] A braking current acquisition module 603, configured to obtain the first braking current of the robotic arm motor of the first surgical navigation device when the emergency stop command is sent;

[0151] An emergency stop duration acquisition module 604, configured to obtain the first emergency stop duration required for the end of the robotic arm of the first surgical navigation device to stop moving after the emergency stop command is sent;

[0152] A theoretical emergency stop displacement calculation module 605, configured to calculate the theoretical emergency stop displacement corresponding to the robotic arm of the first surgical navigation device;

[0153] An AI model processing module 605, configured to input the first initial speed, the first braking current, and the first emergency stop duration into a pre-trained AI model for safety detection; wherein, this AI model is trained by the training device of the embodiment of the present invention shown in Figure 5 ; obtain the displacement deviation corresponding to the robotic arm of the first surgical navigation device output by the AI model for safety detection;

[0154] An emergency stop displacement evaluation module 606, configured to add the theoretical emergency stop displacement corresponding to the robotic arm of the first surgical navigation device to the displacement deviation corresponding to the robotic arm of the first surgical navigation device, and compare the calculated sum value with a pre-set emergency stop displacement threshold; if the sum value is less than the pre-set emergency stop displacement threshold, it is determined that the emergency stop displacement of the robotic arm of the first surgical navigation device based on robotic technology meets the safety requirements; if the sum value is not less than the pre-set emergency stop displacement threshold, it is determined that the emergency stop displacement of the robotic arm of the first surgical navigation device based on robotic technology does not meet the safety requirements.

[0155] In Figure 6 In one embodiment of the system of the present invention shown, the emergency stop command transmission module 601 is configured to execute: simulate the emergency stop command through a step signal.

[0156] In Figure 6In an embodiment of the system of the present invention as shown, the initial velocity acquisition module 602 is configured to perform: a controller that controls the robotic arm of the first surgical navigation device receives the emergency stop instruction, and uses the current velocity of the end of the robotic arm of the first surgical navigation device monitored by the controller when the controller receives the emergency stop instruction as the first initial velocity.

[0157] In Figure 6 In an embodiment of the system of the present invention as shown, the emergency stop duration acquisition module 604 is configured to perform:

[0158] Record the first time point when the emergency stop instruction is issued, use an encoder or an IMU (inertial measurement unit) to detect the second time point when the velocity of the robotic arm drops to zero, and use the duration from the first time point to the second time point as the first emergency stop duration;

[0159] Or,

[0160] Divide the first initial velocity by the deceleration of the motor braking, so as to obtain the first emergency stop duration required for the end of the robotic arm of the first surgical navigation device to stop moving after the emergency stop instruction is sent.

[0161] In Figure 6 In an embodiment of the system of the present invention as shown, the theoretical emergency stop displacement calculation module 605 is configured to perform:

[0162]

[0163] x theory is the theoretical emergency stop displacement of the robotic arm of the first surgical navigation device; v0 is the first initial velocity, t is the first emergency stop duration, and a is the deceleration of the motor braking.

[0164] An embodiment of the present invention provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed on a computer, the computer is made to execute the method in any one of the embodiments in the specification.

[0165] An embodiment of the present invention provides a computing device, including a memory and a processor. An executable code is stored in the memory. When the processor executes the executable code, the method in any one of the embodiments in the specification is implemented.

[0166] It can be understood that the structure schematically shown in the embodiments of the present invention does not constitute a specific limitation on the device of the embodiments of the present invention. In other embodiments of the specification, the above device may include more or fewer components than shown in the figure, or combine certain components, or split certain components, or have different component arrangements. The components shown in the figure can be implemented in hardware, software, or a combination of software and hardware.

[0167] Each embodiment in the present invention is described in a progressive manner. For the same or similar parts among the embodiments, reference can be made to each other. Each embodiment focuses on the differences from other embodiments. In particular, for the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple. For the relevant parts, reference can be made to the description of the method embodiment.

[0168] Those skilled in the art should be able to realize that in one or more of the above examples, the functions described in the present invention can be implemented by hardware, software, add-ons, or any combination thereof. When implemented using software, these functions can be stored in a computer-readable medium or transmitted as one or more instructions or codes on a computer-readable medium.

[0169] The specific embodiments described above further elaborate on the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above is only the specific embodiment of the present invention and is not used to limit the protection scope of the present invention. Any modifications, equivalent replacements, improvements, etc. made on the basis of the technical solution of the present invention should be included in the protection scope of the present invention.

Claims

1. A training method for an AI model for security detection, characterized in that: The AI ​​model is used to detect the safety of the emergency stop displacement of the robotic arm of the surgical navigation device; the method includes: Obtain multiple sets of training samples; Each set of training samples includes: the initial speed of the end of the robotic arm of the surgical navigation device when the emergency stop is triggered, the braking current of the robotic arm motor of the surgical navigation device when the emergency stop is triggered, the time required for the robotic arm end to stop urgently, and the actual value of the displacement deviation as the sample label; Input each group of training samples into the AI ​​model to be trained for safety detection, and obtain the displacement deviation prediction value for each group of training samples output by the AI ​​model for safety detection; According to the difference between the actual displacement deviation value and the predicted displacement deviation value corresponding to each set of training samples, the parameters of the AI ​​model used for safety detection are adjusted so that the AI ​​model learns to predict the deviation between the actual emergency stop displacement and the theoretical emergency stop displacement; Get the trained AI model for security testing.

2. The method according to claim 1, characterized in that The actual displacement deviation value indicates a deviation between an actual emergency stop displacement and a theoretical emergency stop displacement caused by the movement of the robot arm of the surgical navigation device.

3. The method according to claim 1, characterized in that The formula of the AI ​​model for security detection is expressed as: △x=w1·υ0+w2·Ⅰ+w3·t+b w1, w2, w3: weights (parameters that the model needs to learn) b: Bias term Δx: predicted value of displacement deviation; v0: initial velocity of the end of the robotic arm of the surgical navigation device when the emergency stop is triggered; I: Braking current of the robotic arm motor of the surgical navigation device when the emergency stop is triggered; t: The time required for the end of the robot to stop suddenly.

4. A safety detection method for a robotic arm of a surgical navigation device based on robotics technology, characterized in that: The method includes: Generate an emergency stop command; Sending an emergency stop command to a first surgical navigation device based on robotics technology to be tested; The first surgical navigation device triggers the motor of the robotic arm to brake the robotic arm; Obtaining a first initial speed of the end of the robotic arm of the first surgical navigation device when the emergency stop command is sent; Obtaining a first braking current of a motor of a manipulator arm of a first surgical navigation device when an emergency stop command is sent; Obtaining a first emergency stop time required from when the emergency stop command is sent until the end of the robot arm of the first surgical navigation device stops moving; Calculating a theoretical emergency stop displacement corresponding to the robot arm of the first surgical navigation device; Inputting the first initial speed, the first braking current, and the first emergency stop duration into a pre-trained AI model for safety detection; wherein the AI ​​model is trained using the method described in any one of claims 1 to 3; Obtaining a displacement deviation of a robotic arm of the first surgical navigation device corresponding to an output of the AI ​​model for safety detection; adding a theoretical emergency stop displacement of the robot arm corresponding to the first surgical navigation device and a displacement deviation of the robot arm corresponding to the first surgical navigation device, and comparing the calculated sum with a preset emergency stop displacement threshold; If the sum is less than a preset emergency stop displacement threshold, it is determined that the emergency stop displacement of the robot arm of the first surgical navigation device based on robotics meets the safety requirements; If the sum is not less than a preset emergency stop displacement threshold, it is determined that the emergency stop displacement of the robot arm of the first surgical navigation device based on robotics does not meet the safety requirements.

5. The method according to claim 4, characterized in that The generating of the emergency stop command comprises: simulating the emergency stop command through a step signal.

6. The method according to claim 4, characterized in that The obtaining of the initial speed of the end of the robotic arm of the first surgical navigation device when the emergency stop command is sent includes: a controller that controls the robotic arm of the first surgical navigation device receives the emergency stop command, and uses the current speed of the end of the robotic arm of the first surgical navigation device monitored by the controller when the controller receives the emergency stop command as the first initial speed.

7. The method according to claim 4, characterized in that The obtaining of the emergency stop time required from when the emergency stop instruction is sent until the end of the robot arm of the first surgical navigation device stops moving includes: Record the first time point when the emergency stop command is issued, use an encoder or IMU (inertial measurement unit) to detect the second time point when the speed of the robotic arm drops to zero, and use the duration from the first time point to the second time point as the first emergency stop duration; or, The first initial speed is divided by the deceleration of the motor brake to obtain the first emergency stop time required from the time the emergency stop instruction is sent until the end of the robot arm of the first surgical navigation device stops moving.

8. The method according to claim 4, characterized in that The calculating corresponds to a theoretical emergency stop displacement of the robot arm of the first surgical navigation device, comprising: x theory is the theoretical emergency stop displacement of the robotic arm of the first surgical navigation device; v0 is the first initial speed, t is the first emergency stop duration, and a is the deceleration of the motor braking.

9. A training device for an AI model for security detection, characterized in that: The AI ​​model is used to detect the safety of the emergency stop displacement of the robotic arm of the surgical navigation device; the device includes: The training sample acquisition module is configured to acquire multiple sets of training samples; wherein each set of training samples includes: the initial speed of the end of the robotic arm of the surgical navigation device when the emergency stop is triggered, the braking current of the robotic arm motor of the surgical navigation device when the emergency stop is triggered, the time required for the robotic arm end to stop urgently, and the actual value of the displacement deviation as a sample label; A training execution module is configured to input each set of training samples into the AI ​​model for safety detection to be trained, and obtain the displacement deviation prediction value for each set of training samples output by the AI ​​model for safety detection; according to the difference between the actual displacement deviation value and the displacement deviation prediction value corresponding to each set of training samples, adjust the parameters of the AI ​​model for safety detection so that the AI ​​model learns to predict the deviation between the actual emergency stop displacement and the theoretical emergency stop displacement; The AI ​​model generation module is configured to obtain a trained AI model for security testing.

10. A safety detection system for the robotic arm of a surgical navigation device based on robotic technology, characterized in that: The method includes: The emergency stop instruction transmission module is configured to generate an emergency stop instruction; send the emergency stop instruction to the first surgical navigation device based on robotic technology to be detected, so that the first surgical navigation device triggers the motor of the robotic arm to brake the robotic arm; An initial speed acquisition module is configured to obtain a first initial speed of the end of the robotic arm of the first surgical navigation device when the emergency stop instruction is sent; A braking current acquisition module, configured to obtain a first braking current of a motor driving a first surgical navigation device mechanical arm when an emergency stop command is sent; An emergency stop duration acquisition module is configured to obtain a first emergency stop duration required from when the emergency stop instruction is sent until the end of the robot arm of the first surgical navigation device stops moving; a theoretical emergency stop displacement calculation module, configured to calculate a theoretical emergency stop displacement corresponding to the robot arm of the first surgical navigation device; The AI ​​model processing module is configured to input the first initial speed, the first braking current, and the first emergency stop duration into a pre-trained AI model for safety detection; wherein the AI ​​model is trained using the device described in claim 9; and obtain the displacement deviation of the robotic arm of the first surgical navigation device output by the AI ​​model for safety detection; The emergency stop displacement evaluation module is configured to add the theoretical emergency stop displacement of the robotic arm corresponding to the first surgical navigation device and the displacement deviation of the robotic arm corresponding to the first surgical navigation device, and compare the calculated sum with a preset emergency stop displacement threshold; if the sum is less than the preset emergency stop displacement threshold, it is determined that the emergency stop displacement of the robotic arm of the first surgical navigation device based on robotic technology meets the safety requirements; if the sum is not less than the preset emergency stop displacement threshold, it is determined that the emergency stop displacement of the robotic arm of the first surgical navigation device based on robotic technology does not meet the safety requirements.

Images