Surgical robot target tracking compensation control method and system in dynamic breathing environment

Abstract

The invention provides a target tracking compensation control method and system for a surgical robot in a dynamic breathing environment, and the method comprises the steps: collecting an ultrasonic image of a patient in real time, and carrying out the self-adaptive admittance control algorithm based on force feedback; force control output is adjusted in a self-adaptive mode according to the contact force between the ultrasonic probe and the skin surface and the change of the displacement and speed state of the ultrasonic probe, rigidity parameters of an admittance control model are adjusted in real time, constant contact force between the ultrasonic probe and the skin surface is kept, and it is ensured that the ultrasonic probe and the body surface keep stable contact force and meanwhile adapt to the dynamic change of the body surface; a focus is continuously identified and positioned by utilizing a real-time ultrasonic image through template matching and a dynamic template updating algorithm, real-time position information of an internal focus target spot is obtained and serves as a target pose, the motion compensation amount of a puncture mechanical arm is calculated, the track of a puncture needle point and the motion of the target spot are kept synchronous, and the target pose is accurately positioned. And target tracking and compensation control in a dynamic breathing environment are realized. Target tracking compensation control of the surgical robot can be carried out.

Description

Technical Field

[0001] This invention relates to the fields of surgical robot control and medical image navigation technology, and in particular to a method and system for target tracking compensation control of surgical robots in dynamic respiratory environments. Background Technology

[0002] One of the biggest challenges in minimally invasive thoracotomy and abdominal puncture surgery is the displacement of the target site caused by the patient's respiratory movements. This physiological movement not only reduces puncture accuracy and increases surgical risks, but also places higher demands on robot-assisted puncture systems.

[0003] Existing methods primarily focus on punctures in static or quasi-static environments, with insufficient research on real-time tracking and compensation control in dynamic environments. Compared to static punctures, robotic punctures in dynamic environments require real-time sensing, tracking, and compensation of target displacement to ensure surgical safety and accuracy. How to achieve real-time response while maintaining puncture accuracy under the influence of respiratory motion is a major challenge currently facing technology. Summary of the Invention

[0004] To address the technical problems existing in the prior art, the present invention provides a method and system for target tracking compensation control of a surgical robot under dynamic respiratory conditions, the technical solution of which is as follows: On the one hand, a target tracking compensation control method for a surgical robot under dynamic respiratory conditions is provided. The surgical robot has two robotic arms: an ultrasound robotic arm holding an ultrasound probe to acquire real-time ultrasound images of the patient to obtain the location information of internal lesion targets; and a puncture robotic arm holding a puncture needle to perform puncture surgery under ultrasound image guidance. The method includes: S1. Real-time acquisition of patient ultrasound images. During the real-time acquisition of patient ultrasound images, an adaptive admittance control algorithm based on force feedback is used to adaptively adjust the force control output according to the changes in the contact force between the ultrasound probe and the skin surface, as well as the displacement and velocity state of the ultrasound probe. The stiffness parameters of the admittance control model are adjusted in real time to maintain a constant contact force between the ultrasound probe and the skin surface, preventing probe drift or tissue deformation under pressure. This ensures that the ultrasound probe maintains a stable contact force with the body surface while adapting to dynamic changes in the body surface, thereby improving image quality. S2. Using real-time ultrasound images, and through template matching and dynamic template updating algorithms, lesions are continuously identified and located to obtain real-time internal lesion target location information. S3. Using the target position information as the target pose, calculate the motion compensation amount of the puncture robotic arm to keep the puncture needle tip trajectory synchronized with the target movement, thereby realizing target tracking and compensation control under dynamic breathing conditions.

[0005] Optionally, the force feedback-based adaptive admittance control algorithm in S1 specifically includes: A six-dimensional force sensor installed at the end of the ultrasonic robotic arm can acquire the contact force between the ultrasonic probe and the skin surface in real time. ; Calculate the current contact force error: in, yes The estimated value; This refers to the actual position of the ultrasonic probe; It is an environmental stiffness parameter; It is an estimated value of the environmental stiffness parameter; This is the desired position of the ultrasound probe; It is an estimate of the desired position of the ultrasonic probe; The contact force error is input into the admittance control model: in, It is the first Contact force error in the next iteration , , These represent the inertial parameter matrix, damping parameter matrix, and stiffness parameter matrix of the admittance control model, respectively. Determine the correction amount of the ultrasonic probe Through iteration Approaching 0, thus making Approaching 0, so that the probe maintains a stable contact force despite body surface undulations and respiratory movements; stiffness parameters Treating the parameters to be estimated, a Lyapunov function is established based on the error. The online update law is derived through derivation. To ensure that the parameter estimation error converges; Stiffness parameters are adjusted in real time according to the update law. This allows for dynamic adjustment of the admittance control model, enabling the ultrasound probe to maintain the desired constant contact force under different soft tissue contact conditions.

[0006] Optionally, S2 specifically includes: S21. In the initial stage of the ultrasound-guided puncture procedure, the physician manually selects the lesion area of ​​interest on the ultrasound image as the initial template. The initial template Including characteristic information of the lesions; S22. Real-time acquisition of image sequences And preprocess each frame of image; S23. Process the preprocessed image sequence With the current template Based on this, a deformation-tolerant template including scaling and rotation is used for normalized cross-correlation NCC matching to obtain the spatial similarity matrix of each candidate location, and then the gray-level similarity matrix is ​​obtained to determine the lesion region and target location. S24. Merge the identified lesion areas with the current template according to their weights to obtain the updated template. By allowing the template to rotate and scale within a limited range, it can adapt to the deformation of lesions in ultrasound imaging, thereby improving the robustness of tracking.

[0007] Optionally, S23 specifically includes: For the current template By applying a scale factor to it and rotation angle Generate deformation-tolerant template ; For each deformation-tolerant template With the current image Calculate spatial similarity: in, Represents the pixel coordinates inside the template. Indicates the current template After applying deformation according to the scaling factor s and rotation angle θ, the coordinates inside the deformation template The grayscale value at that location; This represents the average grayscale value of all pixels within the template after deformation. This indicates that image I is at the current matching position. Below, corresponding to the pixels inside the template Pixel values; This represents the average gray level of all pixels in image I; Finally, the maximum correlation value among all scale and angle combinations is taken. As a matching result: Maximum correlation value corresponding and Apply to the current template This yields a new template image T. The new template image T and the current image are then compared. Calculate the grayscale similarity between the two: in, and This represents the top-left corner coordinates of the current matching position of the template in image I. and Represents the pixel coordinates inside the template. Indicates the new template image T in grayscale value, This indicates that image I is at the current matching position. Below, corresponding to the pixels inside the template pixel values, This indicates that the new template image T and image I are at the same position. The matching degree at the location ranges from [-1, 1], and the closer the value is to 1, the higher the matching degree. Regions with a matching degree greater than a preset threshold are designated as lesion regions, and the location with the highest matching degree is designated as the target location.

[0008] Optionally, S3 specifically includes: S31. Represent the target pose as... The rotation angles of each joint of the puncture robot arm were solved using inverse kinematics of the robot arm. ; S32. Obtain the current end-effector pose using forward kinematics. and calculate and The pose error is calculated to obtain the joint compensation amount. Compensation amount of each joint constitute : S33. Update the joint angles as follows: Repeat steps S32 and S33, iterating until the convergence condition is met, resulting in the end-effector pose. Approaching the target pose infinitely .

[0009] On the other hand, a target tracking compensation control system for a surgical robot under dynamic respiratory conditions is provided. The surgical robot has two robotic arms: an ultrasound robotic arm holding an ultrasound probe to acquire real-time ultrasound images of the patient to obtain the location information of internal lesion targets; and a puncture robotic arm holding a puncture needle to perform puncture surgery under ultrasound image guidance. The system includes: The admittance control module is used to acquire patient ultrasound images in real time. During the real-time acquisition of patient ultrasound images, the adaptive admittance control algorithm based on force feedback adaptively adjusts the force control output according to the changes in the contact force between the ultrasound probe and the skin surface, as well as the displacement and velocity state of the ultrasound probe. It adjusts the stiffness parameters of the admittance control model in real time to maintain a constant contact force between the ultrasound probe and the skin surface, preventing probe drift or tissue deformation due to pressure. This ensures that the ultrasound probe maintains a stable contact force with the body surface while adapting to dynamic changes in the body surface, thereby improving image quality. The identification and localization module is used to continuously identify and locate lesions using real-time ultrasound images and through template matching and dynamic template update algorithms, thereby obtaining real-time internal lesion target location information. The tracking compensation control module is used to take the target position information as the target pose, calculate the motion compensation amount of the puncture robot arm, so that the trajectory of the puncture needle tip is synchronized with the movement of the target, and realize target tracking and compensation control in dynamic breathing environment.

[0010] Optionally, the admittance control module is specifically used for: A six-dimensional force sensor installed at the end of the ultrasonic robotic arm can acquire the contact force between the ultrasonic probe and the skin surface in real time. ; Calculate the current contact force error: in, yes The estimated value; This refers to the actual position of the ultrasonic probe; It is an environmental stiffness parameter; It is an estimated value of the environmental stiffness parameter; This is the desired position of the ultrasound probe; It is an estimate of the desired position of the ultrasonic probe; The contact force error is input into the admittance control model: in, It is the first Contact force error in the next iteration , , These represent the inertial parameter matrix, damping parameter matrix, and stiffness parameter matrix of the admittance control model, respectively. Determine the correction amount of the ultrasonic probe Through iteration Approaching 0, thus making Approaching 0, so that the probe maintains a stable contact force despite body surface undulations and respiratory movements; stiffness parameters Treating the parameters to be estimated, a Lyapunov function is established based on the error. The online update law is derived through derivation. To ensure that the parameter estimation error converges; Stiffness parameters are adjusted in real time according to the update law. This allows for dynamic adjustment of the admittance control model, enabling the ultrasound probe to maintain the desired constant contact force under different soft tissue contact conditions.

[0011] Optionally, the identification and positioning module is specifically used for: S21. In the initial stage of the ultrasound-guided puncture procedure, the physician manually selects the lesion area of ​​interest on the ultrasound image as the initial template. The initial template Including characteristic information of the lesions; S22. Real-time acquisition of image sequences And preprocess each frame of image; S23. Process the preprocessed image sequence With the current template Based on this, a deformation-tolerant template including scaling and rotation is used for normalized cross-correlation NCC matching to obtain the spatial similarity matrix of each candidate location, and then the gray-level similarity matrix is ​​obtained to determine the lesion region and target location. S24. Merge the identified lesion areas with the current template according to their weights to obtain the updated template. By allowing the template to rotate and scale within a limited range, it can adapt to the deformation of lesions in ultrasound imaging, thereby improving the robustness of tracking.

[0012] Optionally, S23 specifically includes: For the current template By applying a scale factor to it and rotation angle Generate deformation-tolerant template ; For each deformation-tolerant template With the current image Calculate spatial similarity: in, Represents the pixel coordinates inside the template. Indicates the current template After applying deformation according to the scaling factor s and rotation angle θ, the coordinates inside the deformation template The grayscale value at that location; This represents the average grayscale value of all pixels within the template after deformation. This indicates that image I is at the current matching position. Below, corresponding to the pixels inside the template Pixel values; This represents the average gray level of all pixels in image I; Finally, the maximum correlation value among all scale and angle combinations is taken. As a matching result: Maximum correlation value corresponding and Apply to the current template This yields a new template image T. The new template image T and the current image are then compared. Calculate the grayscale similarity between the two: in, and This represents the top-left corner coordinates of the current matching position of the template in image I. and Represents the pixel coordinates inside the template. Indicates the new template image T in grayscale value, This indicates that image I is at the current matching position. Below, corresponding to the pixels inside the template pixel values, This indicates that the new template image T and image I are at the same position. The matching degree at the location ranges from [-1, 1], and the closer the value is to 1, the higher the matching degree. Regions with a matching degree greater than a preset threshold are designated as lesion regions, and the location with the highest matching degree is designated as the target location.

[0013] Optionally, the tracking compensation control module is specifically used for: S31. Represent the target pose as... The rotation angles of each joint of the puncture robot arm were solved using inverse kinematics of the robot arm. ; S32. Obtain the current end-effector pose using forward kinematics. and calculate and The pose error is calculated to obtain the joint compensation amount. Compensation amount of each joint constitute : S33. Update the joint angles as follows: Repeat steps S32 and S33, iterating until the convergence condition is met, resulting in the end-effector pose. Approaching the target pose infinitely .

[0014] On the other hand, an electronic device is provided, comprising a processor and a memory, wherein the memory stores at least one instruction, which is loaded and executed by the processor to implement the above-described target tracking compensation control method for surgical robots under dynamic respiratory conditions.

[0015] On the other hand, a computer-readable storage medium is provided, wherein at least one instruction is stored in the storage medium, the at least one instruction being loaded and executed by a processor to implement the above-described surgical robot target tracking compensation control method under dynamic respiratory environment.

[0016] The beneficial effects of the technical solution provided by this invention include at least the following: In dynamic surgical environments where lesions shift due to respiratory movements, this invention demonstrates higher real-time response capabilities and motion-following accuracy, resulting in a higher overall puncture success rate and effectively ensuring operational stability and safety. Specifically: (1) This invention proposes a target location tracking method under real-time ultrasound images. It uses real-time ultrasound images to obtain internal target location information and continuously identifies and locates lesions through template matching and dynamic template updating algorithms. This improves the identification stability and robustness under tissue deformation and noise interference caused by respiration, and realizes real-time detection of moving targets.

[0017] (2) The present invention designs a dynamic compensation control strategy. It obtains the patient’s body surface motion information through an optical positioning system, constructs a motion compensation control model using visual navigation technology, and estimates the lesion displacement and the mechanical arm motion compensation in real time to keep the puncture needle tip trajectory synchronized with the lesion movement, thereby realizing target tracking and compensation control in a dynamic respiratory environment.

[0018] (3) This invention studies the admittance control algorithm of ultrasonic robotic arm. In view of the problem of unstable contact force between ultrasonic probe and body surface, an adaptive admittance control algorithm is designed. The force control output is adaptively adjusted according to the changes in contact force, displacement and velocity state, and the environmental stiffness parameter is adjusted in real time to keep the probe and skin surface constant contact pressure, prevent probe drift or tissue deformation under pressure, ensure stable contact between ultrasonic probe and body surface, and improve image quality. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a flowchart of a target tracking compensation control method for a surgical robot under dynamic respiratory conditions, provided by an embodiment of the present invention. Figure 2 This is a template matching and recognition effect diagram provided in an embodiment of the present invention; Figure 3This is a flowchart of the error compensation process provided in an embodiment of the present invention; Figure 4 This is a block diagram of a target tracking compensation control system for a surgical robot under dynamic respiratory conditions, provided by an embodiment of the present invention. Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0021] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0022] This invention addresses the issues of lesion displacement and tissue deformation caused by respiratory motion in ultrasound-guided thoracic and abdominal puncture surgeries. It proposes a comprehensive solution based on real-time ultrasound imaging, visual navigation and adaptive admittance control, and motion compensation control. The core idea is to unify respiratory motion modeling, target dynamic identification, motion compensation control, and force-controlled stable imaging within a single robot control framework. This enables synchronous response and precise puncture by the robot during the respiratory cycle. Specifically, real-time ultrasound images are used to acquire the true location of the target within the body, optical navigation provides external spatial reference, and control algorithms complete dynamic trajectory planning and compensation. This achieves closed-loop control from target identification and position calculation to robotic arm movements, maintaining consistency between the needle tip trajectory and target displacement during continuous motion caused by the patient's breathing, thereby significantly improving puncture accuracy and real-time performance.

[0023] like Figure 1 As shown, this embodiment of the invention provides a target tracking compensation control method for a surgical robot in a dynamic respiratory environment. The surgical robot has two robotic arms: an ultrasound robotic arm holding an ultrasound probe to acquire real-time ultrasound images of the patient to obtain the location information of internal lesion targets; and a puncture robotic arm holding a puncture needle to perform puncture surgery under ultrasound image guidance. The method includes: S1. Real-time acquisition of patient ultrasound images. During the real-time acquisition of patient ultrasound images, an adaptive admittance control algorithm based on force feedback is used to adaptively adjust the force control output according to the changes in the contact force between the ultrasound probe and the skin surface, as well as the displacement and velocity state of the ultrasound probe. The stiffness parameters of the admittance control model are adjusted in real time to maintain a constant contact force between the ultrasound probe and the skin surface, preventing probe drift or tissue deformation under pressure. This ensures that the ultrasound probe maintains a stable contact force with the body surface while adapting to dynamic changes in the body surface, thereby improving image quality. Contact between the ultrasound probe and the skin surface is a prerequisite for acquiring high-quality ultrasound images. However, the undulations of the body surface caused by respiratory movements can affect the contact force between the probe and the skin, thus affecting image quality. To address this issue, this invention presents a force feedback-based admittance control algorithm that ensures a stable contact force between the ultrasound probe and the body surface while adapting to dynamic changes in the body surface.

[0024] Admittance control, by simulating a physical spring-damping system, endows a robotic arm with controllable compliance, enabling it to adapt to changes in the shape of the human body surface while maintaining stable contact force. The basic idea of ​​admittance control is to acquire external force using a six-dimensional force sensor at the end effector, compare it with the desired force and desired position, calculate the positional offset requiring real-time compensation based on the admittance control model, and superimpose it onto the desired position to control the ultrasound probe's offset. This process is repeated until the desired pose and desired contact force are met. Traditional admittance control algorithms typically require manually setting fixed stiffness parameters to control the interaction force between the ultrasound probe and the human (or phantom) surface during ultrasound probe scanning. This fixed-parameter design has the following limitations: the surface characteristics of the human body or phantom are complex and highly variable; different parts have different hardness, curvature, and coefficient of friction, making it difficult for fixed parameters to maintain ideal contact force and tracking performance in all scanning areas. Dynamic environmental changes, such as patient breathing and minor body movements, can cause the surface state to change constantly, making it difficult for fixed parameters to adapt in real time. It is also difficult to balance scanning quality and safety: excessively large parameters may cause the probe to compress tissue, while excessively small parameters may cause the probe to detach from the skin, affecting image quality.

[0025] The adaptive admittance control algorithm in this embodiment of the invention dynamically adjusts the admittance parameters by sensing the contact force, velocity and other state information between the probe and the surface in real time. It can adjust the control parameters in real time according to the current contact force and motion state, so that the probe always maintains a suitable contact force with the surface, thereby improving image quality and security.

[0026] Optionally, the force feedback-based adaptive admittance control algorithm in S1 specifically includes: A six-dimensional force sensor installed at the end of the ultrasonic robotic arm can acquire the contact force between the ultrasonic probe and the skin surface in real time. ; Calculate the current contact force error: in, yes The estimated value; This refers to the actual position of the ultrasonic probe; It is an environmental stiffness parameter; It is an estimated value of the environmental stiffness parameter; This is the desired position of the ultrasound probe; It is an estimate of the desired position of the ultrasonic probe; The contact force error is input into the admittance control model: in, It is the first Contact force error in the next iteration , , These represent the inertial parameter matrix, damping parameter matrix, and stiffness parameter matrix of the admittance control model, respectively. Determine the correction amount of the ultrasonic probe Through iteration Approaching 0, thus making Approaching 0, so that the probe maintains a stable contact force despite body surface undulations and respiratory movements; stiffness parameters Treating the parameters to be estimated, a Lyapunov function is established based on the error. The online update law is derived through derivation. To ensure that the parameter estimation error converges; Stiffness parameters are adjusted in real time according to the update law. This allows for dynamic adjustment of the admittance control model, enabling the ultrasound probe to maintain the desired constant contact force under different soft tissue contact conditions.

[0027] Furthermore, force sensors are susceptible to various factors during actual operation, such as environmental noise, electromagnetic interference, and hardware defects, resulting in high-frequency noise and interference components in the acquired signals. This noise can affect the accuracy of subsequent data processing and analysis. Therefore, this embodiment of the invention also performs filtering on the force sensor signals to remove useless interference, extract true mechanical information, and improve the reliability and stability of the measurement results. Specifically, this embodiment of the invention uses Kalman filtering to filter the acquired force sensor signals, reducing noise interference and making the signal smoother and more stable.

[0028] S2. Using real-time ultrasound images, and through template matching and dynamic template updating algorithms, lesions are continuously identified and located to obtain real-time internal lesion target location information. This invention employs real-time ultrasound imaging to directly acquire precise location information of internal target points. The ultrasound probe maintains stable contact with the body surface through robotic arm admittance control, ensuring stable ultrasound image acquisition. Simultaneously, it identifies and locates lesion areas in the images in real time. The target point tracking method based on real-time ultrasound has significant advantages, enabling direct observation of the actual location of tissues within the body and avoiding the error accumulation problem present in prediction models. Tissue deformation caused by respiratory movements is complex and varies from person to person; relying solely on surface movement or respiratory signals to infer internal organ displacement often fails to achieve the clinically required accuracy. Specifically, firstly, image sequences are acquired in real-time using a high-frame-rate ultrasound imaging system; secondly, the original ultrasound images undergo preprocessing such as noise reduction and enhancement to improve image quality; then, the preprocessed images are identified, regional features are automatically extracted, and grayscale templates are generated. The template selection is based on manually confirmed regions from the initial calibration phase. Subsequently, correlation matching is automatically performed in consecutive frames to accurately locate lesion areas and target point positions, and the template is updated.

[0029] Embodiments of the present invention also include converting the target position in the ultrasound image to the world coordinate system through calibration relationships to obtain the three-dimensional spatial position information of the target.

[0030] Optionally, S2 specifically includes: S21. In the initial stage of the ultrasound-guided puncture procedure, the physician manually selects the lesion area of ​​interest on the ultrasound image as the initial template. The initial template This includes characteristic information about the lesion (such as shape, texture, and edge features); S22. Real-time acquisition of image sequences And preprocess each frame of image; S23. Process the preprocessed image sequence With the current template Based on this, a deformation-tolerant template, including scaling and rotation (to improve computational efficiency, this embodiment of the invention employs a multi-scale search strategy, first performing coarse matching on lower-resolution images to determine the approximate location, and then performing fine matching in local areas of high-resolution images. Furthermore, considering lesion deformation caused by respiratory motion, this embodiment of the invention introduces a deformation-tolerant mechanism, allowing the template to rotate and scale within a certain range, enhancing the robustness of the algorithm), is used for normalized cross-correlation NCC matching to obtain the spatial similarity matrix of each candidate location, thereby obtaining the gray-level similarity matrix and determining the lesion region and target location;

[0031] S24. Merge the identified lesion areas with the current template according to their weights to obtain the updated template. By allowing the template to rotate and scale within a limited range, it can adapt to the deformation of lesions in ultrasound imaging, thereby improving the robustness of tracking.

[0032] During respiratory motion, the appearance of lesions in ultrasound images may change with probe position and tissue deformation. To address this challenge, this invention proposes a dynamic template update mechanism that adjusts the template in real time to adapt to minute changes in lesion appearance. This dynamic update strategy maintains the memory of the original template features while adapting to minor changes in lesion appearance, effectively solving the problem of lesion imaging characteristics changing during respiratory motion. The target recognition effect under real-time ultrasound is shown in the image below. Figure 2 As shown.

[0033] Optionally, S23 specifically includes: For the current template By applying a scale factor to it and rotation angle Generate deformation-tolerant template ; For each deformation-tolerant template With the current image Calculate spatial similarity: in, Represents the pixel coordinates inside the template. Indicates the current template After applying deformation according to the scaling factor s and rotation angle θ, the coordinates inside the deformation template The grayscale value at that location; This represents the average grayscale value of all pixels within the template after deformation. This indicates that image I is at the current matching position. Below, corresponding to the pixels inside the template Pixel values; This represents the average gray level of all pixels in image I; Finally, the maximum correlation value among all scale and angle combinations is taken. As a matching result: Maximum correlation value corresponding and Apply to the current template This yields a new template image T. The new template image T and the current image are then compared. Calculate the grayscale similarity between the two: in, and This represents the top-left corner coordinates of the current matching position of the template in image I. and Represents the pixel coordinates inside the template. Indicates the new template image T in grayscale value, This indicates that image I is at the current matching position. Below, corresponding to the pixels inside the template pixel values, This indicates that the new template image T and image I are at the same position. The matching degree at the location ranges from [-1, 1], and the closer the value is to 1, the higher the matching degree. Regions with a matching degree greater than a preset threshold (e.g., a preset threshold of 0.85) are designated as lesion regions, and the location with the highest matching degree is designated as the target location.

[0034] S3. Using the target position information as the target pose, calculate the motion compensation amount of the puncture robotic arm to keep the puncture needle tip trajectory synchronized with the target movement, thereby realizing target tracking and compensation control under dynamic breathing conditions.

[0035] The motion compensation control proposed in this embodiment of the invention uses the target position information obtained from real-time ultrasound images as input. By inputting the planned trajectory of the robotic arm end effector, the position and posture of the robotic arm end effector are adjusted in real time to achieve precise tracking of respiratory movements. The core of the compensation control is to calculate the difference between the target point and the current point, and to adjust the position and direction of the puncture needle through a servo controller.

[0036] The pose error is calculated based on the change in target position, and the posture of the robotic arm's end effector is continuously adjusted through a servo control mechanism. Within each control cycle, the robotic arm's motion path is replanned based on the deviation between the current position and the target position, so that the trajectory of the puncture needle tip remains consistent with the displacement of the target point.

[0037] Optionally, such as Figure 3 As shown, S3 specifically includes: S31. Represent the target pose as... The rotation angles of each joint of the puncture robot arm were solved using inverse kinematics of the robot arm. ; S32. Obtain the current end-effector pose using forward kinematics. and calculate and The pose error is calculated to obtain the joint compensation amount. Compensation amount of each joint constitute : S33. Update the joint angles as follows: Repeat steps S32 and S33, iterating until the convergence condition is met, resulting in the end-effector pose. Approaching the target pose infinitely .

[0038] To improve stability, this embodiment of the invention also introduces an adaptive gain adjustment mechanism, which dynamically adjusts the control gain according to the target's movement speed: in, Based on the gain, The attenuation coefficient is... The target's velocity is used as the reference point. When the target's velocity is high, the control gain is appropriately reduced to avoid system overshoot; conversely, when the target's velocity is low, the control gain is increased to enhance tracking accuracy.

[0039] To achieve more precise motion control, this embodiment of the invention also implements an appropriate filtering strategy to reduce the impact of measurement noise while preserving the basic characteristics of respiratory motion. Considering the requirements of real-time control and the noise characteristics in the data, this embodiment of the invention selects the Kalman filter algorithm for position data processing. The Kalman filter is used, and parameters are adjusted for the respiratory motion characteristics. The process noise covariance matrix Q characterizes the uncertainty of the system's dynamic characteristics, while the measurement noise covariance matrix R reflects the sensor's measurement accuracy. By reasonably setting these parameters, the filter effectively suppresses high-frequency noise interference while preserving the main characteristics of respiratory motion, significantly improving the smoothness of the processed trajectory.

[0040] like Figure 4 As shown, this embodiment of the invention also provides a target tracking compensation control system for a surgical robot under dynamic respiratory conditions. The surgical robot has two robotic arms: an ultrasound robotic arm holding an ultrasound probe to acquire real-time ultrasound images of the patient to obtain the location information of internal lesion targets; and a puncture robotic arm holding a puncture needle to perform puncture surgery under ultrasound image guidance. The system includes: The admittance control module 410 is used to acquire patient ultrasound images in real time. During the real-time acquisition of patient ultrasound images, the adaptive admittance control algorithm based on force feedback adaptively adjusts the force control output according to the changes in the contact force between the ultrasound probe and the skin surface, as well as the displacement and velocity state of the ultrasound probe. It adjusts the stiffness parameters of the admittance control model in real time to maintain a constant contact force between the ultrasound probe and the skin surface, prevents probe drift or tissue deformation under pressure, and ensures that the ultrasound probe maintains a stable contact force with the body surface while adapting to dynamic changes in the body surface, thereby improving image quality. The identification and positioning module 420 is used to continuously identify and locate lesions using real-time ultrasound images and through template matching and dynamic template updating algorithms, and to obtain real-time internal lesion target location information. The tracking compensation control module 430 is used to take the target position information as the target pose, calculate the motion compensation amount of the puncture robot arm, so that the trajectory of the puncture needle tip is synchronized with the target movement, and realize target tracking and compensation control in dynamic breathing environment.

[0041] Optionally, the admittance control module is specifically used for: A six-dimensional force sensor installed at the end of the ultrasonic robotic arm can acquire the contact force between the ultrasonic probe and the skin surface in real time. ; Calculate the current contact force error: in, yes The estimated value; This refers to the actual position of the ultrasonic probe; It is an environmental stiffness parameter; It is an estimated value of the environmental stiffness parameter; This is the desired position of the ultrasound probe; It is an estimate of the desired position of the ultrasonic probe; The contact force error is input into the admittance control model: in, It is the first Contact force error in the next iteration , , These represent the inertial parameter matrix, damping parameter matrix, and stiffness parameter matrix of the admittance control model, respectively. Determine the correction amount of the ultrasonic probe Through iteration Approaching 0, thus making Approaching 0, so that the probe maintains a stable contact force despite body surface undulations and respiratory movements; stiffness parameters Treating the parameters to be estimated, a Lyapunov function is established based on the error. The online update law is derived through derivation. To ensure that the parameter estimation error converges; Stiffness parameters are adjusted in real time according to the update law. This allows for dynamic adjustment of the admittance control model, enabling the ultrasound probe to maintain the desired constant contact force under different soft tissue contact conditions.

[0042] Optionally, the identification and positioning module is specifically used for: S21. In the initial stage of the ultrasound-guided puncture procedure, the physician manually selects the lesion area of ​​interest on the ultrasound image as the initial template. The initial template Including characteristic information of the lesions; S22. Real-time acquisition of image sequences And preprocess each frame of image; S23. Process the preprocessed image sequence With the current template Based on this, a deformation-tolerant template including scaling and rotation is used for normalized cross-correlation NCC matching to obtain the spatial similarity matrix of each candidate location, and then the gray-level similarity matrix is ​​obtained to determine the lesion region and target location. S24. Merge the identified lesion areas with the current template according to their weights to obtain the updated template. By allowing the template to rotate and scale within a limited range, it can adapt to the deformation of lesions in ultrasound imaging, thereby improving the robustness of tracking.

[0043] Optionally, S23 specifically includes: For the current template By applying a scale factor to it and rotation angle Generate deformation-tolerant template ; For each deformation-tolerant template With the current image Calculate spatial similarity: in, Represents the pixel coordinates inside the template. Indicates the current template After applying deformation according to the scaling factor s and rotation angle θ, the coordinates inside the deformation template The grayscale value at that location; This represents the average grayscale value of all pixels within the template after deformation. This indicates that image I is at the current matching position. Below, corresponding to the pixels inside the template Pixel values; This represents the average gray level of all pixels in image I; Finally, the maximum correlation value among all scale and angle combinations is taken. As a matching result: Maximum correlation value corresponding and Apply to the current template This yields a new template image T. The new template image T and the current image are then compared. Calculate the grayscale similarity between the two: in, and This represents the top-left corner coordinates of the current matching position of the template in image I. and Represents the pixel coordinates inside the template. Indicates the new template image T in grayscale value, This indicates that image I is at the current matching position. Below, corresponding to the pixels inside the template pixel values, This indicates that the new template image T and image I are at the same position. The matching degree at the location ranges from [-1, 1], and the closer the value is to 1, the higher the matching degree. Regions with a matching degree greater than a preset threshold are designated as lesion regions, and the location with the highest matching degree is designated as the target location.

[0044] Optionally, the tracking compensation control module is specifically used for: S31. Represent the target pose as... The rotation angles of each joint of the puncture robot arm were solved using inverse kinematics of the robot arm. ; S32. Obtain the current end-effector pose using forward kinematics. and calculate and The pose error is calculated to obtain the joint compensation amount. Compensation amount of each joint constitute : S33. Update the joint angles as follows: Repeat steps S32 and S33, iterating until the convergence condition is met, resulting in the end-effector pose. Approaching the target pose infinitely .

[0045] The target tracking compensation control system for surgical robots in a dynamic breathing environment provided in this embodiment of the invention has a functional structure that corresponds to the target tracking compensation control method for surgical robots in a dynamic breathing environment provided in this embodiment of the invention, and will not be described again here.

[0046] Figure 5 This is a schematic diagram of the structure of an electronic device 500 provided in an embodiment of the present invention. The electronic device 500 may vary considerably due to different configurations or performance. It may include one or more central processing units (CPUs) 501 and one or more memories 502. The memory 502 stores at least one instruction, which is loaded and executed by the processor 501 to implement the steps of the above-described target tracking compensation control method for surgical robots under dynamic breathing conditions.

[0047] In an exemplary embodiment, a computer-readable storage medium is also provided, such as a memory including instructions that can be executed by a processor in a terminal to complete the above-described surgical robot target tracking compensation control method under dynamic respiratory conditions. For example, the computer-readable storage medium may be a ROM, random access memory (RAM), CD-ROM, magnetic tape, floppy disk, or optical data storage device.

[0048] Those skilled in the art will understand that all or part of the steps of the above embodiments can be implemented by hardware or by a program instructing related hardware. The program can be stored in a computer-readable storage medium, such as a read-only memory, a disk, or an optical disk.

[0049] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A dynamic respiratory environment under the control method of surgical robot target tracking compensation, characterized in that, The surgical robot has two mechanical arms, one ultrasonic mechanical arm holding an ultrasonic probe, which collects ultrasonic images of a patient in real time to obtain internal lesion target point position information, and one puncture mechanical arm holding a puncture needle, which performs puncture surgery under the guidance of the ultrasonic images, and the method comprises the following steps: S1, real-time collection of ultrasonic images of a patient, and in the process of real-time collection of ultrasonic images of the patient, an adaptive admittance control algorithm based on force feedback is used to adaptively adjust force control output according to contact force between the ultrasonic probe and the skin surface, and changes in displacement and speed state of the ultrasonic probe, to real-time adjust stiffness parameters of an admittance control model, so that the ultrasonic probe and the skin surface maintain constant contact force, prevent probe drift or tissue deformation under pressure, ensure that the ultrasonic probe and the body surface maintain stable contact force while adapting to dynamic changes of the body surface, and improve image quality; S2, using real-time ultrasonic images, and continuously identifying and positioning the lesion by a template matching and dynamic template updating algorithm to obtain real-time internal lesion target point position information; S3, using the target point position information as a target pose, calculating a motion compensation amount of the puncture mechanical arm, so that a puncture needle tip trajectory and target point motion are kept in synchronization, to realize target point tracking and compensation control in a dynamic breathing environment.

2. The method of claim 1, wherein, The adaptive admittance control algorithm based on force feedback in S1 specifically comprises the following steps: The contact force between the ultrasonic probe and the skin surface is acquired in real time through a six-dimensional force sensor installed at the end of an ultrasonic mechanical arm ; calculating a current contact force error: wherein is an estimate of is the actual position of the ultrasound probe; is an environmental stiffness parameter; is an estimate of the environmental stiffness parameter; is the desired position of the ultrasound probe; is an estimate of the desired position of the ultrasound probe; inputting the contact force error into an admittance control model: wherein, is the contact force error of the th iteration, , , , respectively, denote the inertia, damping and stiffness parameter matrices of the mobility control model. The ultrasonic probe correction amount is obtained by iteration so that approaches 0, and further so that approaches 0, so that the probe maintains stable contact force under body surface fluctuation and respiratory motion. The stiffness parameter A Lyapunov function is established based on the error The online updating law is derived To ensure the convergence of the parameter estimation error Adjusting stiffness parameters in real time according to an update law Thus, the admittance control model is dynamically adjusted, so that the ultrasound probe can maintain the desired constant contact force to be kept under different soft tissue contact conditions.

3. The method of claim 1, wherein, S2 specifically comprises the following steps: S21, in an initial stage of the ultrasound-guided puncture process, manually selecting, by a doctor, a lesion region of interest on the ultrasound image as an initial template , the initial template includes feature information of the lesion; S22, acquiring the image sequence in real time and pre-processing each frame of image; S23, the pre-processed image sequence , using the current template as a reference, normalized cross-correlation (NCC) matching is performed using a deformation tolerant template including scaling and rotation to obtain a spatial similarity matrix of each candidate position, and then a gray scale similarity matrix is obtained, and a lesion region and a target position are determined; S24, fusing the determined lesion region and the current template by weight to obtain an updated template By allowing the template to rotate and scale within a limited range, the robustness of the tracking can be improved by adapting to the deformation of the lesion in ultrasound imaging.

4. The method of claim 3, wherein, S23 specifically comprises the following steps: to the current template by applying a scale factor and a rotation angle to generate a deformation tolerant template ; for each morphable template with the current image computing spatial similarity: wherein, represents the pixel coordinate inside the template, represents the pixel value of the pixel inside the template after applying the warping with the scaling factor s and the rotation angle θ; at the coordinate represents the average value of the gray scale of all pixels inside the warped template; represents the pixel value of the pixel inside the template at the coordinate in the image I at the current matching position represents the average value of the gray scale of all pixels in the image I. Final take all scale and angle combinations with maximum correlation value As a result of the match: the maximum correlation value corresponding and applied to the current template , a new template image T is obtained, and the gray similarity between the new template image T and the current image is calculated. in, and This represents the top-left corner coordinates of the current matching position of the template in image I. and Represents the pixel coordinates inside the template. This indicates that the new template image T is in grayscale value, This indicates that image I is at the current matching position. Below, corresponding to the pixels inside the template pixel values, This indicates that the new template image T and image I are at the same position. The matching degree at a given point ranges from [-1, 1], with a value closer to 1 indicating a higher matching degree. regarding a region with a matching degree greater than a preset threshold as a lesion region, and regarding a position with the highest matching degree in the region as a position of the target point.

5. The method of claim 1, wherein, S3 specifically comprises the following steps: S31, representing the target pose as , solving the rotation angles of each joint of the puncture robot arm by using the inverse kinematics of the robot arm ; S32, obtain the current end position by forward kinematics , and calculate the pose error of , so as to obtain the joint compensation amount , each joint compensation amount comprises : S33, updating joint angles in the following manner: Steps S32 and S33 are repeated, iterating until a convergence condition is met, such that the end pose unlimitedly approximates the target pose .

6. A surgical robotic target tracking compensation control system in a dynamic breathing environment, characterized by, The surgical robot has two mechanical arms, one ultrasonic mechanical arm holding an ultrasonic probe, which collects ultrasonic images of a patient in real time to obtain internal lesion target point position information, and one puncture mechanical arm holding a puncture needle, which performs puncture surgery under the guidance of the ultrasonic images, and the system comprises the following modules: An admittance control module is configured to collect ultrasonic images of a patient in real time, and in the process of real-time collection of ultrasonic images of the patient, an adaptive admittance control algorithm based on force feedback is used to adaptively adjust force control output according to contact force between the ultrasonic probe and the skin surface, and changes in displacement and speed state of the ultrasonic probe, to real-time adjust stiffness parameters of an admittance control model, so that the ultrasonic probe and the skin surface maintain constant contact force, prevent probe drift or tissue deformation under pressure, ensure that the ultrasonic probe and the body surface maintain stable contact force while adapting to dynamic changes of the body surface, and improve image quality; An identification and positioning module is configured to use real-time ultrasonic images, and continuously identify and position the lesion by a template matching and dynamic template updating algorithm to obtain real-time internal lesion target point position information; A tracking and compensation control module is configured to use the target point position information as a target pose, calculate a motion compensation amount of the puncture mechanical arm, so that a puncture needle tip trajectory and target point motion are kept in synchronization, to realize target point tracking and compensation control in a dynamic breathing environment.

7. The system of claim 6, wherein, The admittance control module is specifically configured to: The contact force between the ultrasonic probe and the skin surface is acquired in real time through a six-dimensional force sensor installed at the end of an ultrasonic mechanical arm ; calculate a current contact force error: wherein is an estimate of is the actual position of the ultrasound probe; is the environmental stiffness parameter; is an estimate of the environmental stiffness parameter; is the desired position of the ultrasound probe; is an estimate of the desired position of the ultrasound probe; input the contact force error into an admittance control model: wherein is the contact force error of the th iteration, , , , respectively, denote the inertia, damping and stiffness parameter matrices of the mobility control model. The ultrasonic probe correction amount is obtained by iteration so that approaches 0, and further so that approaches 0, so that the probe maintains stable contact force under body surface fluctuation and respiratory motion. stiffness parameters Treating the parameters to be estimated, a Lyapunov function is established based on the error. The online update law is derived through derivation. To ensure that the parameter estimation error converges; Adjusting stiffness parameters in real time according to an update law Thus, the admittance control model is dynamically adjusted so that the ultrasound probe can maintain the desired constant contact force to be kept under different soft tissue contact conditions.

8. The system of claim 6, wherein, The identification and positioning module is specifically configured to: S21, in an initial stage of the ultrasound-guided puncture process, manually selecting, by a doctor, a lesion region of interest on the ultrasound image as an initial template , the initial template includes feature information of the lesion; S22, acquiring the image sequence in real time and pre-processing each frame of image; S23, normalizing cross-correlation (NCC) matching is performed on the pre-processed image sequence with the current template as a reference, to obtain a spatial similarity matrix of each candidate position, and then a gray scale similarity matrix, and to determine a lesion region and a target position; S24, fusing the determined lesion region and the current template by weight to obtain an updated template By allowing the template to rotate and scale within a limited range, the robustness of the tracking can be improved by adapting to the deformation of the lesion in the ultrasound imaging.

9. The system of claim 8, wherein, The S23 specifically comprises: to the current template by applying a scale factor and a rotation angle to generate a deformation tolerant template ; for each morphable template with the current image computing spatial similarity: wherein, represents the pixel coordinates inside the template, represents the pixel value of the pixel inside the template after applying the warping with the scaling factor s and the rotation angle θ; at the pixel coordinates represents the average of the pixel values of all pixels inside the warped template; represents the pixel value of the pixel inside the template at the pixel coordinates after applying the warping with the scaling factor s and the rotation angle θ; represents the average of the pixel values of all pixels inside the warped template; Final take all scale and angle combinations with maximum correlation value As a result of the match: the maximum correlation value corresponding and applied to the current template , a new template image T is obtained, and the gray similarity between the new template image T and the current image is calculated. wherein, with denotes the top-left corner coordinate of the current matching position of the template on the image I, with denotes the pixel coordinate inside the template, denotes the gray value of the new template image T at denotes the pixel value of the image I at the current matching position denotes the matching degree of the new template image T and the image I at position , the value closer to 1 indicates the higher matching degree.​​​ The region with the matching degree greater than the preset threshold is taken as the lesion region, and the position with the highest matching degree therein is taken as the position of the target point.

10. The system of claim 6, wherein, The tracking compensation control module is specifically configured to: S31, representing the target pose as , solving the rotation angles of each joint of the puncture robot arm by using the inverse kinematics of the robot arm ; S32, obtain the current end position by forward kinematics , and calculate the pose error of , so as to obtain the joint compensation amount , each joint compensation amount comprises : S33, update the joint angle in the following manner: Steps S32 and S33 are repeated, iterating until a convergence condition is met, such that the end pose unlimitedly approximates the target pose .

Images