Spinal surgery mechanical arm precise regulation and control system integrating deep learning and dynamic calibration

Abstract

The invention relates to the technical field of artificial intelligence, and discloses a spinal surgery mechanical arm precise regulation and control system integrating deep learning and dynamic calibration, and the system comprises a digital twinborn body construction module which is used for constructing a dynamic spinal digital twinborn body of a target object; the motion path analysis module is used for generating a planned motion path of the mechanical arm; the coarse regulation and control module is used for calculating a translation vector and a rotation vector of the mechanical arm and executing coarse regulation and control of the mechanical arm when the path error is not smaller than a preset first threshold value; the fine regulation and control module is used for calculating the rigidity coefficient of the mechanical arm and executing fine regulation and control of the mechanical arm when the path error is smaller than a first threshold value and not smaller than a preset second threshold value; and the micron-order regulation and control module is used for calculating pulse voltage required by the micro-motion platform and executing micron-order regulation and control of the mechanical arm when the path error is smaller than a second threshold value and not smaller than a preset safety threshold value. The control precision and safety of the spine surgery mechanical arm can be improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of artificial intelligence, and in particular to a spine surgery robot arm precision control system fusing deep learning and dynamic calibration. BACKGROUND

[0002] The spine surgery robot arm refers to an intelligent robot system applied to spine surgery, which assists doctors to complete complex spine surgery operations such as pedicle screw implantation, vertebroplasty, and decompression fixation through high-precision motion control, real-time navigation, and force feedback technology. Spine surgery robot arm precision control can replicate the ideal path planned before the operation to the patient with sub-millimeter accuracy, ensuring that the surgical tool strictly follows the preset path, thereby minimizing the risk of nerve and blood vessel damage and reducing the risk of surgery.

[0003] The traditional control method of the surgical robot arm is to use a pre-programmed path, generate a pre-programmed path of the surgical robot arm according to the registration structure through preoperative image registration, which cannot adapt to the deformation of the anatomical structure caused by respiratory motion and instrument interaction in real time, so that effective adjustment cannot be made in the case of intraoperative tissue displacement, resulting in large positioning errors during the operation. SUMMARY

[0004] The present application provides a spine surgery robot arm precision control system fusing deep learning and dynamic calibration, which mainly aims to improve the control accuracy and safety of the spine surgery robot arm.

[0005] To achieve the above-mentioned purpose, the present application provides a spine surgery robot arm precision control system fusing deep learning and dynamic calibration, which comprises:

[0006] A digital twin construction module is configured to obtain multi-modal data of a target object, wherein the multi-modal data comprises CT image data, three-dimensional point cloud data, and instrument contact force data, so as to construct a dynamic spine digital twin of the target object;

[0007] A motion path analysis module is configured to calculate a dynamic deformation field of a corresponding spine of the target object according to the dynamic spine digital twin, so as to generate a planned motion path of a corresponding robot arm of the spine;

[0008] A coarse-level control module is configured to obtain a real-time motion path of the robot arm, calculate a path error between the real-time motion path and the planned motion path, calculate a translation vector and a rotation vector of the robot arm when the path error is not less than a preset first threshold, generate a pose calibration instruction of the robot arm according to the translation vector and the rotation vector, and perform coarse-level control of the robot arm according to the pose calibration instruction;

[0009] a fine control module, configured to, when the path error is less than the first threshold value and not less than a preset second threshold value, calculate a stiffness coefficient of the robot arm, determine a displacement compensation amount of the robot arm according to the stiffness coefficient, generate an impedance calibration instruction of the robot arm according to the impedance calibration instruction, and perform fine control of the robot arm according to the impedance calibration instruction;

[0010] a micrometer control module, configured to, when the path error is less than the second threshold value and not less than a preset safety threshold value, calculate a pulse voltage required by a micro-motion platform corresponding to the robot arm, generate a micrometer calibration instruction of the robot arm according to the micrometer calibration instruction, and perform micrometer control of the robot arm according to the micrometer calibration instruction.

[0011] Optionally, the calculation of the translation vector and the rotation vector of the robot arm when the path error is not less than the preset first threshold value comprises:

[0012] determining an actual pose and a planned pose of the robot arm according to the path error;

[0013] calculating a translation vector of the robot arm according to the actual pose and the planned pose;

[0014] respectively determining an actual attitude quaternion of the actual pose and a planned pose quaternion of the planned pose;

[0015] calculating a relative rotation quaternion of the robot arm according to the actual attitude quaternion and the planned pose quaternion;

[0016] calculating a rotation angle and a rotation unit vector of the robot arm according to the relative rotation quaternion;

[0017] determining a rotation vector of the robot arm according to the rotation angle and the rotation unit vector.

[0018] Optionally, the calculation of the stiffness coefficient of the robot arm when the path error is less than the first threshold value and not less than a preset second threshold value comprises:

[0019] extracting a CT grayscale value and a bone-implant contact force of multi-modal data corresponding to the robot arm;

[0020] analyzing a current regional bone density of a spine region corresponding to the robot arm according to the CT grayscale value;

[0021] calculating a stiffness coefficient of the robot arm according to the current regional bone density, the bone-implant contact force, and the path error

[0022] Optionally, the determination of the displacement compensation amount of the robot arm according to the stiffness coefficient comprises:

[0023] According to the stiffness coefficient, analyze the damping coefficient of the mechanical arm;

[0024] Combine the damping coefficient and the stiffness coefficient, fit the impedance model of the mechanical arm;

[0025] According to the path error corresponding to the mechanical arm, calculate the static compensation amount of the mechanical arm through the impedance model;

[0026] Determine the breathing motion law of the target object corresponding to the mechanical arm;

[0027] According to the breathing motion law and the static compensation amount, analyze the displacement compensation amount of the mechanical arm.

[0028] Optionally, when the path error is less than the second threshold and not less than a preset safety threshold, the required pulse voltage of the micro-motion platform corresponding to the mechanical arm is calculated, comprising:

[0029] Determine the mapping coefficient of the mechanical arm coordinate system corresponding to the mechanical arm and the micro-motion platform coordinate system of the micro-motion platform;

[0030] According to the path error and the mapping coefficient, determine the displacement amount required to be compensated by the mechanical arm;

[0031] Analyze the displacement-voltage linear relationship of the micro-motion platform to determine the displacement-voltage conversion coefficient of the micro-motion platform;

[0032] According to the displacement-voltage conversion coefficient and the displacement amount, determine the pulse voltage of the micro-motion platform.

[0033] Optionally, the dynamic spine digital twin of the target object is constructed, comprising:

[0034] Respectively construct the image coordinate system corresponding to the CT image data of the target object and the surgical three-dimensional coordinate system corresponding to the three-dimensional point cloud data of the CT image data of the target object;

[0035] Determine the transformation matrix of the image coordinate system and the surgical three-dimensional coordinate system;

[0036] Based on the transformation matrix, perform spatio-temporal registration on the multi-modal data corresponding to the target object to obtain multi-modal registration data;

[0037] Extract the multi-modal features of the multi-modal registration data to perform data fusion on the multi-modal registration data to obtain fusion data;

[0038] Based on the fusion data, fit the spine three-dimensional model of the target object;

[0039] simulate a dynamic behavior state of the spine three-dimensional model to construct a dynamic spine digital twin of the spine three-dimensional model.

[0040] Optionally, the calculating, according to the dynamic spine digital twin, of a dynamic deformation field of a spine corresponding to the target object comprises:

[0041] simulate an initial static equilibrium state of the spine through the dynamic spine digital twin;

[0042] calculate an initial displacement field of the initial static equilibrium state;

[0043] define a time step and a dynamic load of the dynamic spine digital twin;

[0044] simulate a dynamic change state of the spine according to the time step and the dynamic load;

[0045] analyze a displacement vector of a node corresponding to the spine in the dynamic change state based on the initial displacement field;

[0046] determine a dynamic deformation field of the spine according to the displacement vector.

[0047] Optionally, the generating of the planning motion path of the mechanical arm corresponding to the spine comprises:

[0048] identify a workspace of the mechanical arm;

[0049] map the dynamic deformation field of the spine to the workspace to obtain a mapped deformation field;

[0050] calculate a target position sequence and a posture sequence of the mechanical arm according to the mapped deformation field;

[0051] generate the planning motion path of the mechanical arm according to the target position sequence and the posture sequence.

[0052] Optionally, the calculating of a path error of the real-time motion path and the planning motion path comprises:

[0053] perform time sequence discretization processing on the real-time motion path and the planning motion path respectively to obtain a real-time path sequence and a planning path sequence;

[0054] construct a real-time path vector set of the real-time path sequence and a planning path vector set of the planning path sequence respectively;

[0055] calculate a position error, a posture error and an angle error of the real-time motion path and the planning motion path according to the real-time path vector set and the planning path vector set;

[0056] According to the position error, the attitude error and the angle error, a path error of the real-time motion path and the planned motion path is determined.

[0057] Optionally, the calculating the static compensation amount of the mechanical arm through the impedance model comprises:

[0058] The mechanical arm external force is calculated through the impedance model;

[0059] According to the mechanical arm external force and the corresponding stiffness coefficient of the mechanical arm, a static displacement of the mechanical arm is calculated;

[0060] Based on the static displacement, a static compensation amount of the mechanical arm is determined.

[0061] The embodiment of the present application can reflect the geometric and mechanical behavior of the spine in a dynamic environment by constructing the dynamic spine digital twin of the target object, realize adaptive response to the dynamic environment, and lay a data foundation for subsequent mechanical arm path planning. Optionally, the embodiment of the present application can dynamically compensate for real-time deformation (such as micro-displacement caused by respiration, heartbeat and muscle contraction) of the spine during the operation process by generating the planned motion path of the mechanical arm corresponding to the spine, thereby avoiding positioning deviation caused by anatomical structure movement. The embodiment of the present application can significantly reduce trajectory tracking error and improve motion control accuracy by generating the pose calibration instruction of the mechanical arm, thereby improving the calibration robustness of the mechanical arm and ensuring motion stability and reliability. The embodiment of the present application can correct path errors caused by inertia, friction and external disturbance of the mechanical arm in high-speed or high-load motion in real time by determining the displacement compensation amount of the mechanical arm according to the stiffness coefficient, thereby ensuring that the mechanical arm moves strictly according to the planned path. The embodiment of the present application can realize micro-displacement adjustment of the mechanical arm by calculating the pulse voltage required by the micro-motion platform corresponding to the mechanical arm when the path error is less than the second threshold value and not less than a preset safety threshold, thereby ensuring that the mechanical arm does not overshoot and oscillate when approaching the target position, and realizing nanometer-level positioning accuracy. Finally, the embodiment of the present application can accurately control the displacement of the end effector of the mechanical arm to achieve micro-meter-level precision of the target position by executing micro-meter-level regulation and control of the mechanical arm according to the micro-meter-level calibration instruction, thereby avoiding damage to surrounding tissues and improving surgical safety. Therefore, the control accuracy and safety of the spine surgery mechanical arm are improved. BRIEF DESCRIPTION OF DRAWINGS

[0062] Figure 1 A function module diagram of a spine surgery mechanical arm precise regulation and control system fusing deep learning and dynamic calibration is provided for an embodiment of the present application;

[0063] Figure 2A flowchart of a precise regulation method of a spine surgery robot arm fusing deep learning and dynamic calibration is provided for an embodiment of the present application.

[0064] The implementation, functional features and advantages of the present application will be further described with reference to the embodiments and the accompanying drawings. DETAILED DESCRIPTION

[0065] To make the objectives, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings of the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work fall within the protection scope of the present application.

[0066] In addition, the step sequence in each method embodiment described below is only an example, not a strict limitation.

[0067] In fact, the server device deployed by the precise regulation system of a spine surgery robot arm fusing deep learning and dynamic calibration can be composed of one or more devices. The precise regulation system of a spine surgery robot arm fusing deep learning and dynamic calibration can be implemented as a business instance, a virtual machine or a hardware device. For example, the precise regulation system of a spine surgery robot arm fusing deep learning and dynamic calibration can be implemented as a business instance deployed on one or more devices in a cloud node. In short, the precise regulation system of a spine surgery robot arm fusing deep learning and dynamic calibration can be understood as a software deployed on a cloud node, which is used to provide the service of precise regulation of a spine surgery robot arm fusing deep learning and dynamic calibration for each user terminal. Alternatively, the precise regulation system of a spine surgery robot arm fusing deep learning and dynamic calibration can also be implemented as a virtual machine deployed on one or more devices in a cloud node. The virtual machine has application software installed therein for managing each user terminal. Alternatively, the precise regulation system of a spine surgery robot arm fusing deep learning and dynamic calibration can also be implemented as a server composed of a plurality of same or different types of hardware devices, and one or more hardware devices are arranged to provide the service of precise regulation of a spine surgery robot arm fusing deep learning and dynamic calibration for each user terminal.

[0068] In an implementation form, the spinal surgery robot arm precision regulation system fusing deep learning and dynamic calibration and the user end are mutually adaptive. That is, the spinal surgery robot arm precision regulation system fusing deep learning and dynamic calibration is installed as an application on a cloud service platform, and the user end is a client establishing a communication connection with the application; or the spinal surgery robot arm precision regulation system fusing deep learning and dynamic calibration is implemented as a website, and the user end is a webpage; or the spinal surgery robot arm precision regulation system fusing deep learning and dynamic calibration is implemented as a cloud service platform, and the user end is a mini-program in an instant messaging application.

[0069] Referring to Figure 1 Fig. 1 is a functional module diagram of a spinal surgery robot arm precision regulation system fusing deep learning and dynamic calibration according to an embodiment of the present application.

[0070] The spinal surgery robot arm precision regulation system fusing deep learning and dynamic calibration 100 can be provided in a cloud server, and in an implementation form, can be one or more service devices, or can be installed as an application on a cloud (such as a server, a server cluster, etc. for the spinal surgery robot arm precision regulation fusing deep learning and dynamic calibration), or can be developed as a website. According to the functions implemented, the spinal surgery robot arm precision regulation system fusing deep learning and dynamic calibration 100 includes a digital twin construction module 101, a motion path analysis module 102, a coarse level regulation module 103, a fine level regulation module 104, and a micrometer level regulation module 105.

[0071] In the embodiment of the present application, each of the modules in the tracking of the spinal surgery robot arm precision regulation fusing deep learning and dynamic calibration can be independently implemented and called by other modules. The calling here can be understood as that a module can connect multiple modules of another type and provide corresponding services for the connected multiple modules. In the spinal surgery robot arm precision regulation system fusing deep learning and dynamic calibration provided by the embodiment of the present application, without modifying the program code, the application range of the spinal surgery robot arm precision regulation architecture fusing deep learning and dynamic calibration can be adjusted by adding modules and directly calling, to realize cluster-level expansion, so as to achieve the purpose of quickly and flexibly expanding the spinal surgery robot arm precision regulation system fusing deep learning and dynamic calibration.

[0072] The following will describe the components and specific work flow of the spinal surgery robot arm precision regulation system fusing deep learning and dynamic calibration with reference to specific embodiments.

[0073] The digital twin construction module 101 is configured to acquire multi-modal data of a target object, wherein the multi-modal data comprises CT image data, three-dimensional point cloud data and instrument contact force data, so as to construct a dynamic spine digital twin of the target object.

[0074] According to the embodiment of the present application, the multi-modal data of the target object can be acquired to provide a data basis for subsequent data analysis, path planning and accurate regulation. The multi-modal data refers to a data set with different physical properties, information dimensions or expression forms acquired by different types of sensors, devices and acquisition methods. The CT image data refers to high-resolution three-dimensional medical image data acquired by computed tomography. The three-dimensional point cloud data refers to data providing real-time and dynamic spatial position and posture of the spine during the operation process. The instrument contact force data refers to real-time information of physical interaction between the mechanical arm and the spine tissue, including the size, direction and torque of the contact force.

[0075] According to the embodiment of the present application, the dynamic spine digital twin of the target object can be constructed to truly reflect the geometric and mechanical behavior of the spine in a dynamic environment, realize adaptive response to the dynamic environment, and lay a data foundation for subsequent mechanical arm path planning. The dynamic spine digital twin refers to a virtual digital model based on real patient spine data, which can be dynamically updated in real time and simulate the deformation and mechanical behavior of the spine during the movement process.

[0076] As an embodiment of the present application, the dynamic spine digital twin of the target object is constructed, comprising:

[0077] An image coordinate system corresponding to the CT image data of the target object and a surgical three-dimensional coordinate system corresponding to the three-dimensional point cloud data of the CT image data of the target object are respectively constructed;

[0078] A transformation matrix of the image coordinate system and the surgical three-dimensional coordinate system is determined;

[0079] Based on the transformation matrix, multi-modal data corresponding to the target object is time-space registered to obtain multi-modal registration data;

[0080] Multi-modal features of the multi-modal registration data are extracted to perform data fusion on the multi-modal registration data to obtain fusion data;

[0081] Based on the fusion data, a spine three-dimensional model of the target object is fitted;

[0082] A dynamic behavior state of the spine three-dimensional model is simulated to construct a dynamic spine digital twin of the spine three-dimensional model.

[0083] The image coordinate system refers to a coordinate system used to describe the spatial position of pixels in a CT slice. The surgical three-dimensional coordinate system refers to a coordinate system based on a 3D point cloud acquisition device (such as a laser scanner) in surgery. The transformation matrix refers to a matrix used to map points from the image coordinate system to the surgical coordinate system. The multi-modal registration data refers to a unified data set formed by spatio-temporal alignment of data from different sources (CT, point cloud, force sensation). The multi-modal feature refers to representative information extracted from the registered multi-modal data, such as geometric features (such as curvature, surface normal vector), texture features (such as gray scale distribution), statistical features (such as mean, variance), etc. The fusion data refers to a data set integrated from different modalities. The spinal three-dimensional model refers to a three-dimensional geometric model with true anatomical structure features reconstructed based on the fusion data. The dynamic behavior state refers to the mechanical response and deformation state of the spinal three-dimensional model under different loads, postures and surgical operations, such as displacement, stress, strain, rotation angle and other physical quantities.

[0084] Optionally, the transformation matrix of the image coordinate system and the surgical three-dimensional coordinate system can be determined by a paired point method, such as a singular value decomposition algorithm, an iterative closest point algorithm, etc.

[0085] Optionally, the fusion data can be obtained by multi-modal feature fusion technology, such as weighted fusion, deep learning fusion, etc.

[0086] Optionally, the dynamic spinal digital twin of the spinal three-dimensional model can be constructed by deep learning technology, such as convolutional neural network, recurrent neural network, generative adversarial network, etc.

[0087] The motion path analysis module 102 is configured to calculate a dynamic deformation field of the corresponding spine of the target object according to the dynamic spinal digital twin, to generate a planned motion path of the corresponding mechanical arm of the spine.

[0088] According to the dynamic spinal digital twin, the dynamic deformation field of the corresponding spine of the target object can be calculated to reflect the slight deformation of the spine caused by breathing, muscle contraction, instrument operation, etc. during the operation process in real time, and the operation path can be adjusted in real time to avoid misoperation caused by anatomical structure deviation. The dynamic deformation field refers to a mathematical expression describing the change of the geometric shape and spatial position of the spine with time under the action of external force, motion and physiological activity (such as breathing, heartbeat, muscle contraction).

[0089] As an embodiment of the present application, the dynamic deformation field of the corresponding spine of the target object is calculated according to the dynamic spinal digital twin, comprising:

[0090] simulate an initial static equilibrium state of the spine by the dynamic spine digital twin;

[0091] calculate an initial displacement field of the initial static equilibrium state;

[0092] define a time step and a dynamic load of the dynamic spine digital twin;

[0093] simulate a dynamic changing state of the spine according to the time step and the dynamic load;

[0094] analyze displacement vectors of corresponding nodes of the spine in the dynamic changing state based on the initial displacement field;

[0095] determine a dynamic deformation field of the spine according to the displacement vectors.

[0096] The initial static equilibrium state refers to the state of the spine under no external force (except gravity) and all tissues (vertebral body, ligament, muscle) in mechanical equilibrium. The initial displacement field refers to a set of small displacement vectors of nodes of the spine relative to the three-dimensional model of the spine in the initial static equilibrium state. The time step refers to the length of discretizing continuous time into a series of small time intervals in dynamic simulation. The dynamic load refers to the external force that changes with time and is applied to the spine model to simulate dynamic disturbance in the surgical process. The dynamic changing state refers to the continuously changing mechanical state of the spine over time under the action of the dynamic load. The displacement vector refers to the position change of the node in the dynamic process relative to the initial static equilibrium state.

[0097] Optionally, the initial displacement field of the initial static equilibrium state can be calculated by a machine learning model, such as random forest, GNN, etc.

[0098] Optionally, the displacement vectors of corresponding nodes of the spine in the dynamic changing state can be analyzed by deep learning technology, such as long short-term memory network, CNN-LSTM hybrid model, etc.

[0099] As another embodiment, the dynamic load is represented by the following formula:

[0100]

[0101] wherein, represents the dynamic load at time t, represents the amplitude of the load, represents the sine function, represents the circular constant, represents the load frequency, represents the time t. ​

[0102] For example, there is a periodic force on the spine caused by breathing, with a load amplitude of 10N and a load frequency of 0.25Hz, and a time of 4s, then Similarly, , , .

[0103] The embodiment of the present application can dynamically compensate for real-time deformation of the spine (such as micro-displacement caused by breathing, heartbeat, muscle contraction) during the operation process by generating a planning motion path of the mechanical arm corresponding to the spine, avoiding positioning deviation caused by the movement of anatomical structures. The planning motion path refers to an optimal motion trajectory designed for the surgical mechanical arm from the starting position to the target position according to the spatiotemporal variation characteristics of the spine in the dynamic deformation field.

[0104] As an embodiment of the present application, the generation of the planning motion path of the mechanical arm corresponding to the spine comprises:

[0105] Identifying the workspace of the mechanical arm;

[0106] Mapping the dynamic deformation field of the spine to the workspace to obtain a mapped deformation field;

[0107] According to the mapped deformation field, calculating a target position sequence and an attitude sequence of the mechanical arm;

[0108] According to the target position sequence and the attitude sequence, generating a planning motion path of the mechanical arm.

[0109] The workspace refers to the set of all spatial points that the mechanical arm can reach. The mapped deformation field refers to the deformation field converted from the dynamic deformation field of the spine to the workspace coordinate system of the mechanical arm. The target position sequence refers to the set of spatial position points that the mechanical arm needs to reach in sequence during the motion process. The attitude sequence refers to the set of attitudes that the mechanical arm needs to maintain in sequence during the motion process.

[0110] Optionally, the target position sequence and the attitude sequence of the mechanical arm can be calculated by a deformation field interpolation algorithm, such as a radial basis function neural network.

[0111] Optionally, the planning motion path of the mechanical arm corresponding to the spine can be generated by a path planning algorithm, such as a rapidly-exploring random tree, a probabilistic roadmap, an optimization algorithm, etc.

[0112] The coarse regulation module 103 is configured to acquire a real-time motion path of the robot arm, calculate a path error of the real-time motion path and the planned motion path, calculate a translation vector and a rotation vector of the robot arm when the path error is not less than a preset first threshold, generate a pose calibration instruction of the robot arm, and perform coarse regulation of the robot arm according to the pose calibration instruction.

[0113] According to the embodiment of the present application, the real-time motion path of the robot arm can be acquired to realize accurate tracking of the path, and avoid surgical errors caused by motion deviation of the robot arm. The real-time motion path refers to a space motion trajectory recorded and fed back by the robot arm in real time during surgery.

[0114] According to the embodiment of the present application, the path error of the real-time motion path and the planned motion path can be calculated in real time to accurately evaluate the deviation of the end effector of the robot arm from the predetermined path, and provide a basis for subsequent error compensation and path calibration. The path error refers to the difference between the actual motion trajectory and the preset planned trajectory of the robot arm during execution of the surgical task.

[0115] As an embodiment of the present application, the calculation of the path error of the real-time motion path and the planned motion path comprises:

[0116] The real-time motion path and the planned motion path are respectively subjected to time sequence discretization processing to obtain a real-time path sequence and a planned path sequence.

[0117] A real-time path vector set of the real-time path sequence and a planned path vector set of the planned path sequence are respectively constructed.

[0118] According to the real-time path vector set and the planned path vector set, a position error, a pose error and an angle error of the real-time motion path and the planned motion path are calculated.

[0119] According to the position error, the pose error and the angle error, the path error of the real-time motion path and the planned motion path is determined.

[0120] The real-time path sequence refers to an ordered point set formed after time discretization processing of a real-time motion path. The planning path sequence refers to an ordered point set formed after discretization in the time dimension of an ideal motion trajectory of the robot arm. The real-time path vector set refers to a feature vector set extracted from the real-time path sequence for error calculation. The planning path vector set refers to a feature vector set extracted from the planning path sequence for error comparison. The position error refers to the Euclidean distance deviation between the position of a specific time in the real-time path sequence and the position of the corresponding time in the planning path sequence. The attitude error refers to the deviation between the attitude of a specific time in the real-time path sequence and the attitude of the corresponding time in the planning path sequence. The angle error refers to the angle deviation between the real-time path vector set and the planning path vector set.

[0121] Optionally, the real-time path sequence can be obtained by a dynamic time warping discretization method.

[0122] Optionally, the position error of the real-time motion path and the planning motion path can be calculated by a Euclidean distance calculation method. The attitude error can be calculated by a quaternion difference method.

[0123] The embodiment of the application can accurately quantify the position deviation and direction deviation of the end effector of the robot arm in space by calculating the translation vector and rotation vector of the robot arm when the path error is not less than a preset first threshold, providing an accurate mathematical basis for subsequent pose calibration, and ensuring that the robot arm can be accurately adjusted to the target position and attitude. The preset first threshold refers to an upper limit of error tolerance preset in the robot arm control system, which is used to judge whether the deviation between the real-time motion path and the planning path of the robot arm exceeds a safe range, which can be set to 0.5mm in this application. The translation vector refers to a position change amount required for the robot arm to move in three-dimensional space. The rotation vector refers to a mathematical tool for representing the rotation change required for the robot arm in space.

[0124] As an embodiment of the application, when the path error is not less than the preset first threshold, the translation vector and the rotation vector of the robot arm are calculated, comprising:

[0125] determining the actual pose and the planning pose of the robot arm according to the path error;

[0126] calculating the translation vector of the robot arm according to the actual pose and the planning pose;

[0127] determining actual attitude quaternions of the actual pose and planning pose quaternions of the planning pose, respectively;

[0128] According to the actual pose quaternion and the planned pose quaternion, a relative rotation quaternion of the robot arm is calculated;

[0129] According to the relative rotation quaternion, a rotation angle and a rotation unit vector of the robot arm are calculated;

[0130] According to the rotation angle and the rotation unit vector, a rotation vector of the robot arm is determined.

[0131] The actual pose refers to the real position and attitude of the robot arm in space. The planned pose refers to the position and attitude that the robot arm is expected to reach in motion planning. The actual pose quaternion refers to the quaternion describing the actual attitude of the robot arm. The planned pose quaternion refers to the quaternion describing the expected attitude of the robot arm. The relative rotation quaternion refers to the amount of rotation required to rotate from the actual attitude to the planned attitude. The rotation angle refers to the angle that the robot arm needs to rotate around an axis. The rotation unit vector refers to the unit vector of the rotation axis direction.

[0132] Optionally, the translation vector of the robot arm can be calculated by direct difference method.

[0133] As another embodiment, the relative rotation quaternion can be calculated by the following formula:

[0134]

[0135] wherein, represents the relative rotation quaternion, represents the planned pose quaternion, represents the actual pose quaternion, represents the actual pose quaternion, represents the quaternion multiplication operator.

[0136] As another embodiment, the rotation angle and the rotation unit vector are calculated by the following formula:

[0137]

[0138]

[0139] wherein, represents the rotation angle, represents the inverse cosine function, represents the real part of the relative rotation quaternion, represents the component of the corresponding imaginary part of the relative rotation quaternion on the x-axis, represents the component of the corresponding imaginary part of the relative rotation quaternion on the y-axis, represents the component of the corresponding imaginary part of the relative rotation quaternion on the z-axis, represents the component of the corresponding imaginary part of the relative rotation quaternion on the z-axis, denotes the component of the relative rotation quaternion on the imaginary axis, denotes the component of the relative rotation quaternion on the imaginary axis, denotes the sine function.

[0140] Exemplarily, there are planning pose quaternions , actual pose quaternions , a relative rotation quaternion , a rotation angle , a rotation unit vector .

[0141] The embodiment of the application can significantly reduce the trajectory tracking error, improve the motion control accuracy, and thus improve the mechanical arm calibration robustness, ensure the motion stability and reliability by generating the pose calibration instruction of the mechanical arm. The pose calibration instruction of the mechanical arm is a control instruction for dynamically adjusting the pose of the mechanical arm generated according to the translation vector and the rotation vector of the mechanical arm.

[0142] Optionally, the pose calibration instruction of the mechanical arm can be generated by a reinforcement learning algorithm.

[0143] The embodiment of the application can adopt a larger step size and gain coefficient by performing the coarse regulation of the mechanical arm according to the pose calibration instruction, so that the mechanical arm can approach the target pose faster, gain time for subsequent fine adjustment, and improve the overall control efficiency. The coarse regulation refers to a stage in the motion control of the mechanical arm for preliminarily adjusting the pose (position and attitude) of the mechanical arm in a short time and with a large amplitude, so as to quickly approach the target pose.

[0144] The fine regulation module 104 is configured to calculate a stiffness coefficient of the mechanical arm when the path error is less than the first threshold value and not less than a preset second threshold value, determine a displacement compensation amount of the mechanical arm according to the stiffness coefficient, generate an impedance calibration instruction of the mechanical arm according to the impedance calibration instruction, and perform fine regulation of the mechanical arm according to the impedance calibration instruction.

[0145] The embodiment of the application can automatically adjust the end stiffness according to the bone density and the real-time contact force by calculating the stiffness coefficient of the mechanical arm when the path error is less than the first threshold value and not less than a preset second threshold value, thereby avoiding instrument slipping caused by sudden change of tissue characteristics, and thus improving the safety of the operation. The preset second threshold value is a key demarcation value for triggering different regulation modes of the mechanical arm, which can be set to 0.1 mm in the present application.

[0146] As an embodiment of the application, when the path error is less than the first threshold value and not less than a preset second threshold value, the stiffness coefficient of the mechanical arm is calculated, including:

[0147] extracting a CT grayscale value and a bone-instrument contact force of the multi-modal data corresponding to the robot arm;

[0148] analyzing a current regional bone density of a spine region corresponding to the robot arm according to the CT grayscale value;

[0149] calculating a stiffness coefficient of the robot arm according to the current regional bone density, the bone-instrument contact force, and the path error.

[0150] The CT grayscale value refers to the relative density value of a pixel in a CT. The bone-instrument contact force refers to the real-time interaction force between the robot arm and the bone tissue. The current regional bone density refers to the bone tissue density of the current operation region. The healthy cortical bone refers to the part with the highest density in the normal human skeleton. The cortical bone density threshold refers to the density limit value for distinguishing the healthy cortical bone from other bone tissues, which can be set to 1200HU in the present application. The bone density-stiffness nonlinear relationship coefficient refers to the coefficient describing the nonlinear relationship between the bone density and the stiffness. The force feedback gain coefficient refers to the weight for adjusting the influence of the bone-instrument contact force on the stiffness coefficient.

[0151] Optionally, the CT grayscale value can be extracted by a deep learning model, such as U-Net, VisionTransformer, etc.

[0152] Optionally, the current regional bone density can be analyzed by establishing a calibration curve between the CT grayscale value and the bone density.

[0153] As another embodiment, the stiffness coefficient is calculated by the following formula:

[0154]

[0155] wherein, represents the stiffness coefficient, represents the basic stiffness of the healthy cortical bone corresponding to the spine, represents the current regional bone density, represents the cortical bone density threshold corresponding to the spine, represents the bone density-stiffness nonlinear relationship coefficient, represents the force feedback gain coefficient, represents the bone-instrument contact force, represents the path error.

[0156] It should be explained that in the present application, the formula represents the bone density-stiffness nonlinear relationship coefficient, and the value is 1.5, represents the force feedback gain coefficient, and the value is (0, 1], the greater the value, the more significant the influence of force feedback on stiffness.

[0157] For example, if the base stiffness is 500 N / mm, the bone-implant contact force is 12 N, the path error is 0.4 mm, the bone density-stiffness nonlinearity coefficient is 1.5, the force feedback gain coefficient is 0.2, the current region bone density is 850 HU, and the cortical bone density threshold is 1200 HU, then the stiffness coefficient is N / mm.

[0158] According to the stiffness coefficient, the displacement compensation amount of the mechanical arm can be determined in real time to correct the path error caused by inertia, friction and external disturbance of the mechanical arm during high-speed or high-load movement, and ensure that the mechanical arm moves strictly according to the planned path. The displacement compensation amount refers to the displacement vector that needs to be adjusted in real time by the end effector of the mechanical arm.

[0159] As an embodiment of the present application, the determination of the displacement compensation amount of the mechanical arm according to the stiffness coefficient comprises:

[0160] According to the stiffness coefficient, the damping coefficient of the mechanical arm is analyzed;

[0161] The impedance model of the mechanical arm is fitted in combination with the damping coefficient and the stiffness coefficient;

[0162] According to the path error corresponding to the mechanical arm, the static compensation amount of the mechanical arm is calculated through the impedance model;

[0163] The breathing motion law of the target object corresponding to the mechanical arm is determined;

[0164] According to the breathing motion law and the static compensation amount, the displacement compensation amount of the mechanical arm is analyzed.

[0165] The damping coefficient refers to a parameter describing the energy dissipation rate of the mechanical arm system during movement. The impedance model refers to a mathematical model describing the dynamic response behavior of the end effector of the mechanical arm under external force. The static compensation amount refers to the displacement adjustment amount calculated by a control algorithm to compensate for the end position deviation caused by factors such as gravity, static load and structural deformation during low-speed movement of the mechanical arm. The breathing motion law refers to the law of periodic three-dimensional spatial displacement, deformation and velocity change of the spine with the breathing rhythm during natural breathing of the human body.

[0166] Optionally, the impedance model of the mechanical arm can be fitted through meta-learning, such as MAML, Reptile, etc.

[0167] Optionally, the calculation of the static compensation amount of the mechanical arm through the impedance model comprises:

[0168] calculating an external force of the robot arm through the impedance model;

[0169] calculating a static displacement of the robot arm according to the external force of the robot arm and a stiffness coefficient corresponding to the robot arm;

[0170] determining a static compensation amount of the robot arm based on the static displacement.

[0171] The external force of the robot arm refers to an acting force on the end of the robot arm from an external environment and a load, such as a gravity load, a contact force, and the like. The static displacement refers to a displacement amount generated when the end of the robot arm reaches a steady state (i.e., both the velocity and the acceleration are zero) under the action of the external force. The static compensation amount refers to an additional position adjustment amount that needs to be applied to offset the static displacement generated by the robot arm under the action of the external force.

[0172] As another embodiment, the damping coefficient can be calculated by the following formula:

[0173]

[0174] wherein, denotes the damping coefficient, denotes the damping ratio, denotes the stiffness coefficient, denotes the virtual mass.

[0175] It needs to be explained that in the present application, the formula in the formula denotes the damping ratio, and the value is 0.7, denotes the virtual mass, and the value is 0.1 kg.

[0176] The embodiment of the present application can smoothly switch between position control and force control by generating the impedance calibration instruction of the robot arm, adapt to complex task requirements, and avoid damage caused by excessive rigidity and insufficient precision caused by excessive softness. The impedance calibration instruction refers to a set of control instructions for adjusting the impedance parameters of the robot arm in the robot arm control system to make the dynamic response characteristics (such as stiffness, damping, and inertia) of the robot arm meet specific task requirements.

[0177] The embodiment of the present application can make the robot arm generate a small displacement under the action of an external force, avoid rigid collision, and improve the safety and comfort of human-robot collaboration by performing fine-level regulation and control of the robot arm according to the impedance calibration instruction. The fine-level regulation and control refers to high-precision, small-amplitude, and fine position and force control adjustment of the end effector of the robot arm after the preliminary coarse-level regulation and control is completed.

[0178] The micro-level regulation module 105 is configured to calculate a pulse voltage required by a micro-drive platform corresponding to the robot arm when the path error is less than the second threshold value and not less than a preset safety threshold value, to generate a micro-level calibration instruction of the robot arm, and to perform micro-level regulation of the robot arm according to the micro-level calibration instruction.

[0179] The embodiment of the present application can realize the adjustment of the slight displacement of the robot arm by calculating the pulse voltage required by the micro-drive platform corresponding to the robot arm when the path error is less than the second threshold value and not less than a preset safety threshold value, and ensure that the robot arm does not overshoot and oscillate when approaching the target position, thereby realizing the nanometer-level positioning accuracy. The preset safety threshold value refers to a key error limit set in the motion control of the robot arm to ensure the safety, stability and reliability of the system, which can be set to 0.01 mm in the present application.

[0180] As an embodiment of the present application, when the path error is less than the second threshold value and not less than a preset safety threshold value, the calculation of the pulse voltage required by the micro-drive platform corresponding to the robot arm includes:

[0181] determining a mapping coefficient of a robot arm coordinate system corresponding to the robot arm and a micro-drive platform coordinate system of the micro-drive platform;

[0182] determining a displacement amount required to be compensated by the robot arm according to the path error and the mapping coefficient;

[0183] analyzing a displacement-voltage linear relationship of the micro-drive platform to determine a displacement-voltage conversion coefficient of the micro-drive platform;

[0184] determining the pulse voltage of the micro-drive platform according to the displacement-voltage conversion coefficient and the displacement amount.

[0185] The mapping coefficient refers to the conversion relationship between the robot arm coordinate system and the micro-drive platform coordinate system. The displacement amount refers to the displacement required to be compensated by the micro-drive platform. The displacement-voltage linear relationship refers to an approximate linear function relationship between the driving voltage input by the micro-drive platform and the mechanical displacement generated thereby. The displacement-voltage conversion coefficient refers to the displacement amount that can be driven by a unit voltage.

[0186] Optionally, the mapping coefficient of the robot arm coordinate system corresponding to the robot arm and the micro-drive platform coordinate system of the micro-drive platform can be determined by an augmented reality assisted calibration method, such as HoloLens, Magic Leap, etc.

[0187] Optionally, the displacement-voltage linear relationship of the micro-drive platform can be analyzed by a recursive least squares algorithm.

[0188] This invention enables micron- or even nanometer-level displacement control by generating micron-level calibration commands for the robotic arm, thereby improving surgical precision and enhancing surgical quality and safety. Specifically, the micron-level calibration commands refer to control commands used to drive the robotic arm's end effector to perform micron-level (μm) or even sub-micron-level precision displacement or attitude adjustments.

[0189] This invention, through micron-level calibration instructions, enables precise control of the robotic arm's end effector displacement to achieve micron-level accuracy at the target position, thereby avoiding damage to surrounding tissues and improving surgical safety. The micron-level control refers to precise control of the robotic arm's end effector's displacement, velocity, force, or posture at micron or even sub-micron levels.

[0190] This invention, through constructing a dynamic digital twin of the target spine, can realistically reflect the geometric and mechanical behavior of the spine in a dynamic environment, achieving adaptive response to the dynamic environment and laying a data foundation for subsequent robotic arm path planning. Optionally, this invention, by generating the planned motion path of the robotic arm corresponding to the spine, can dynamically compensate for real-time deformations of the spine during surgery (such as minute displacements caused by breathing, heartbeat, and muscle contraction), avoiding positioning deviations caused by anatomical structure movement. This invention, by generating the pose calibration command for the robotic arm, can significantly reduce trajectory tracking errors, improve motion control accuracy, thereby enhancing the robustness of robotic arm calibration and ensuring motion stability and reliability. This invention, by determining the stiffness coefficient, can further improve the stability and reliability of the robotic arm. The displacement compensation of the robotic arm can correct path errors caused by inertia, friction, and external disturbances during high-speed or high-load movements in real time, ensuring that the robotic arm moves strictly according to the planned path. In this embodiment, when the path error is less than a second threshold and not less than a preset safety threshold, the pulse voltage required for the corresponding micro-motion platform of the robotic arm can be calculated to achieve micro-displacement adjustment of the robotic arm, ensuring that the robotic arm does not overshoot or oscillate when approaching the target position, thereby achieving nanometer-level positioning accuracy. Finally, in this embodiment, by executing micrometer-level control of the robotic arm according to the micrometer-level calibration command, the displacement of the end effector of the robotic arm can be precisely controlled to achieve micrometer-level accuracy at the target position, thereby avoiding damage to surrounding tissues and improving surgical safety. Therefore, this invention can improve the control accuracy and safety of robotic arms in spinal surgery.

[0191] like Figure 2 The diagram shown is a flowchart illustrating a method for precise control of a spinal surgery robotic arm integrating deep learning and dynamic calibration, according to an embodiment of the present invention. In this embodiment, the method for precise control of a spinal surgery robotic arm integrating deep learning and dynamic calibration includes:

[0192] S1, acquire multi-modal data of a target object, wherein the multi-modal data comprises CT image data, three-dimensional point cloud data and instrument contact force data, to construct a dynamic spine digital twin of the target object;

[0193] S2, calculate a dynamic deformation field of a corresponding spine of the target object according to the dynamic spine digital twin, to generate a planned motion path of a corresponding mechanical arm of the spine;

[0194] S3, acquire a real-time motion path of the mechanical arm, and calculate a path error between the real-time motion path and the planned motion path, when the path error is not less than a preset first threshold, calculate a translation vector and a rotation vector of the mechanical arm, to generate a pose calibration instruction of the mechanical arm, and perform coarse-level regulation and control of the mechanical arm according to the pose calibration instruction;

[0195] S4, when the path error is less than the first threshold and not less than a preset second threshold, calculate a stiffness coefficient of the mechanical arm, determine a displacement compensation amount of the mechanical arm according to the stiffness coefficient, to generate an impedance calibration instruction of the mechanical arm, and perform fine-level regulation and control of the mechanical arm according to the impedance calibration instruction;

[0196] S5, when the path error is less than the second threshold and not less than a preset safety threshold, calculate a required pulse voltage of a micro-motion platform corresponding to the mechanical arm, to generate a micron-level calibration instruction of the mechanical arm, and perform micron-level regulation and control of the mechanical arm according to the micron-level calibration instruction.

[0197] In several embodiments provided in the present application, it should be understood that the provided system and method can be implemented in other manners. For example, the system embodiments described above are merely schematic. For example, the division of the modules is merely a logical function division. In actual implementation, another division manner can be adopted.

[0198] In addition, each function module in each embodiment of the present application can be integrated in a processing unit, or each unit can exist physically independently, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware, or in the form of hardware plus software function module.

[0199] Finally, it should be noted that in the above embodiments, each embodiment can be combined with or independent of each other, and deleting any one of them does not affect the technical implementation of the other embodiments. The above embodiments are only used to illustrate the technical solutions of the present application and are not limiting. Although the present application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or replaced by equivalents without departing from the spirit and scope of the present application.

Claims

1. A precision control system for a spinal surgery robotic arm integrating deep learning and dynamic calibration, characterized in that, The precision control system for spinal surgery robotic arms that integrates deep learning and dynamic calibration includes: A digital twin construction module is used to acquire multimodal data of a target object, wherein the multimodal data includes CT image data, three-dimensional point cloud data, and instrument contact force data, in order to construct a dynamic spinal digital twin of the target object; The motion path analysis module is used to calculate the dynamic deformation field of the spine corresponding to the target object based on the dynamic digital twin of the spine, so as to generate the planned motion path of the robotic arm corresponding to the spine. The coarse-level control module is used to acquire the real-time motion path of the robotic arm and calculate the path error between the real-time motion path and the planned motion path. When the path error is not less than a preset first threshold, the module calculates the translation vector and rotation vector of the robotic arm to generate the pose calibration command of the robotic arm. Based on the pose calibration command, the module executes the coarse-level control of the robotic arm. The precision control module is used to calculate the stiffness coefficient of the robotic arm when the path error is less than the first threshold and not less than the preset second threshold, determine the displacement compensation amount of the robotic arm based on the stiffness coefficient, generate the impedance calibration command of the robotic arm, and execute the precision control of the robotic arm based on the impedance calibration command. The micrometer-level control module is used to calculate the pulse voltage required by the micro-motion platform corresponding to the robotic arm when the path error is less than the second threshold and not less than the preset safety threshold, so as to generate the micrometer-level calibration command of the robotic arm and execute the micrometer-level control of the robotic arm according to the micrometer-level calibration command.

2. The precision control system for a spinal surgery robotic arm integrating deep learning and dynamic calibration as described in claim 1, characterized in that, When the path error is not less than a preset first threshold, the translation vector and rotation vector of the robotic arm are calculated, including: Based on the path error, the actual pose and planned pose of the robotic arm are determined; Calculate the translation vector of the robotic arm based on the actual pose and the planned pose; Determine the actual pose four elements of the actual pose and the planned pose four elements of the planned pose, respectively; The relative rotation quaternion of the robotic arm is calculated based on the actual pose quaternion and the planned pose quaternion. Based on the relative rotation quaternion, calculate the rotation angle and rotation unit vector of the robotic arm; The rotation vector of the robotic arm is determined based on the rotation angle and the rotation unit vector.

3. The precision control system for a spinal surgery robotic arm integrating deep learning and dynamic calibration as described in claim 1, characterized in that, When the path error is less than the first threshold and not less than a preset second threshold, the stiffness coefficient of the robotic arm is calculated, including: Extract the CT grayscale values ​​and bone-device contact force from the multimodal data corresponding to the robotic arm; Based on the CT grayscale values, analyze the current bone density of the spinal region corresponding to the robotic arm; The stiffness coefficient of the robotic arm is calculated based on the current bone density in the region, the bone-device contact force, and the path error.

4. The precision control system for a spinal surgery robotic arm integrating deep learning and dynamic calibration as described in claim 1, characterized in that, Determining the displacement compensation amount of the robotic arm based on the stiffness coefficient includes: Based on the stiffness coefficient, analyze the damping coefficient of the robotic arm; By combining the damping coefficient and the stiffness coefficient, an impedance model of the robotic arm is fitted. Based on the path error corresponding to the robotic arm, the static compensation amount of the robotic arm is calculated using the impedance model. Determine the respiratory motion pattern of the target object corresponding to the robotic arm; Based on the breathing motion pattern and the static compensation amount, the displacement compensation amount of the robotic arm is analyzed.

5. The precision control system for a spinal surgery robotic arm integrating deep learning and dynamic calibration as described in claim 1, characterized in that, When the path error is less than the second threshold and not less than a preset safety threshold, the step of calculating the pulse voltage required by the micro-motion platform corresponding to the robotic arm includes: Determine the mapping coefficients between the robotic arm's coordinate system and the micro-motion platform's coordinate system; Based on the path error and the mapping coefficient, the amount of displacement that the robotic arm needs to compensate is determined; The displacement-voltage linear relationship of the micro-motion platform is analyzed to determine the displacement-voltage conversion coefficient of the micro-motion platform; The pulse voltage of the micro-motion platform is determined based on the displacement-voltage conversion coefficient and the displacement amount.

6. The precision control system for a spinal surgery robotic arm integrating deep learning and dynamic calibration as described in claim 1, characterized in that, The construction of the dynamic spinal digital twin of the target object includes: Construct an image coordinate system for the CT image data corresponding to the target object and a surgical three-dimensional coordinate system for the three-dimensional point cloud data of the CT image data corresponding to the target object, respectively; Determine the transformation matrix between the image coordinate system and the surgical three-dimensional coordinate system; Based on the transformation matrix, spatiotemporal registration is performed on the multimodal data corresponding to the target object to obtain multimodal registration data; Multimodal features are extracted from the multimodal registration data to perform data fusion on the multimodal registration data, resulting in fused data. Based on the fused data, a three-dimensional model of the spine of the target object is fitted; The dynamic behavior of the three-dimensional spinal model is simulated to construct a dynamic digital twin of the spine.

7. The precision control system for a spinal surgery robotic arm integrating deep learning and dynamic calibration as described in claim 1, characterized in that, The step of calculating the dynamic deformation field of the spine corresponding to the target object based on the dynamic digital twin of the spine includes: The initial static equilibrium state of the spine is simulated through the dynamic digital twin of the spine. Calculate the initial displacement field of the initial static equilibrium state; Define the time step and dynamic load of the dynamic spinal digital twin; The dynamic changes of the spine are simulated based on the time step and the dynamic load. Based on the initial displacement field, analyze the displacement vector of the corresponding node of the spine in the dynamic change state; The dynamic deformation field of the spine is determined based on the displacement vector.

8. The precision control system for a spinal surgery robotic arm integrating deep learning and dynamic calibration as described in claim 1, characterized in that, The generation of the planned motion path for the robotic arm corresponding to the spine includes: Identify the workspace of the robotic arm; The dynamic deformation field of the spine is mapped to the workspace to obtain the mapped deformation field; Based on the mapped deformation field, the target position sequence and attitude sequence of the robotic arm are calculated; The planned motion path of the robotic arm is generated based on the target position sequence and the posture sequence.

9. The precision control system for a spinal surgery robotic arm integrating deep learning and dynamic calibration as described in claim 1, characterized in that, The calculation of the path error between the real-time motion path and the planned motion path includes: The real-time motion path and the planned motion path are respectively subjected to time-series discretization processing to obtain the real-time path sequence and the planned path sequence; Construct the real-time path vector set of the real-time path sequence and the planned path vector set of the planned path sequence respectively; Based on the real-time path vector set and the planned path vector set, calculate the position error, attitude error, and angle error of the real-time motion path and the planned motion path; The path error between the real-time motion path and the planned motion path is determined based on the position error, the attitude error, and the angle error.

10. The precision control system for a spinal surgery robotic arm integrating deep learning and dynamic calibration as described in claim 4, characterized in that, The step of calculating the static compensation amount of the robotic arm using the impedance model includes: The external force on the robotic arm is calculated using the impedance model. The static displacement of the robotic arm is calculated based on the external force on the robotic arm and the stiffness coefficient corresponding to the robotic arm. Based on the static displacement, the static compensation amount of the robotic arm is determined.

Images