Data fusion orthopedic operating room indoor positioning method and system

Abstract

The application belongs to the technical field of surgical navigation indoor positioning, and particularly relates to a data fusion orthopedic surgery indoor positioning method and system, which comprises the following steps: acquiring the acceleration and angular velocity of a surgical knife in a surgical area; calculating a step length according to the acceleration; constructing an incremental quaternion according to the angular velocity, rotating and combining the incremental quaternion to a previous attitude quaternion through quaternion multiplication to obtain a final attitude quaternion; correcting the final attitude quaternion by using the gravity direction measured by an accelerometer and the geomagnetic direction measured by a magnetometer to obtain a compensated attitude angle; deploying four same wireless positioning base stations, and obtaining the positioning data of the wireless positioning base stations on the surgical knife by using a multilateral positioning algorithm; and fusing the step length, the compensated attitude angle and the positioning data of the wireless positioning base stations by using an extended Kalman filter to iteratively calculate the positioning of the surgical knife. The application can maintain high-precision positioning and significantly improve the robustness in a complex environment of a surgical room.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of surgical navigation indoor positioning, and particularly relates to a data fusion orthopedic surgery indoor positioning method and system. BACKGROUND

[0002] With the rapid development of wireless network technology and intelligent sensing technology, dynamic real-time tracking of the target affected area in surgery has become one of the key problems that need to be solved in clinical surgical diagnosis and treatment. In the process of orthopedic surgery, the surgeon needs to achieve accurate positioning and continuous tracking of the surgical area (such as the fracture site, lesion point, and screw implantation point) through the navigation system without directly exposing all the bone structures.

[0003] At present, the orthopedic surgery navigation tracking methods widely used in clinical practice mainly include optical navigation, magnetic navigation, and navigation based on an inertial measurement unit (IMU). Among them, optical navigation relies on an external camera array and a reflective marker to achieve high-precision positioning, but it has strict requirements for the surgical field and is easily affected by occlusion, which can cause loss of positioning ability; magnetic navigation can achieve positioning under non-direct vision conditions, but the magnetic field is easily affected by metal instruments and electromagnetic interference in the surgical environment, which can cause a decrease in precision; the navigation system based on an inertial measurement unit can achieve portability and continuous tracking, but due to the cumulative error of inertial drift, the positioning accuracy is difficult to maintain at the millimeter level for a long time.

[0004] In actual clinical operations, orthopedic surgery needs to be performed under a very small surgical approach and a complex anatomical environment, and the position of the affected area can be slightly shifted due to breathing, traction, or external force operation. Therefore, the existing navigation technology still has deficiencies in dynamic real-time performance, millimeter-level precision maintenance, and anti-occlusion ability, and it is difficult to fully meet the needs of orthopedic fine surgery. SUMMARY

[0005] The application embodiment provides a data fusion orthopedic surgery indoor positioning method for achieving real-time positioning and tracking of the affected area with millimeter-level precision in a complex clinical environment, solving the deficiencies of the existing navigation technology in dynamic real-time performance, surgical precision maintenance, and anti-occlusion ability, and making it difficult to fully meet the needs of orthopedic fine surgery.

[0006] Another aspect of the application provides a data fusion orthopedic surgery indoor positioning system.

[0007] In order to achieve the above-mentioned purpose, the application adopts the following technical scheme:

[0008] A data fusion orthopedic surgery indoor positioning method, comprising:

[0009] Obtaining the acceleration and angular velocity of the surgical knife in the surgical area;

[0010] Calculating the step length according to the acceleration;

[0011] Based on the angular velocity, an incremental quaternion is constructed. The incremental quaternion is then rotated and compounded into the previous attitude quaternion through quaternion multiplication to obtain the final attitude quaternion.

[0012] The final attitude quaternion is corrected using the gravity direction measured by the accelerometer and the geomagnetic direction measured by the magnetometer to obtain the compensated attitude angle.

[0013] Four identical wireless positioning base stations were deployed, and the positioning data of the scalpel by the wireless positioning base stations was obtained using a multilateral positioning algorithm.

[0014] The scalpel positioning is obtained by fusing the step size, the compensated attitude angle, and the positioning data from the wireless positioning base station using an extended Kalman filter and iteratively calculating.

[0015] Furthermore, the step size is calculated based on the acceleration, including:

[0016] The resultant acceleration is obtained from the acceleration in the three directions.

[0017] Obtain the maximum and minimum values ​​of the resultant acceleration;

[0018] The step size is calculated based on the maximum and minimum values: ,in, For the first The maximum value of the resultant acceleration during the step. For the first The minimum value of the resultant acceleration during the step. Indicates the calibration coefficient. For the first The stride length of a step.

[0019] Furthermore, an incremental quaternion is constructed based on the angular velocity. This incremental quaternion is then multiplied and rotated to be combined with the previous attitude quaternion to obtain the final attitude quaternion, which includes:

[0020] Calculate the rotation increment angle and rotation axis based on the angular velocity and sampling period;

[0021] Construct incremental quaternions based on the rotation increment angle and rotation axis;

[0022] The incremental quaternion rotation is compounded to the previous attitude by quaternion multiplication to obtain the intermediate attitude without external observation correction.

[0023] The intermediate poses are normalized to obtain the final pose quaternion.

[0024] Furthermore, the final attitude quaternion is corrected using the gravity direction measured by the accelerometer and the geomagnetic direction measured by the magnetometer to obtain the compensated attitude angles, including:

[0025] The gravity direction measured by the accelerometer is used as the observation error to compensate for the gyroscope integration result. This includes: using a rotation matrix to transform the standard gravity vector from the geographic coordinate system to the coordinate system of the device where the scalpel is located to obtain the gravity vector.

[0026] The acceleration measured in the device coordinate system is normalized by the magnitude of the resultant acceleration to obtain the unit vector of acceleration.

[0027] Multiplying the gravity vector by the unit vector of acceleration yields the error vector of the gravity direction.

[0028] Furthermore, the geomagnetic direction measured by the magnetometer is used to compensate for errors in the heading, including:

[0029] The magnetic field strength measured by the magnetometer is transformed from the geographic coordinate system to the coordinate system of the device where the scalpel is located using a rotation matrix, so as to obtain the magnetic field strength in the device coordinate system;

[0030] Calculate the magnitude of the magnetic field strength measured by the magnetometer, normalize the magnetic field strength measured by the magnetometer based on the magnitude, and obtain the unit vector of the magnetic field strength.

[0031] Multiply the magnetic field strength in the device coordinate system by the unit vector of the magnetic field strength to obtain the magnetometer data after error compensation.

[0032] Furthermore, a PI controller is used to perform total error compensation based on the error vector in the direction of gravity and the magnetometer data after error compensation. The formula is expressed as: ,in, Indicates proportional gain. Indicates integral gain. The sum of the error vector representing the direction of gravity and the magnetometer data after error compensation. This represents the total error.

[0033] Furthermore, the step size, compensated attitude angle, and positioning data from the wireless positioning base station are fused using an extended Kalman filter, and the scalpel positioning is obtained through iterative calculation, including:

[0034] Establish the state equation and the observation equation;

[0035] The state equation and observation equation are linearized, the state transition matrix is ​​calculated, and the state transition matrix is ​​used to predict the state. During the prediction, the process noise covariance is used to calculate the prediction covariance and the prediction covariance is updated according to the Joseph covariance form.

[0036] The diagonal elements of the process noise covariance and observation noise covariance are adjusted using the hybrid tunicate Kepler optimization algorithm.

[0037] Furthermore, the predicted covariance is updated according to the Joseph covariance form, expressed as: ,in To predict covariance, For Kalman gain, For the observation matrix, To measure the noise covariance, Indicates the first At that moment, For transpose;

[0038] Kalman gain is calculated using innovative covariance: , The innovation covariance is represented by the calculation process, which includes: calculating the innovation covariance. innovation covariance Perform numerical corrections: And innovative covariance is corrected by diagonal loading or eigenvalue correction.

[0039] Furthermore, the diagonal elements of the process noise covariance and observation noise covariance are adjusted using a hybrid tunicate Keplerian optimization algorithm, including:

[0040] The fitness function is established by taking the minimization of the localization error as the optimization objective for the parameter vector to be optimized.

[0041] A local spiral search is performed using the tunicate search algorithm;

[0042] When the local spiral search fails to improve within a preset number of iterations, a cross-regional jump global search is performed based on the Kepler algorithm.

[0043] Repeat the local spiral search and global search, and apply the optimal parameter vector to the process noise covariance and observation noise covariance after each round of update.

[0044] Another aspect of this application provides a data fusion-based intraoperative positioning system for orthopedic operating rooms, comprising:

[0045] The data acquisition module is used to acquire the acceleration and angular velocity of the scalpel in the surgical area;

[0046] The step size calculation module is used to calculate the step size based on acceleration.

[0047] The attitude calculation module is used to construct an incremental quaternion based on the angular velocity, and then rotate and combine the incremental quaternion with the previous attitude quaternion through quaternion multiplication to obtain the final attitude quaternion.

[0048] The compensation module uses the gravity direction measured by the accelerometer and the geomagnetic direction measured by the magnetometer to correct the final attitude quaternion and obtain the compensated attitude angle.

[0049] The positioning module deploys four identical wireless positioning base stations and uses a multilateral positioning algorithm to obtain the positioning data of the scalpel by the wireless positioning base stations;

[0050] The parameter adaptive optimization module is used to fuse the step size, the compensated attitude angle, and the positioning data from the wireless positioning base station using an extended Kalman filter, and iteratively calculate the scalpel positioning.

[0051] Compared with the prior art, this application has the following beneficial effects: the application can maintain high-precision positioning even in the presence of obstruction, complex electromagnetic environment or multipath interference, which significantly improves the robustness in the complex environment of the operating room; it effectively solves the complementary problem of the cumulative error of inertial sensor drift and the instantaneous fluctuation of wireless positioning base station, and achieves complementary advantages through multi-source data fusion. Attached Figure Description

[0052] Figure 1 A flowchart illustrating a data fusion-based intraoperative positioning method for orthopedic operating rooms, provided as an embodiment of this application;

[0053] Figure 2 A structural block diagram of a data fusion-based orthopedic operating room positioning system provided in this application embodiment;

[0054] Figure 3 A comparison diagram of the actual trajectory and the estimated trajectory provided in the embodiments of this application;

[0055] Figure 4 The convergence curve provided for the embodiments of this application. Detailed Implementation

[0056] To make the technical solution of this application clearer, the following will describe this application in further detail with reference to the accompanying drawings in the embodiments of this application. The embodiments here are only used to explain a part of this application, not all of it.

[0057] One embodiment of this application proposes a hardware structure for a data fusion-based intraoperative positioning method in orthopedic operating rooms, comprising: an inertial measurement unit (IMU): including an accelerometer, gyroscope, and magnetometer, with a sampling frequency of 100Hz, used to acquire real-time acceleration, angular velocity, and magnetic field data of the scalpel. It can be a miniaturized structure and fixed to the scalpel, or it can be embedded in the scalpel handle. Ultra-wideband wireless positioning base stations: four ultra-wideband wireless positioning base stations are deployed at the four corners of the operating room, with tags installed on the scalpel handle.

[0058] The host computer acquires data from the inertial measurement unit and the ultra-wideband wireless positioning base station, and fuses the data from the inertial sensor and the ultra-wideband wireless positioning base station using an extended Kalman filter (EKF). Through a parameter adaptive optimization module, a hybrid search-by-analyze (SSA+KOA) algorithm is implemented, adjusting the positive definite enhancement introduced during the extended Kalman filter filtering process in real time. This achieves a global search for the logarithmic parameter space of the noise covariance and the observation noise covariance using radial tangential spiral perturbation and jump operations.

[0059] See Figure 1 As shown in the embodiment of this application, a data fusion-based intraoperative positioning method for orthopedic operating rooms includes:

[0060] S1 obtains the acceleration and angular velocity of the scalpel in the surgical area;

[0061] The method of acquisition is to use inertial sensors to obtain the acceleration and angular velocity of the surgical area in real time to guide the scalpel tracking and positioning. The acceleration is obtained through an accelerometer, and the acquired acceleration includes acceleration values ​​in three axes. The angular velocity is obtained through a gyroscope.

[0062] S2 calculates the step size based on acceleration; including:

[0063] The resultant acceleration is obtained from the acceleration in the three directions.

[0064] Obtain the maximum and minimum values ​​of the resultant acceleration;

[0065] The step size is calculated based on the maximum and minimum values: ,in, For the first The maximum value of the resultant acceleration during the step. For the first The minimum value of the resultant acceleration during the step. Indicates the calibration coefficient. For the first The stride length of a step.

[0066] Acceleration is obtained by summing and averaging the accelerations in three directions, and the resultant acceleration is expressed as: In the formula, These represent the output values ​​of linear acceleration in three different directions. By analyzing the relationship between acceleration statistical characteristics such as step size and step frequency, a nonlinear step size estimation model based on the acceleration amplitude difference is established, namely: .

[0067] S3 constructs an incremental quaternion based on the angular velocity, and then rotates and combines the incremental quaternion with the previous attitude quaternion through quaternion multiplication to obtain the final attitude quaternion.

[0068] S4 uses the gravity direction measured by the accelerometer and the geomagnetic direction measured by the magnetometer to correct the final attitude quaternion and obtain the compensated attitude angle.

[0069] The S5 deploys four identical wireless positioning base stations and uses a multilateral positioning algorithm to obtain the positioning data of the scalpel by the wireless positioning base stations.

[0070] S6 uses an extended Kalman filter to fuse the step size, the compensated attitude angle, and the positioning data from the wireless positioning base station, and iteratively calculates the scalpel positioning.

[0071] In one embodiment, an incremental quaternion is constructed based on the angular velocity, and the incremental quaternion is rotated and compounded to the previous attitude quaternion through quaternion multiplication to obtain the final attitude quaternion, including:

[0072] Calculate the rotation increment angle and rotation axis based on the angular velocity and sampling period;

[0073] Construct incremental quaternions based on the rotation increment angle and rotation axis;

[0074] The incremental quaternion rotation is compounded to the previous attitude by quaternion multiplication to obtain the intermediate attitude without external observation correction.

[0075] The intermediate poses are normalized to obtain the final pose quaternion.

[0076] This step is heading estimation, which involves constructing an incremental quaternion using angular velocity, converting it into Euler angles, and then obtaining the attitude angles. , is represented as: In the formula, It belongs to the scalar (real part). It is a vector (imaginary part). This represents the roll angle obtained after updating using quaternions. This represents the pitch angle obtained after updating using quaternions. This represents the heading angle (yaw) obtained after updating using quaternions.

[0077] In one embodiment, calculating the rotation increment angle and rotation axis based on the angular velocity and sampling period includes: gyroscope integral update: assuming the gyroscope is at the angular velocity and sampling period... Each sampling time in the device coordinate system Angular velocity measured below (Unit is) ), , and The angular velocities are for the three axes, and the sampling period is... ; denote the rotation increment angle Rotation axis ,like Then take , This represents the magnitude of the angular velocity. (Superscript here) This indicates transpose, used to write a row vector as a column vector.

[0078] Construct an incremental quaternion based on the rotation increment angle and rotation axis, expressed by the following formula: For quaternions, a scalar-first order is used. Quaternion multiplication is then performed. Composite the incremental quaternion rotation to the previous pose. The intermediate attitude without external observation correction is obtained. To eliminate numerical errors, intermediate attitudes should be addressed. After normalization, the final pose quaternion is obtained. ,in, This represents an incremental quaternion.

[0079] In one embodiment, the final attitude quaternion is corrected using the gravity direction measured by the accelerometer and the geomagnetic direction measured by the magnetometer to obtain the compensated attitude angle, including:

[0080] The gravity direction measured by the accelerometer is used as the observation error to compensate for the gyroscope integration result. This includes: using a rotation matrix to transform the standard gravity vector from the geographic coordinate system to the coordinate system of the device where the scalpel is located to obtain the gravity vector.

[0081] The acceleration measured in the device coordinate system is normalized by the magnitude of the resultant acceleration to obtain the unit vector of acceleration.

[0082] Multiplying the gravity vector by the unit vector of acceleration yields the error vector of the gravity direction.

[0083] The direction of gravity measured by the accelerometer is used as the observation error to compensate for the error in the gyroscope integration result. This reflects the deviation between the attitude obtained by the gyroscope integration and the actual measurement by the accelerometer, i.e., the error vector of the direction of gravity, which is expressed as follows: ,in, It is the standard gravity vector Using rotation matrix Depend on Switch to (geographic coordinate system) The gravity vector obtained from the system (equipment coordinate system, i.e., the coordinate system of the device where the scalpel is located), The unit vector representing acceleration is the unit vector of acceleration measured in the equipment coordinate system after magnitude normalization. It represents the estimate of the direction of gravity in the equipment coordinate system.

[0084] The accelerations along the three axes are as follows: The magnitude of the resultant acceleration is The normalized unit vector of acceleration is: .

[0085] In one embodiment, the geomagnetic direction measured by the magnetometer is used to compensate for errors in the heading, including:

[0086] The magnetic field strength measured by the magnetometer is transformed from the geographic coordinate system to the coordinate system of the device where the scalpel is located using a rotation matrix, so as to obtain the magnetic field strength in the device coordinate system;

[0087] Calculate the magnitude of the magnetic field strength measured by the magnetometer, normalize the magnetic field strength measured by the magnetometer based on the magnitude, and obtain the unit vector of the magnetic field strength.

[0088] Multiply the magnetic field strength in the device coordinate system by the unit vector of the magnetic field strength to obtain the magnetometer data after error compensation.

[0089] The heading error is compensated using the geomagnetic direction measured by the magnetometer, reflecting the deviation between the attitude angle obtained by gyroscope integration and the heading actually measured by the magnetometer, as shown below: ,in, Indicates using rotation matrix in The magnetic field strength obtained under the system. This represents the unit vector of the magnetometer after normalization. This represents the magnetometer data after error compensation;

[0090] Using rotation matrix in The magnetic field strength obtained under the system is expressed as: , express Magnetic field strength in the direction, for Magnetic field strength in the direction, for The magnetic field strength in the direction of the magnetometer is marked as follows: The modulus is The normalized unit vector of the magnetometer is: .

[0091] A PI controller is used to perform total error compensation based on the error vector in the direction of gravity and the magnetometer data after error compensation. The formula is expressed as: ,in, Indicates proportional gain. Indicates integral gain. The sum of the error vector representing the direction of gravity and the magnetometer data after error compensation. This is the compensation amount generated after processing by the PI controller, i.e., the total error.

[0092] The compensation amount is added to the gyroscope angular velocity as a virtual angular velocity term: ,

[0093] And update the final pose quaternion:

[0094] ,

[0095] in The direction of rotation axis The component in the direction of the unit rotation axis. This indicates the angular velocity of the gyroscope before compensation. The direction is the rotation axis. In this way, the final attitude quaternion automatically includes the error information of the accelerometer and magnetometer during the integration update, thereby compensating for gyroscope drift.

[0096] The attitude angles are extracted from the corrected final attitude quaternion for subsequent calculations. The extraction process is a conventional method and will not be described in detail here.

[0097] In one embodiment, four identical wireless positioning base stations are deployed, and a multilateral positioning algorithm is used to obtain the positioning data of the scalpel from the wireless positioning base stations; the wireless positioning base stations are arranged in the operating room. The distance from the scalpel to the wireless positioning base station is expressed as: , This represents the location coordinates of the wireless positioning base station, where , Indicates the position of the scalpel. It refers to a scalpel and the first Ranging values ​​between wireless positioning base stations.

[0098] Based on the formula for the distance between two points, substituting the coordinates of each wireless positioning base station and the distance between the wireless positioning base station and the scalpel into the equations, we obtain the following system of equations: Using the second to the third equations in the system Subtracting each equation from the first equation yields the following form: ,

[0099] in, , , ;

[0100] Each row in the matrix represents the first row. The first wireless positioning base station and the first wireless positioning base station are in direction and Twice the difference in coordinates of the direction is used to linearize the nonlinear ranging equation. This represents the unknown two-dimensional position coordinates of the scalpel, i.e., the position of the scalpel to be estimated. Each item in the matrix represents the first wireless base station and the second... The squared difference of the ranging distances from each wireless base station, plus the squared difference of the coordinates, is used to solve the right-hand vector of the linear equation for the position of the scalpel.

[0101] Using the least squares method to find The solution yields the coordinates of the scalpel position. .

[0102] In one embodiment, the scalpel positioning is obtained by fusing the step size, compensated attitude angle, and positioning data from the wireless positioning base station using an extended Kalman filter and iteratively calculating the scalpel positioning, including:

[0103] Establish the state equation and the observation equation;

[0104] The state equation and observation equation are linearized, the state transition matrix is ​​calculated, and the state transition matrix is ​​used to predict the state. During the prediction, the process noise covariance is used to calculate the prediction covariance and the prediction covariance is updated according to the Joseph covariance form.

[0105] The diagonal elements of the process noise covariance and observation noise covariance are adjusted using the hybrid tunicate Kepler optimization algorithm.

[0106] The state equation is expressed as:

[0107] , ,

[0108] ;

[0109] The observation equation is: , ;

[0110] Among them, input quantity , Indicates the first The step size is calculated based on the acceleration within each sampling period; Indicates the first Within each sampling period, respectively direction and The acceleration component in the direction, Angular velocity; express Attitude angle (orientation angle) within each sampling period. express Attitude angle within each sampling period, For attitude angle, This represents the location coordinates of the wireless positioning base station, where , Representing the The position of the scalpel at -1 moment. Representing the -1 moment's scalpel speed, Indicates the first The theoretical distance from a wireless positioning base station to a scalpel Indicates the first The wireless positioning base station is in the first The distance measurement value at each sampling time. Indicates the sampling period. Indicates the first The state quantity at time -1 represents the two-dimensional position of the scalpel. ,speed and attitude angle . express -1 is the state value at time step -1, representing the state of the scalpel at the previous time step. Indicates the first Observed variables at each time point Indicates the first The first wireless positioning base station Observed variables at each time point Indicates the first The system process noise introduced at time -1 is used to describe the uncertainty of the system model; The statistical characteristics are represented by a mean of 0 and a process noise covariance matrix of... Gaussian distribution; This represents the measurement noise vector, used to describe the measurement error of the wireless positioning base station; For statistical properties, let represent a mean of 0 and an observation noise covariance of . The Gaussian distribution. This represents the actual state of the system; each filter update provides an estimate of this state. With error covariance .

[0111] The above equations are linearized, and the state transition matrix is ​​calculated. Simultaneously, state prediction is performed on the state transition matrix. During prediction, the process noise covariance is used to calculate the predicted covariance, which is updated in real time. To ensure numerical stability, Joseph covariance updates are used, and symmetry and eigenvalue lower bound constraints are applied to the innovation covariance. Since the Kalman gain depends on both the measurement noise covariance and the predicted covariance, a hybrid tunic Keplerian optimization algorithm is used to adjust the diagonal elements of the process noise covariance and the observation noise covariance, thereby changing the growth and convergence behavior of the state covariance. Simultaneously, the diagonal elements of the measurement noise covariance are optimized, indirectly affecting the Kalman gain calculation and state / covariance update. The parameter vector to be optimized is defined as follows:

[0112] , , , This represents the diagonal weight parameter. Represents process noise covariance The diagonal elements of the matrix correspond to position, velocity, and angular noise. Represents the observation noise covariance The diagonal elements of the matrix, corresponding to Ranging noise of each base station.

[0113] When updating measurements using the Extended Kalman Filter (EKF), the predicted covariance is updated in the form of Joseph covariance: ,in To predict covariance, For Kalman gain, For the observation matrix, To measure the noise covariance, This represents the identity matrix, used to ensure dimensional consistency in matrix operations; and is also used in calculating the innovation covariance. Afterwards, Perform numerical correction processing And, if necessary, ensure that the innovation covariance is strictly positive definite by diagonal loading or eigenvalue correction, that is, when the innovation covariance is minimized. Less than the preset threshold season , Small positive numbers (e.g.) To ensure numerical stability and the feasibility of subsequent matrix decomposition, Represent the identity matrix, and guarantee All eigenvalues ​​are positive, ensuring that the matrix is ​​strictly positive definite.

[0114] In one embodiment, the diagonal elements of the process noise covariance and the observation noise covariance are adjusted using a hybrid tunicate Keplerian optimization algorithm, including:

[0115] The fitness function is established by taking the minimization of the localization error as the optimization objective for the parameter vector to be optimized.

[0116] A local spiral search is performed using the tunicate search algorithm;

[0117] When the local spiral search fails to improve within a preset number of iterations, a cross-regional jump global search is performed based on the Kepler algorithm.

[0118] Repeat the local spiral search and global search, and apply the optimal parameter vector to the process noise covariance and observation noise covariance after each round of update.

[0119] The fitness function is defined based on the parameter vector to be optimized, with minimizing the localization error as the optimization objective: ,in It is a predicted location. It is the actual location. Represents the set of real reference variables; Indicates the population size.

[0120] Based on the search algorithm for tunicates, an orthogonal spiral local trajectory perturbation is performed near the logarithmic domain of the current solution. This perturbation includes "radial components pointing towards the optimum and tangential components orthogonal to it," and is further modulated exponentially with each iteration. The noise weighting coefficients are then fine-tuned to obtain the optimal solution. The parameter update formula for the search algorithm for tunicates is: ,in, For the search step size, It is an exponential function. It is a spiral growth factor. For angle variables, The norm is the difference between the global optimal solution and the individual solutions. The unit vector is , A random unit vector orthogonal to the unit vector. For the radial component pointing towards the optimal solution, The tangential component is orthogonal to the direction of the optimal solution. This represents the optimal parameter solution obtained by the tunic search algorithm in the (t+1)th iteration.

[0121] When the local spiral search fails to improve the local search within a preset number of iterations, a cross-regional jump global search is performed based on the Kepler algorithm to achieve local-global switching and avoid getting trapped in local minima. The global jump update formula of the Kepler algorithm is as follows:

[0122] in, This is the jump scaling factor. For the solution randomly sampled within the parameter search space, This represents the optimal solution randomly sampled within the parameter search space.

[0123] The search algorithm for the tunicate is used for most iterations. When there is no improvement after several rounds, the Kepler algorithm is triggered to escape the local optimum. The two-stage search is repeated, and the optimal solution is applied in real time to the process noise covariance of the extended Kalman filter after each round of update. Covariance of observation noise This allows for dynamic adjustment of filter parameters and improvement of positioning accuracy.

[0124] Perform the following steps within each iteration cycle:

[0125] A fine-grained search is performed near the optimal solution using the tunic search algorithm;

[0126] Check if the fitness function has been improved;

[0127] If no improvement is made, then the Kepler algorithm global jump search is performed;

[0128] The optimal solution is updated to the extended Kalman filter and the next iteration is performed.

[0129] By using the Kepler hybrid optimization algorithm of the sea squirt, the process noise covariance and observation noise covariance parameters in the extended Kalman filter (EKF) are dynamically adjusted, real-time adaptive optimization of the filter parameters is achieved, avoiding the degradation of filtering performance caused by the traditional fixed parameter method.

[0130] On the other hand, see Figure 2 As shown, this application provides a data fusion-based orthopedic operating room positioning system, comprising:

[0131] The data acquisition module is used to acquire the acceleration and angular velocity of the scalpel in the surgical area;

[0132] The step size calculation module is used to calculate the step size based on acceleration.

[0133] The attitude calculation module is used to construct an incremental quaternion based on the angular velocity, and then rotate and combine the incremental quaternion with the previous attitude quaternion through quaternion multiplication to obtain the final attitude quaternion.

[0134] The compensation module uses the gravity direction measured by the accelerometer and the geomagnetic direction measured by the magnetometer to correct the final attitude quaternion and obtain the compensated attitude angle.

[0135] The positioning module deploys four identical wireless positioning base stations and uses a multilateral positioning algorithm to obtain the positioning data of the scalpel by the wireless positioning base stations;

[0136] The parameter adaptive optimization module is used to fuse the step size, the compensated attitude angle, and the positioning data from the wireless positioning base station using an extended Kalman filter, and iteratively calculate the scalpel positioning.

[0137] This application is applied to an orthopedic positioning surgical system dominated by an ultra-wideband wireless positioning base station. It achieves seamless collaborative positioning with inertial sensors through parameter adaptive optimization, ensuring high accuracy and robustness. Multi-source data provides high-frequency attitude and motion acceleration information, as well as global spatial location information of the affected area. An extended Kalman filter framework is employed to achieve dynamic state estimation of the affected area's position, effectively compensating for the shortcomings of a single sensor in terms of accuracy attenuation and accumulated error. A hybrid tunicate Keplerian optimization algorithm is introduced, using real-time error as the optimization objective, and updating parameters with the logarithmic domain of the diagonal elements of the adaptive noise covariance, ensuring the filter maintains optimal state estimation accuracy during surgery. This not only significantly improves the accuracy and stability of the positioning results but also enhances the system's robustness and adaptability in complex environments. See also... Figure 3 As shown, the actual trajectory is highly consistent with the trajectory estimated in this application, the positioning error is significantly reduced, and the root mean square error reaches the millimeter level, verifying the high accuracy of this application. Figure 4 To optimize the convergence curve of the process, as the number of iterations increases, the root mean square error decreases rapidly and tends to stabilize, indicating that this application has good convergence and robustness.

[0138] Finally, it should be noted that the above description is only a preferred embodiment of this application and is not intended to limit this application. Although this application has been described in detail with reference to the embodiments, any modifications, equivalent substitutions and improvements made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A data fusion based intraoperative localization method for orthopedic surgery, characterized in that, The method comprises the following steps: Obtaining the acceleration and angular velocity of the surgical knife in the surgical area; Calculating the step length according to the acceleration; Constructing an incremental quaternion according to the angular velocity, and rotating and combining the incremental quaternion to the previous attitude quaternion through quaternion multiplication to obtain a final attitude quaternion; Correcting the final attitude quaternion by using the gravity direction measured by the accelerometer and the geomagnetic direction measured by the magnetometer to obtain a compensated attitude angle; Deploying four identical wireless positioning base stations, and obtaining the positioning data of the surgical knife by the wireless positioning base stations through a multilateration algorithm; Fusing the step length, the compensated attitude angle and the positioning data of the wireless positioning base stations by using an extended Kalman filter to iteratively calculate the positioning of the surgical knife; Calculating the step length according to the acceleration, comprising: Calculating the resultant acceleration according to the acceleration in three directions; Obtaining the maximum and minimum values of the resultant acceleration; The step size is calculated from the maximum and minimum values as: wherein is the maximum value of the combined acceleration in step is the minimum value of the combined acceleration in step denotes a calibration factor, is the step size of step is the step size of step denotes a calibration factor, is the step size of step Constructing an incremental quaternion according to the angular velocity, and rotating and combining the incremental quaternion to the previous attitude quaternion through quaternion multiplication to obtain a final attitude quaternion, comprising: Calculating a rotation incremental angle and a rotation axis according to the angular velocity and the sampling period; Constructing an incremental quaternion according to the rotation incremental angle and the rotation axis; Rotating and combining the incremental quaternion to the previous attitude through quaternion multiplication to obtain an intermediate attitude which is not corrected by external observation; Unitizing the intermediate attitude to obtain a final attitude quaternion; Correcting the final attitude quaternion by using the gravity direction measured by the accelerometer and the geomagnetic direction measured by the magnetometer to obtain a compensated attitude angle, comprising: Using the gravity direction measured by the accelerometer as an observation error to compensate the error of the gyroscope integral result, comprising: converting the standard gravity vector from the geographic coordinate system to the device coordinate system of the surgical knife by using a rotation matrix to obtain a gravity vector; After the acceleration measured in the device coordinate system is normalized by the modulus value of the resultant acceleration, the unit vector of the acceleration is obtained; The error vector of the gravity direction is obtained by multiplying the gravity vector and the unit vector of the acceleration; The geomagnetic direction measured by the magnetometer is used to compensate the error of the heading, comprising: Converting the magnetic field strength measured by the magnetometer from the geographic coordinate system to the device coordinate system of the surgical knife by using a rotation matrix to obtain the magnetic field strength in the device coordinate system; Calculating the modulus value of the magnetic field strength measured by the magnetometer, and normalizing the magnetic field strength measured by the magnetometer according to the modulus value to obtain the unit vector of the magnetic field strength; The error-compensated magnetometer data is obtained by multiplying the magnetic field strength in the device coordinate system and the unit vector of the magnetic field strength; The PI controller is used to compensate the total error according to the error vector of gravity direction and the compensated magnetometer data, which is expressed as: wherein, represents the proportional gain, represents the integral gain, represents the sum of the error vector of gravity direction and the compensated magnetometer data, is the total error.

2. The data fusion based intraoperative localization method for orthopedic surgery of claim 1, wherein, Fusing the step length, the compensated attitude angle and the positioning data of the wireless positioning base stations by using an extended Kalman filter to iteratively calculate the positioning of the surgical knife, comprising: Establishing a state equation and an observation equation; Linearizing the state equation and the observation equation, calculating a state transition matrix, and predicting the state by using the state transition matrix, wherein the process noise covariance is used to calculate the prediction covariance and update the prediction covariance in the Joseph covariance form; Adjusting the diagonal elements of the process noise covariance and the observation noise covariance by using the hybrid tentacled sea slug Kepler optimization algorithm.

3. The data fusion based intraoperative localization method for orthopedic surgery of claim 2, wherein, The prediction covariance is updated in accordance with the Joseph covariance form, denoted as: where is the prediction covariance, is the Kalman gain, is the observation matrix, is the measurement noise covariance, denotes the time instant, is the transpose; Kalman gain is calculated using innovative covariance: , The innovation covariance is represented by the calculation process, which includes: calculating the innovation covariance. innovation covariance Perform numerical corrections: And innovative covariance is corrected by diagonal loading or eigenvalue correction.

4. The data fusion based intra-operative positioning method for orthopedic surgery of claim 2, wherein, Adjusting the diagonal elements of the process noise covariance and the observation noise covariance by using the hybrid tentacled sea slug Kepler optimization algorithm, comprising: A fitness function is established by taking a parameter vector to be optimized as an optimization objective to minimize positioning error; A local spiral search is performed by using a sea squirt search algorithm; When the local spiral search fails to obtain improvement within a preset number of iterations, a cross-region jump global search is performed based on a Kepler algorithm; The local spiral search and the global search are repeated, and the optimal parameter vector is applied to process noise covariance and observation noise covariance after each round of update.

5. A data fusion based orthopedic operating room indoor positioning system for performing a data fusion based orthopedic operating room indoor positioning method according to any one of claims 1 to 4, characterized in that, Comprise: A data acquisition module for acquiring acceleration and angular velocity of a surgical knife in a surgical area; A step length calculation module for calculating step length according to acceleration; A posture calculation module for constructing an incremental quaternion according to angular velocity, rotating and compounding the incremental quaternion to a previous attitude quaternion through quaternion multiplication to obtain a final attitude quaternion; A compensation module for correcting the final attitude quaternion by using a gravity direction measured by an accelerometer and a geomagnetic direction measured by a magnetometer to obtain a compensated attitude angle; A positioning module for deploying four identical wireless positioning base stations and obtaining positioning data of the surgical knife by the wireless positioning base stations through a multilateral positioning algorithm; A parameter self-adaptive optimization module for fusing step length, the compensated attitude angle and the positioning data of the wireless positioning base stations by using an extended Kalman filter to iteratively calculate the positioning of the surgical knife.

Images