CN118217012B - Magnetic control intervention control method, system and storage medium based on robot guide wire - Google Patents
Magnetic control intervention control method, system and storage medium based on robot guide wire Download PDFInfo
- Publication number
- CN118217012B CN118217012B CN202410650722.5A CN202410650722A CN118217012B CN 118217012 B CN118217012 B CN 118217012B CN 202410650722 A CN202410650722 A CN 202410650722A CN 118217012 B CN118217012 B CN 118217012B
- Authority
- CN
- China
- Prior art keywords
- magnetic control
- model
- guide wire
- blood vessel
- robot
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 53
- 210000004204 blood vessel Anatomy 0.000 claims abstract description 143
- 238000013528 artificial neural network Methods 0.000 claims abstract description 51
- 238000013507 mapping Methods 0.000 claims abstract description 13
- 238000004806 packaging method and process Methods 0.000 claims abstract description 11
- 238000012858 packaging process Methods 0.000 claims abstract description 6
- 238000004364 calculation method Methods 0.000 claims description 32
- 230000003044 adaptive effect Effects 0.000 claims description 24
- 230000000007 visual effect Effects 0.000 claims description 21
- 238000005457 optimization Methods 0.000 claims description 18
- 238000002583 angiography Methods 0.000 claims description 17
- 230000005540 biological transmission Effects 0.000 claims description 9
- 238000010276 construction Methods 0.000 claims description 7
- 238000007781 pre-processing Methods 0.000 claims description 3
- 230000002792 vascular Effects 0.000 description 21
- 238000005516 engineering process Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 230000000694 effects Effects 0.000 description 3
- 230000017531 blood circulation Effects 0.000 description 2
- 230000000295 complement effect Effects 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 238000003062 neural network model Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000000903 blocking effect Effects 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 238000013152 interventional procedure Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000005855 radiation Effects 0.000 description 1
- 238000012800 visualization Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B34/00—Computer-aided surgery; Manipulators or robots specially adapted for use in surgery
- A61B34/20—Surgical navigation systems; Devices for tracking or guiding surgical instruments, e.g. for frameless stereotaxis
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B34/00—Computer-aided surgery; Manipulators or robots specially adapted for use in surgery
- A61B34/30—Surgical robots
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61M—DEVICES FOR INTRODUCING MEDIA INTO, OR ONTO, THE BODY; DEVICES FOR TRANSDUCING BODY MEDIA OR FOR TAKING MEDIA FROM THE BODY; DEVICES FOR PRODUCING OR ENDING SLEEP OR STUPOR
- A61M25/00—Catheters; Hollow probes
- A61M25/01—Introducing, guiding, advancing, emplacing or holding catheters
- A61M25/0105—Steering means as part of the catheter or advancing means; Markers for positioning
- A61M25/0116—Steering means as part of the catheter or advancing means; Markers for positioning self-propelled, e.g. autonomous robots
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B34/00—Computer-aided surgery; Manipulators or robots specially adapted for use in surgery
- A61B34/20—Surgical navigation systems; Devices for tracking or guiding surgical instruments, e.g. for frameless stereotaxis
- A61B2034/2046—Tracking techniques
- A61B2034/2051—Electromagnetic tracking systems
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B34/00—Computer-aided surgery; Manipulators or robots specially adapted for use in surgery
- A61B34/20—Surgical navigation systems; Devices for tracking or guiding surgical instruments, e.g. for frameless stereotaxis
- A61B2034/2046—Tracking techniques
- A61B2034/2065—Tracking using image or pattern recognition
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B34/00—Computer-aided surgery; Manipulators or robots specially adapted for use in surgery
- A61B34/30—Surgical robots
- A61B2034/301—Surgical robots for introducing or steering flexible instruments inserted into the body, e.g. catheters or endoscopes
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Theoretical Computer Science (AREA)
- Biomedical Technology (AREA)
- Surgery (AREA)
- Molecular Biology (AREA)
- Heart & Thoracic Surgery (AREA)
- Animal Behavior & Ethology (AREA)
- Software Systems (AREA)
- Public Health (AREA)
- Veterinary Medicine (AREA)
- General Physics & Mathematics (AREA)
- Biophysics (AREA)
- Robotics (AREA)
- Medical Informatics (AREA)
- Evolutionary Computation (AREA)
- Mathematical Physics (AREA)
- General Engineering & Computer Science (AREA)
- Computing Systems (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- Computational Linguistics (AREA)
- Pulmonology (AREA)
- Anesthesiology (AREA)
- Hematology (AREA)
- Geometry (AREA)
- Computer Graphics (AREA)
- Manipulator (AREA)
Abstract
The invention relates to the technical field of intelligent medical treatment, in particular to a magnetic control intervention control method, a magnetic control intervention control system and a magnetic control intervention control storage medium based on a robot guide wire, which comprise the following steps: packaging the mapping relation between the magnetic control data and the blood vessel shape data on a magnetic control intervention twin platform by using a neural network to obtain a control model for outputting the magnetic control data according to the blood vessel shape data; self-adaptive learning is added in the packaging process of the control model, and the performance of the control model is self-adaptively optimized, so that the magnetic control data output by the control model reaches the optimal control of the robot guide wire. According to the invention, the operation model of the magnetic control data is packaged on the digital twin platform through the neural network, the operation and intervention traveling of the magnetic control data are mutually independent, the intervention process is prevented from being interrupted, the safety of the guide wire intervention process is ensured, and the control model ensures the operation efficiency of the magnetic control data and simultaneously ensures the operation accuracy of the magnetic control data to be higher.
Description
Technical Field
The invention relates to the technical field of intelligent medical treatment, in particular to a magnetic control intervention control method, a magnetic control intervention control system and a storage medium based on a robot guide wire.
Background
In the current medical environment, magnetic navigation technology is increasingly applied to medical practice. In the prior art, magnetic navigation techniques have been applied to interventional procedures. For example: in interventional operation, a magnetic control device (tip) of a magnetic navigation technology is provided with an SLAM system, after a blood vessel is penetrated by a guide wire through a visual sensor and a pressure sensor for sensing the wall of the blood vessel or a laser sensor (applicable to a medical blood vessel), the pressure of the wall of the blood vessel, the blood flow speed and the resistance of the blood vessel are sensed, the blood vessel can be assisted to build a blood vessel running path and a map construction corresponding to sensing obstacle, the visual sensor is assisted to build a blood vessel path diagram, and based on imaging scanning before and during operation, a clinician is assisted to operate and judge more accurately according to parameters (the pressure of the wall of the blood vessel, the obstacle, the blood flow speed and the like) given by the SLAM sensor.
Therefore, the magnetic navigation technology can realize interventional treatment by controlling a magnetic field to drive a robot guide wire to walk in a blood vessel to reach an affected part by using a computer interface, so that the precision of fine interventional operation is greatly improved, and meanwhile, the radiation dose of a patient and medical staff in the operation process can be greatly reduced. The magnetic navigation control effect of the guide wire directly determines the operation effect of the interventional operation.
In the prior art, magnetic control data of magnetic navigation control of a guide wire in an intervention process are generally calculated only by utilizing a blood vessel image acquired before operation, and although the calculation of the magnetic control data does not need to be operated synchronously in the intervention process, i.e. the intervention process is not interrupted, the calculated magnetic control data of the guide wire is difficult to adapt to the real-time condition of blood vessels in the intervention process, such as the occurrence of the real-time conditions of narrowing, blocking and the like, and therefore, the magnetic control data can be calculated by utilizing the blood vessel image before operation, so that the magnetic control accuracy of the guide wire is insufficient, the actual advancing requirement of the guide wire is difficult to be met, and the intervention treatment effect is influenced.
Disclosure of Invention
The invention aims to provide a magnetic control interventional control method based on a robot guide wire, which aims to solve the technical problem that in the prior art, the magnetic control real-time accuracy of the guide wire is insufficient due to the fact that magnetic control data are calculated only by using preoperative blood vessel images.
In order to solve the technical problems, the invention specifically provides the following technical scheme:
a magnetic control intervention control method based on a robot guide wire comprises the following steps:
according to angiography images and a robot guide wire, a magnetically controlled intervention twin platform is built on the digital twin platform, wherein the structure of the robot guide wire comprises: the rear end of the robot guide wire is an unmagnetized part, the front end of the robot guide wire is a magnetized part, and a visual sensor is loaded at the front end of the magnetized part;
Determining the steering angle of the robot guide wire according to the blood vessel shape data in the magnetic control intervention twin platform by using a calculation formula of the steering angle of the guide wire, and synchronously calculating magnetic control data for controlling the robot guide wire to finish the steering angle;
packaging the mapping relation between the magnetic control data and the blood vessel shape data on a magnetic control intervention twin platform by using a neural network to obtain a control model for outputting the magnetic control data according to the blood vessel shape data;
self-adaptive learning is added in the packaging process of the control model, and the performance of the control model is self-adaptively optimized, so that the magnetic control data output by the control model reaches the optimal control of the robot guide wire.
As a preferable scheme of the invention, the construction method of the magnetic control intervention twin platform comprises the following steps:
three-dimensionally reconstructing the angiographic image into a vessel shape three-dimensional model in a digital twin platform;
The robot guide wire is projected and mapped into a three-dimensional model of the robot guide wire in a digital twin platform;
establishing a data transmission channel between a vision sensor carried at the front end of the robot guide wire and a three-dimensional model of the robot guide wire in a digital twin platform;
and obtaining the magnetic control intervention twin platform based on the blood vessel shape-moving three-dimensional model, the robot guide wire three-dimensional model, the data transmission channel and the digital twin platform.
As a preferred aspect of the present invention, the vessel shape data in the magnetically controlled intervention twin table includes: and the three-dimensional model is used for receiving the blood vessel shape data acquired by the visual sensor carried at the front end of the robot guide wire.
As a preferred embodiment of the present invention, the magnetic control data calculation method includes:
Acquiring a starting point coordinate and an ending point coordinate of the movement of the robot guide wire from the blood vessel shape data;
Calculating a three-dimensional deflection angle of the robot guide wire according to the starting point coordinate and the end point coordinate by using a calculation formula of the guide wire steering angle, wherein the three-dimensional deflection angle is as follows:
；
；
Wherein, K xoy is the steering angle in the horizontal plane direction, K yoz is the steering angle in the vertical plane direction, x start is the coordinate value of the starting point coordinate in the x direction, y start is the coordinate value of the starting point coordinate in the y direction, x end is the coordinate value of the ending point coordinate in the x direction, y end is the coordinate value of the ending point coordinate in the y direction, x xyz is the coordinate value of the front end of the robot wire after deflection in the x direction, y xyz is the coordinate value of the front end of the robot wire after deflection in the y direction, and z xyz is the coordinate value of the front end of the robot wire after deflection in the z direction;
calculating and integrating an Euler-Bernoulli equation according to the three-dimensional deflection angle, and calculating magnetic control data for controlling the robot guide wire to reach an end point coordinate from a start point coordinate, wherein the magnetic control data comprises magnetic control field intensity and magnetic control direction;
the magnetic control field intensity is as follows: ；
；
wherein Bxoy is the magnetic control field intensity for completing the steering angle in the horizontal plane direction, byoz is the magnetic control field intensity for completing the steering angle in the c vertical plane direction, A is the cross-sectional area of the robot guide wire, I is the area second moment of the robot guide wire, E is the Young's modulus of the permanent magnet of the magnetic control robot guide wire, K xoy is the steering angle in the horizontal plane direction, and K yoz is the steering angle in the vertical plane direction;
The magnetic control direction is as follows: p= (x end-xxyz,yend-yxyz,zend-zxyz) where P is the magnetically controlled direction of the robot wire from the start point coordinate to the end point coordinate, x end is the coordinate value of the end point coordinate in the x direction, y end is the coordinate value of the end point coordinate in the y direction, z end is the coordinate value of the end point coordinate in the z direction, x xyz is the coordinate value of the front end of the robot wire after deflection in the x direction, y xyz is the coordinate value of the front end of the robot wire after deflection in the y direction, and z xyz is the coordinate value of the front end of the robot wire after deflection in the z direction;
Wherein the method comprises the steps of ,xxyz=xstart*cosKxyz*cosKzoy+ystart*cosKxyz-zstart*sinKxyz;
yxyz=ystart*cosKyoz-xstart*sinKxoy;zxyz=zstart*cosKxoy+xstart*cosKyoz*sinKxoy+ystart*sinKxoy*sinKyoz.
As a preferred embodiment of the present invention, the control model construction method includes:
dividing the blood vessel shape three-dimensional model according to preset length to obtain a plurality of groups of local blood vessels, wherein the starting point coordinates and the end point coordinates in the blood vessel shape data of the local blood vessels respectively correspond to the adjacent two divided node coordinates of the local blood vessels;
taking the blood vessel shape data of the local blood vessel as an input item of the first neural network, and taking the magnetic control field intensity and the magnetic control direction obtained according to the blood vessel shape data of the local blood vessel as an output item of the first neural network;
performing convolution learning on a mapping relation between an input item of the first neural network and an output item of the first neural network by using the first neural network to obtain a generalization model for obtaining magnetic control data according to a blood vessel shape three-dimensional model;
Taking the blood vessel shape data received by the three-dimensional model of the robot guide wire and acquired by the vision sensor carried by the front end of the robot guide wire as an input item of a second neural network, and taking the magnetic control field intensity and the magnetic control direction obtained by the blood vessel shape data received by the three-dimensional model of the robot guide wire and acquired by the vision sensor carried by the front end of the robot guide wire as an output item of the second neural network, wherein the starting point coordinates and the end point coordinates of the blood vessel shape data acquired by the vision sensor correspond to the coordinates of blood vessels at the nearest end and the farthest end in the visual field range of the vision sensor respectively;
convolving and learning a mapping relation between an input item of the second neural network and an output item of the second neural network by using the second neural network to obtain a fine model for obtaining magnetic control data according to blood vessel shape data obtained by a visual sensor;
combining the generalization model and the fine model to obtain the control model;
The control model is as follows: [ Bxoy, byoz, P ] goal=softmax([Bxoy,Byoz,P]1,[Bxoy,Byoz,P]2);
Wherein [ Bxoy, byoz, P ] goal is the magnetic control field intensity and magnetic control direction output by a control model, [ Bxoy, byoz, P ] 1 is the magnetic control field intensity and magnetic control direction output by a generalization model, [ Bxoy, byoz, P ] 2 is the magnetic control field intensity and magnetic control direction output by a fine model; wherein, [ Bxoy, byoz, P ] 1=Model1(S3d);
[Bxoy,Byoz,P]2=Model2(Ssensor)；
Wherein Model1 is a first neural network, model2 is a second neural network, S 3d is the blood vessel shape data of a local blood vessel in the three-dimensional Model of blood vessel shape, and S sensor is the blood vessel shape data acquired by a visual sensor. As a preferred embodiment of the present invention, the adaptive optimization method for the performance of the control model includes:
adding self-adaptive weights to the generalization model and the fine model to realize self-adaptive optimization of the generalization model and the fine model in the control model;
the adaptive optimization result of the generalization model is as follows: [ Bxoy, byoz, P ] 1best=W1(k)*Model1(S3d); the self-adaptive optimization result of the fine model is as follows: [ Bxoy, byoz, P ] 2best=W2(k)*Model1(Ssensor);
The self-adaptive optimization result of the control model is as follows:
[Bxoy,Byoz,P]goalbest=softmax([Bxoy,Byoz,P]1best,[Bxoy,Byoz,P]2best)；
Wherein [ Bxoy, byoz, P ] goalbest is the magnetic control field intensity and magnetic control direction output by the optimized control model, [ Bxoy, byoz, P ] 1best is the magnetic control field intensity and magnetic control direction output by the optimized generalization model, [ Bxoy, byoz, P ] 2best is the magnetic control field intensity and magnetic control direction output by the optimized fine model; w1 (k) is the adaptive weight of the generalization model, and W2 (k) is the adaptive weight of the fine model.
As a preferred embodiment of the present invention, the method for setting adaptive weights includes:
the adaptive weight of the generalization model is as follows: ；
The adaptive weights of the fine model are as follows: ；
Wherein W1 (k) is an adaptive weight of the generalization model, W2 (k) is an adaptive weight of the fine model, x em is a coordinate value of the intervention target in the x direction, y em is a coordinate value of the intervention target in the y direction, z em is a coordinate value of the intervention target in the z direction, x sm is a coordinate value of the intervention entrance in the x direction, y sm is a coordinate value of the intervention entrance in the y direction, z sm is a coordinate value of the intervention entrance in the z direction, x k is a real-time coordinate value of the front end of the robot wire in the x direction, y k is a real-time coordinate value of the front end of the robot wire in the y direction, and z xyz is a real-time coordinate value of the front end of the robot wire in the z direction.
As a preferred embodiment of the present invention, the first neural network and the second neural network have the same network structure.
In a second aspect of the present invention, the present invention provides a magnetic control intervention control method based on a robot guide wire, which is applied to a magnetic control intervention control system based on a robot guide wire, and the system includes:
The platform builds the unit for according to angiography image, robot seal wire, build out the magnetic control and intervene twin platform, wherein, the structure of robot seal wire includes: the rear end of the robot guide wire is an unmagnetized part, the front end of the robot guide wire is a magnetized part, and a visual sensor is loaded at the front end of the magnetized part;
The data preprocessing unit is used for determining the steering angle of the robot guide wire according to the blood vessel shape data in the magnetic control intervention twin platform by utilizing a calculation formula of the steering angle of the guide wire, and synchronously calculating magnetic control data for controlling the robot guide wire to finish the steering angle;
the model packaging unit is used for packaging the mapping relation between the magnetic control data and the blood vessel shape data onto the magnetic control intervention twin platform by using a neural network to obtain a control model for outputting the magnetic control data according to the blood vessel shape data;
The platform running unit is used for adding self-adaptive learning in the packaging process of the control model by utilizing the magnetic control intervention twin platform, and performing self-adaptive optimization on the performance of the control model, so that the magnetic control data output by the control model reach the optimal control on the robot guide wire.
In a third aspect of the present invention, a computer-readable storage medium having stored therein computer-executable instructions that, when executed by a processor, implement a magnetic control intervention control method.
Compared with the prior art, the invention has the following beneficial effects:
According to the invention, the operation model of the magnetic control data is packaged on the digital twin platform through the neural network, so that the magnetic control data can be operated in a virtual space close to a real environment in a development and test stage, a robot guide wire can obtain a steering angle which is suitable for a real-time blood vessel real structure from the virtual space of the digital twin platform in real time, the operation and intervention progress of the magnetic control data are mutually independent, the intervention process is prevented from being interrupted, the safety of the intervention process of the guide wire is ensured, the control model comprises a generalization model and a fine model, the performance of the model is improved through common packaging training, real-time adjustment according to the real-time condition of the blood vessel on the basis of pre-planning is realized, and the operation accuracy of the magnetic control data is higher.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below. It will be apparent to those of ordinary skill in the art that the drawings in the following description are exemplary only and that other implementations can be obtained from the extensions of the drawings provided without inventive effort.
FIG. 1 is a flow chart of a magnetic control intervention control method provided by an embodiment of the invention;
FIG. 2 is a block diagram of a magnetic control intervention control system provided by an embodiment of the invention;
Fig. 3 is a diagram of a structure of a robot guide wire according to an embodiment of the present invention;
fig. 4 is a schematic diagram of calculation of magnetic control data according to an embodiment of the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
As shown in fig. 1, in a first aspect of the present invention, the present invention provides a magnetic control intervention control method based on a robot guide wire, including the steps of:
constructing a magnetic control intervention twin platform on the digital twin platform according to the angiography image and the robot guide wire;
Determining the steering angle of the robot guide wire according to the blood vessel shape data in the magnetic control intervention twin platform by using a calculation formula of the steering angle of the guide wire, and synchronously calculating magnetic control data for controlling the robot guide wire to finish the steering angle;
packaging the mapping relation between the magnetic control data and the blood vessel shape data on a magnetic control intervention twin platform by using a neural network to obtain a control model for outputting the magnetic control data according to the blood vessel shape data;
self-adaptive learning is added in the packaging process of the control model, and the performance of the control model is self-adaptively optimized, so that the magnetic control data output by the control model reaches the optimal control of the robot guide wire.
The invention establishes a virtual model (a three-dimensional model of the robot guide wire and a three-dimensional model of the blood vessel shape variation) which is the same as the real structure of the blood vessel shape variation and the real structure of the robot guide wire in a virtual space by utilizing a digital twin technology, and realizes the visualization of the interventional process by utilizing the interaction between the virtual model in a twin platform and the robot guide wire data in the physical space and simulating the relative position between the robot guide wire and the blood vessel in real time.
According to the invention, the visual sensor carried at the front end of the robot guide wire and the data transmission channel between the three-dimensional model of the robot guide wire in the digital twin platform realize the data interaction between the virtual space and the physical space, so that the control of the magnetic control data required by the vascular advancing of the robot guide wire on the twin platform is realized according to the data interacted in the physical space, the control of the robot guide wire in the physical space and the magnetic control data calculation of the guide wire in the virtual space are independently separated, the intervention process is prevented from being interrupted by the operation process of the magnetic control data, and the safety of the intervention process of the guide wire in the physical space is ensured.
According to the blood vessel shape, the invention determines the steering angle of the robot guide wire, and then calculates the magnetic control data of the control robot guide wire according to the steering degree of the robot guide wire, namely the magnetic field intensity and the magnetic field direction required by the control robot guide wire to realize the steering angle advancing at the blood vessel shape.
In the invention, the correlation between the vascular shape and the magnetic control data of the robot guide wire is established, and the vascular shape and the magnetic control data of the robot guide wire are packaged on a twin platform to form a neural network model for directly obtaining the magnetic control data according to the vascular shape data, wherein the twin platform is packaged with the magnetic control data which can be obtained according to the vascular shape data in the vascular shape three-dimensional model reconstructed in advance by preoperative angiography, the magnetic control data is derived from the vascular shape data recorded by the preoperative angiography, and the twin platform is also provided with the magnetic control data which can be obtained according to the vascular shape data in the real-time vascular image in the intervention process, and the magnetic control data is derived from the vascular shape data obtained in the intervention.
The magnetic control data is obtained according to the blood vessel shape data in the blood vessel shape three-dimensional model reconstructed in advance by preoperative angiography, the magnetic control data is deduced by utilizing the preoperative blood vessel image, the magnetic control data of each blood vessel shape can be grasped in advance, namely the magnetic control data is provided with a rough measurement and calculation, the magnetic control data is derived from the blood vessel shape data recorded by the preoperative angiography and does not necessarily completely match the real condition (real blood vessel shape data) of the blood vessel in the intervention process, the preoperative angiography is the rough calculation or the generalized calculation of the magnetic control data.
The fine model in the magnetic control model for acquiring the magnetic control data is established according to the magnetic control data obtained by the blood vessel shape data in the real-time blood vessel image acquired by the visual sensor loaded at the front end of the guide wire in the intervention process, the magnetic control data is estimated by utilizing the real condition of the blood vessel in the intervention process, the real-time blood vessel image acquired by the visual sensor is mastered in real time by the blood vessel shape, the fine model is the accurate calculation or the fine calculation of the magnetic control data, and the calculation of the fine model needs to carry out the steps of data acquisition, transmission, processing and the like in the intervention process, so that a certain calculation time is consumed in the calculation process, and the calculation efficiency of the magnetic control data is influenced.
The invention combines the fine model and the generalization model, can combine the complementary advantages between the fine model and the generalization model, can roughly calculate the magnetic control data in advance, quickly reduces the solving range, and then provides the solving range as priori knowledge for the fine model, so that the fine model can further carry out fine solving on the basis of the result of the generalization model, thereby ensuring the calculating efficiency in the process of estimating the magnetic control data and simultaneously giving attention to accuracy.
Furthermore, in order to optimally combine the advantages of the fine model and the generalization model, the invention adds the self-adaptive weight to the fine model and the generalization model, so that the generalization model can be provided with a large weight at the starting point of a blood vessel at the robot guide wire, the fine model is provided with a small weight, the weight of the generalization model is reduced along with the advancing of the robot guide wire towards the end point, the weight of the fine model is increased, the generalization model is provided with a small weight at the end point, the fine model is provided with a large weight, so that global search and convergence are accelerated before magnetic control are performed, a solution range is determined on the basis of preoperative angiography rapidly, as a priori solution range, magnetic control data which is matched with the real condition of the blood vessel in intervention is accurately obtained in the prior solution range along with the approaching enhancement of the end point, namely the solution efficiency and the solution precision of the magnetic control data are optimized, and the solution efficiency and the solution precision of the magnetic control data are balanced are adaptively realized.
The invention establishes a virtual model (a robot guide wire three-dimensional model and a blood vessel shape-moving three-dimensional model) which is the same as the real structure of blood vessel shape-moving and the real structure of a robot guide wire in a virtual space by utilizing a digital twin technology, and interacts with the robot guide wire data in the physical space by utilizing the virtual model in a twin platform, and specifically comprises the following steps:
the construction method of the magnetic control intervention twin platform comprises the following steps:
three-dimensionally reconstructing the angiographic image into a vessel shape-moving three-dimensional model in a digital twin platform;
the method comprises the steps of mapping a projection of a robot guide wire into a three-dimensional model of the robot guide wire in a digital twin platform;
establishing a data transmission channel between a vision sensor carried at the front end of the robot guide wire and a three-dimensional model of the robot guide wire in a digital twin platform;
And obtaining the magnetic control intervention twin platform based on the blood vessel shape-moving three-dimensional model, the robot guide wire three-dimensional model, the data transmission channel and the digital twin platform.
The blood vessel shape data in the magnetic control intervention twin platform comprises: and the three-dimensional model is used for receiving the blood vessel shape data acquired by the visual sensor carried at the front end of the robot guide wire.
As shown in fig. 3, the structure of the robot guide wire is: a rear non-magnetized part and a front magnetized part, and a visual sensor is arranged at the front of the magnetized part.
The invention determines the steering angle of the robot guide wire according to the blood vessel shape, and then calculates the magnetic control data of the control robot guide wire according to the steering degree of the robot guide wire, namely the magnetic field intensity and the magnetic field direction required by the control robot guide wire to realize the steering angle advancing at the blood vessel shape, and the invention comprises the following specific steps:
as shown in fig. 4, the magnetic control data calculation method includes:
Acquiring a starting point coordinate and an ending point coordinate of the movement of the robot guide wire from the blood vessel shape data;
calculating the three-dimensional deflection angle of the robot guide wire according to the starting point coordinate and the end point coordinate by using a calculation formula of the guide wire steering angle, wherein the three-dimensional deflection angle is as follows:
；
；
Wherein, K xoy is the steering angle in the horizontal plane direction, K yoz is the steering angle in the vertical plane direction, x start is the coordinate value of the starting point coordinate in the x direction, y start is the coordinate value of the starting point coordinate in the y direction, x end is the coordinate value of the ending point coordinate in the x direction, y end is the coordinate value of the ending point coordinate in the y direction, x xyz is the coordinate value of the front end of the robot wire after deflection in the x direction, y xyz is the coordinate value of the front end of the robot wire after deflection in the y direction, and z xyz is the coordinate value of the front end of the robot wire after deflection in the z direction; calculating and integrating an Euler-Bernoulli equation according to the three-dimensional deflection angle, and calculating magnetic control data for controlling the robot guide wire to reach an end point coordinate from a start point coordinate, wherein the magnetic control data comprises magnetic control field intensity and magnetic control direction;
the magnetic control field intensity is: ；
；
wherein Bxoy is the magnetic control field intensity for completing the steering angle in the horizontal plane direction, byoz is the magnetic control field intensity for completing the steering angle in the c vertical plane direction, A is the cross-sectional area of the robot guide wire, I is the area second moment of the robot guide wire, E is the Young's modulus of the permanent magnet of the magnetic control robot guide wire, K xoy is the steering angle in the horizontal plane direction, and K yoz is the steering angle in the vertical plane direction; the magnetic control direction is as follows: p= (x end-xxyz,yend-yxyz,zend-zxyz) where P is the magnetically controlled direction of the robot wire from the start point coordinate to the end point coordinate, x end is the coordinate value of the end point coordinate in the x direction, y end is the coordinate value of the end point coordinate in the y direction, z end is the coordinate value of the end point coordinate in the z direction, x xyz is the coordinate value of the front end of the robot wire after deflection in the x direction, y xyz is the coordinate value of the front end of the robot wire after deflection in the y direction, and z xyz is the coordinate value of the front end of the robot wire after deflection in the z direction;
Wherein the method comprises the steps of ,xxyz=xstart*cosKxyz*cosKzoy+ystart*cosKxyz-zstart*sinKxyz;
yxyz=ystart*cosKyoz-xstart*sinKxoy；
zxyz=zstart*cosKxoy+xstart*cosKyoz*sinKxoy+ystart*sinKxoy*sinKyoz.
In the invention, the correlation between the vascular shape and the magnetic control data of the robot guide wire is established, and the vascular shape and the magnetic control data of the robot guide wire are packaged on a twin platform to form a neural network model for directly obtaining the magnetic control data according to the vascular shape data, wherein the twin platform is packaged with the magnetic control data which can be obtained according to the vascular shape data in the vascular shape three-dimensional model reconstructed in advance by preoperative angiography, the magnetic control data is derived from the vascular shape data recorded by the preoperative angiography, and the twin platform is also provided with the magnetic control data which can be obtained according to the vascular shape data in the real-time vascular image in the intervention process, and the magnetic control data is derived from the vascular shape data obtained in the intervention.
The control model construction method comprises the following steps:
dividing the blood vessel shape three-dimensional model according to preset length to obtain a plurality of groups of local blood vessels, wherein the starting point coordinates and the end point coordinates in the blood vessel shape data of the local blood vessels respectively correspond to the adjacent two divided node coordinates of the local blood vessels;
Taking the blood vessel shape data of the local blood vessel as an input item of the first neural network, and taking the magnetic control field intensity and the magnetic control direction obtained according to the blood vessel shape data of the local blood vessel as an output item of the first neural network;
performing convolution learning on a mapping relation between an input item of the first neural network and an output item of the first neural network by using the first neural network to obtain a generalization model for obtaining magnetic control data according to a blood vessel shape three-dimensional model;
Taking the blood vessel shape data received by the three-dimensional model of the robot guide wire and acquired by the vision sensor carried by the front end of the robot guide wire as an input item of a second neural network, and taking magnetic control field intensity and magnetic control direction obtained by the blood vessel shape data received by the three-dimensional model of the robot guide wire and acquired by the vision sensor carried by the front end of the robot guide wire as an output item of the second neural network, wherein the starting point coordinates and the end point coordinates of the blood vessel shape data acquired by the vision sensor correspond to the coordinates of blood vessels at the nearest end and the farthest end in the visual field range of the vision sensor respectively;
convolving and learning a mapping relation between an input item of the second neural network and an output item of the second neural network by using the second neural network to obtain a fine model for obtaining magnetic control data according to blood vessel shape data obtained by a visual sensor;
Combining the generalization model and the fine model to obtain a control model;
the control model is as follows:
[ Bxoy, byoz, P ] goal=softmax([Bxoy,Byoz,P]1,[Bxoy,Byoz,P]2); wherein [ Bxoy, byoz, P ] goal is the magnetic control field intensity and magnetic control direction output by the control model, [ Bxoy, byoz, P ] 1 is the magnetic control field intensity and magnetic control direction output by the generalization model, [ Bxoy, byoz, P ] 2 is the magnetic control field intensity and magnetic control direction output by the fine model;
Wherein, [ Bxoy, byoz, P ] 1=Model1(S3d);
[Bxoy,Byoz,P]2=Model2(Ssensor)；
Wherein Model1 is a first neural network, model2 is a second neural network, S 3d is the blood vessel shape data of a local blood vessel in the three-dimensional Model of blood vessel shape, and S sensor is the blood vessel shape data acquired by a visual sensor.
The magnetic control data is obtained according to the blood vessel shape data in the blood vessel shape three-dimensional model reconstructed in advance by preoperative angiography, the magnetic control data is deduced by utilizing the preoperative blood vessel image, the magnetic control data of each blood vessel shape can be grasped in advance, namely the magnetic control data is provided with a rough measurement and calculation, the magnetic control data is derived from the blood vessel shape data recorded by the preoperative angiography and does not necessarily completely match the real condition (real blood vessel shape data) of the blood vessel in the intervention process, the preoperative angiography is the rough calculation or the generalized calculation of the magnetic control data.
The fine model in the magnetic control model for acquiring the magnetic control data is established according to the magnetic control data obtained by the blood vessel shape data in the real-time blood vessel image acquired by the visual sensor loaded at the front end of the guide wire in the intervention process, the magnetic control data is estimated by utilizing the real condition of the blood vessel in the intervention process, the real-time blood vessel image acquired by the visual sensor is mastered in real time by the blood vessel shape, the fine model is the accurate calculation or the fine calculation of the magnetic control data, and the calculation of the fine model needs to carry out the steps of data acquisition, transmission, processing and the like in the intervention process, so that a certain calculation time is consumed in the calculation process, and the calculation efficiency of the magnetic control data is influenced.
The invention combines the fine model and the generalization model, can combine the complementary advantages between the fine model and the generalization model, can roughly calculate the magnetic control data in advance, quickly reduces the solving range, then provides the solving range as priori knowledge for the fine model, and ensures that the fine model further carries out fine solving on the basis of the result of the generalization model, thereby ensuring the calculating efficiency in the process of estimating the magnetic control data and simultaneously giving consideration to the accuracy, and the invention comprises the following specific steps:
The self-adaptive optimization method for the performance of the control model comprises the following steps:
adding self-adaptive weights to the generalization model and the fine model to realize self-adaptive optimization of the generalization model and the fine model in the control model;
The adaptive optimization result of the generalization model is as follows: [ Bxoy, byoz, P ] 1best=W1(k)*Model1(S3d);
The adaptive optimization result of the fine model is as follows: [ Bxoy, byoz, P ] 2best=W2(k)*Model1(Ssensor);
The self-adaptive optimization result of the control model is as follows:
[Bxoy,Byoz,P]goalbest=softmax([Bxoy,Byoz,P]1best,[Bxoy,Byoz,P]2best)；
Wherein [ Bxoy, byoz, P ] goalbest is the magnetic control field intensity and magnetic control direction output by the optimized control model, [ Bxoy, byoz, P ] 1best is the magnetic control field intensity and magnetic control direction output by the optimized generalization model, [ Bxoy, byoz, P ] 2best is the magnetic control field intensity and magnetic control direction output by the optimized fine model;
w1 (k) is the adaptive weight of the generalization model, and W2 (k) is the adaptive weight of the fine model.
The self-adaptive weight setting method comprises the following steps:
The adaptive weights of the generalization model are as follows: ；
The adaptive weights of the fine model are: ；
Wherein W1 (k) is an adaptive weight of the generalization model, W2 (k) is an adaptive weight of the fine model, x em is a coordinate value of the intervention target in the x direction, y em is a coordinate value of the intervention target in the y direction, z em is a coordinate value of the intervention target in the z direction, x sm is a coordinate value of the intervention entrance in the x direction, y sm is a coordinate value of the intervention entrance in the y direction, z sm is a coordinate value of the intervention entrance in the z direction, x k is a real-time coordinate value of the front end of the robot wire in the x direction, y k is a real-time coordinate value of the front end of the robot wire in the y direction, and z xyz is a real-time coordinate value of the front end of the robot wire in the z direction.
In order to optimally combine the advantages of the fine model and the generalization model, the invention adds the self-adaptive weight to the fine model and the generalization model, so that the generalization model of the robot guide wire at the vascular intervention entrance has large weight, the fine model has small weight, the weight of the generalization model is reduced along with the advancing of the robot guide wire towards the intervention target, the weight of the fine model is increased, the generalization model at the intervention target has small weight, the fine model has large weight, the global search is accelerated to converge in advance in the magnetic control, a solution range is determined on the basis of preoperative angiography rapidly, as the prior solution range, the magnetic control data which is matched with the real vascular condition in the intervention is accurately obtained in the prior solution range along with the approaching enhancement of the intervention target, namely the solution efficiency and the solution precision of the magnetic control data are optimized, and the solution efficiency and the solution precision of the magnetic control data are balanced in a self-adaptive manner.
The network structures of the first neural network and the second neural network are the same.
As shown in fig. 2, in a second aspect of the present invention, the present invention provides a magnetic control intervention control method based on a robot wire, which is applied to a magnetic control intervention control system based on a robot wire, the system includes:
the platform building unit is used for building a magnetic control intervention twin platform according to the angiography image and the robot guide wire;
the data preprocessing unit is used for determining the steering angle of the robot guide wire according to the blood vessel shape data in the magnetic control intervention twin platform by utilizing a calculation formula of the steering angle of the guide wire, and synchronously calculating magnetic control data for controlling the robot guide wire to finish the steering angle;
the model packaging unit is used for packaging the mapping relation between the magnetic control data and the blood vessel shape data onto the magnetic control intervention twin platform by using a neural network to obtain a control model for outputting the magnetic control data according to the blood vessel shape data;
The platform running unit is used for adding self-adaptive learning in the packaging process of the control model by utilizing the magnetic control intervention twin platform, and performing self-adaptive optimization on the performance of the control model, so that the magnetic control data output by the control model reach the optimal control on the robot guide wire.
In a third aspect of the present invention, a computer-readable storage medium having stored therein computer-executable instructions that, when executed by a processor, implement a magnetic control intervention control method.
According to the invention, the operation model of the magnetic control data is packaged on the digital twin platform through the neural network, so that the magnetic control data can be operated in a virtual space close to a real environment in a development and test stage, a robot guide wire can obtain a steering angle which is suitable for a real-time blood vessel real structure from the virtual space of the digital twin platform in real time, the operation and intervention progress of the magnetic control data are mutually independent, the intervention process is prevented from being interrupted, the safety of the intervention process of the guide wire is ensured, the control model comprises a generalization model and a fine model, the performance of the model is improved through common packaging training, real-time adjustment according to the real-time condition of the blood vessel on the basis of pre-planning is realized, and the operation accuracy of the magnetic control data is higher.
The above embodiments are only exemplary embodiments of the present application and are not intended to limit the present application, the scope of which is defined by the claims. Various modifications and equivalent arrangements of this application will occur to those skilled in the art, and are intended to be within the spirit and scope of the application.
Claims (4)
1. The utility model provides a magnetic control intervenes control system based on robot seal wire which characterized in that, magnetic control intervenes control system includes:
The platform builds the unit for according to angiography image, robot seal wire, build out the magnetic control and intervene twin platform, wherein, the structure of robot seal wire includes: the rear end of the robot guide wire is an unmagnetized part, the front end of the robot guide wire is a magnetized part, and a visual sensor is loaded at the front end of the magnetized part;
The data preprocessing unit is used for determining the steering angle of the robot guide wire according to the blood vessel shape data in the magnetic control intervention twin platform by utilizing a calculation formula of the steering angle of the guide wire, and synchronously calculating magnetic control data for controlling the robot guide wire to finish the steering angle;
the model packaging unit is used for packaging the mapping relation between the magnetic control data and the blood vessel shape-changing data in the magnetic control intervention twin platform onto the magnetic control intervention twin platform by using a neural network to obtain a control model for outputting the magnetic control data according to the blood vessel shape-changing data in the magnetic control intervention twin platform;
The platform running unit is used for adding self-adaptive learning in the packaging process of the control model by utilizing the magnetic control intervention twin platform, and carrying out self-adaptive optimization on the performance of the control model, so that the magnetic control data output by the control model reach the optimal control on the robot guide wire;
the construction method of the magnetic control intervention twin platform comprises the following steps:
three-dimensionally reconstructing the angiographic image into a vessel shape three-dimensional model in a digital twin platform;
The robot guide wire is projected and mapped into a three-dimensional model of the robot guide wire in a digital twin platform;
establishing a data transmission channel between a vision sensor carried at the front end of the robot guide wire and a three-dimensional model of the robot guide wire in a digital twin platform;
Obtaining the magnetic control intervention twin platform based on the blood vessel shape-moving three-dimensional model, the robot guide wire three-dimensional model, the data transmission channel and the digital twin platform;
The blood vessel shape data in the magnetic control intervention twin platform comprises: the method comprises the steps of (1) blood vessel shape data in a blood vessel shape three-dimensional model and blood vessel shape data received by a robot guide wire three-dimensional model and acquired by a vision sensor carried at the front end of the robot guide wire;
the magnetic control data calculation method comprises the following steps:
Acquiring a start point coordinate and an end point coordinate of the movement of the robot guide wire from the blood vessel shape data in the magnetic control intervention twin platform;
calculating the three-dimensional deflection angle of the robot guide wire according to the starting point coordinate and the end point coordinate by using a calculation formula of the guide wire steering angle, wherein the three-dimensional deflection angle is as follows: ； ; wherein, K xoy is the steering angle in the horizontal plane direction, K yoz is the steering angle in the vertical plane direction, x start is the coordinate value of the starting point coordinate in the x direction, y start is the coordinate value of the starting point coordinate in the y direction, x end is the coordinate value of the ending point coordinate in the x direction, y end is the coordinate value of the ending point coordinate in the y direction, x xyz is the coordinate value of the front end of the robot wire after deflection in the x direction, y xyz is the coordinate value of the front end of the robot wire after deflection in the y direction, and z xyz is the coordinate value of the front end of the robot wire after deflection in the z direction; calculating and integrating an Euler-Bernoulli equation according to the three-dimensional deflection angle, and calculating magnetic control data for controlling the robot guide wire to reach an end point coordinate from a start point coordinate, wherein the magnetic control data comprises magnetic control field intensity and magnetic control direction;
the magnetic control field intensity is: ； ; wherein Bxoy is the magnetic control field intensity for completing the steering angle in the horizontal plane direction, byoz is the magnetic control field intensity for completing the steering angle in the c vertical plane direction, A is the cross-sectional area of the robot guide wire, I is the area second moment of the robot guide wire, E is the Young's modulus of the permanent magnet of the magnetic control robot guide wire, K xoy is the steering angle in the horizontal plane direction, and K yoz is the steering angle in the vertical plane direction; the magnetic control direction is as follows: p= (x end-xxyz,yend-yxyz,zend-zxyz) where P is the magnetically controlled direction of the robot wire from the start point coordinate to the end point coordinate, x end is the coordinate value of the end point coordinate in the x direction, y end is the coordinate value of the end point coordinate in the y direction, z end is the coordinate value of the end point coordinate in the z direction, x xyz is the coordinate value of the front end of the robot wire after deflection in the x direction, y xyz is the coordinate value of the front end of the robot wire after deflection in the y direction, and z xyz is the coordinate value of the front end of the robot wire after deflection in the z direction; wherein the method comprises the steps of ,xxyz=xstart*cosKxyz*cosKzoy+ystart*cosKxyz-zstart*sinKxyz;yxyz=ystart*cosKyoz-xstart*sinKxoy;
zxyz=zstart*cosKxoy+xstart*cosKyoz*sinKxoy+ystart*sinKxoy*sinKyoz; The control model construction method comprises the following steps:
Dividing the blood vessel shape three-dimensional model according to preset length to obtain a plurality of groups of local blood vessels, wherein the starting point coordinates and the ending point coordinates in the blood vessel shape data of the local blood vessels in the blood vessel shape three-dimensional model correspond to the adjacent two divided node coordinates of the local blood vessels respectively;
Taking the blood vessel shape data of the local blood vessel in the blood vessel shape three-dimensional model as an input item of the first neural network, and taking magnetic control field intensity and magnetic control direction obtained according to the blood vessel shape data of the local blood vessel in the blood vessel shape three-dimensional model as an output item of the first neural network;
performing convolution learning on a mapping relation between an input item of the first neural network and an output item of the first neural network by using the first neural network to obtain a generalization model for obtaining magnetic control data according to a blood vessel shape three-dimensional model;
Taking the blood vessel shape data received by the three-dimensional model of the robot guide wire and acquired by the vision sensor carried by the front end of the robot guide wire as an input item of a second neural network, and taking magnetic control field intensity and magnetic control direction obtained by the blood vessel shape data received by the three-dimensional model of the robot guide wire and acquired by the vision sensor carried by the front end of the robot guide wire as an output item of the second neural network, wherein the starting point coordinates and the end point coordinates of the blood vessel shape data acquired by the vision sensor carried by the front end of the robot guide wire correspond to the coordinates of blood vessels at the nearest end and the farthest end in the visual field range of the vision sensor respectively;
Performing convolution learning on a mapping relation between an input item of the second neural network and an output item of the second neural network by using the second neural network to obtain a fine model for obtaining magnetic control data according to blood vessel shape data obtained by a visual sensor carried by the front end of the robot guide wire;
Combining the generalization model and the fine model to obtain a control model;
the control model is as follows: [ Bxoy, byoz, P ] goal=softmax([Bxoy,Byoz,P]1,[Bxoy,Byoz,P]2);
Wherein [ Bxoy, byoz, P ] goal is the magnetic control field intensity and magnetic control direction output by the control model, [ Bxoy, byoz, P ] 1 is the magnetic control field intensity and magnetic control direction output by the generalization model, [ Bxoy, byoz, P ] 2 is the magnetic control field intensity and magnetic control direction output by the fine model; wherein, [ Bxoy, byoz, P ] 1=Model1(S3d);
[ Bxoy, byoz, P ] 2=Model2(Ssensor); in the formula, model1 is a first neural network, model2 is a second neural network, S 3d is the blood vessel shape data of a local blood vessel in the blood vessel shape three-dimensional Model, and S sensor is the blood vessel shape data obtained by a vision sensor carried on the front end of a robot guide wire.
2. The robotic guidewire-based magnetic intervention control system of claim 1, wherein: the self-adaptive optimization method for the performance of the control model comprises the following steps:
adding self-adaptive weights to the generalization model and the fine model to realize self-adaptive optimization of the generalization model and the fine model in the control model;
The adaptive optimization result of the generalization model is as follows: [ Bxoy, byoz, P ] 1best=W1(k)*Model1(S3d);
the self-adaptive optimization result of the fine model is as follows: [ Bxoy, byoz, P ] 2best=W2(k)*Model1(Ssensor);
The self-adaptive optimization result of the control model is as follows: [ Bxoy, byoz, P ] goalbest=softmax([Bxoy,Byoz,P]1best,[Bxoy,Byoz,P]2best);
Wherein [ Bxoy, byoz, P ] goalbest is the magnetic control field intensity and magnetic control direction output by the optimized control model, [ Bxoy, byoz, P ] 1best is the magnetic control field intensity and magnetic control direction output by the optimized generalization model, [ Bxoy, byoz, P ] 2best is the magnetic control field intensity and magnetic control direction output by the optimized fine model;
w1 (k) is the adaptive weight of the generalization model, and W2 (k) is the adaptive weight of the fine model.
3. The robotic guidewire-based magnetic intervention control system of claim 2, wherein: the self-adaptive weight setting method comprises the following steps:
the adaptive weight of the generalization model is as follows: ; the adaptive weights of the fine model are as follows: ; wherein W1 (k) is an adaptive weight of the generalization model, W2 (k) is an adaptive weight of the fine model, x em is a coordinate value of the intervention target in the x direction, y em is a coordinate value of the intervention target in the y direction, z em is a coordinate value of the intervention target in the z direction, x sm is a coordinate value of the intervention entrance in the x direction, y sm is a coordinate value of the intervention entrance in the y direction, z sm is a coordinate value of the intervention entrance in the z direction, x k is a real-time coordinate value of the front end of the robot wire in the x direction, y k is a real-time coordinate value of the front end of the robot wire in the y direction, and z xyz is a real-time coordinate value of the front end of the robot wire in the z direction.
4. The robotic guidewire-based magnetic intervention control system of claim 1, wherein: the network structures of the first neural network and the second neural network are the same.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410650722.5A CN118217012B (en) | 2024-05-24 | 2024-05-24 | Magnetic control intervention control method, system and storage medium based on robot guide wire |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410650722.5A CN118217012B (en) | 2024-05-24 | 2024-05-24 | Magnetic control intervention control method, system and storage medium based on robot guide wire |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN118217012A CN118217012A (en) | 2024-06-21 |
| CN118217012B true CN118217012B (en) | 2024-08-20 |
Family
ID=91513693
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410650722.5A Active CN118217012B (en) | 2024-05-24 | 2024-05-24 | Magnetic control intervention control method, system and storage medium based on robot guide wire |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN118217012B (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2022188881A1 (en) * | 2021-03-12 | 2022-09-15 | 北京阅影科技有限公司 | Method for detecting guidewire-free ffr, guidewire-free imr, and guidewire-free cfr |
| CN116808408A (en) * | 2023-07-31 | 2023-09-29 | 深圳微美机器人有限公司 | Interventional guide wire control method and interventional navigation system |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP3824477A1 (en) * | 2018-08-22 | 2021-05-26 | Siemens Healthcare GmbH | Data-driven estimation of predictive digital twin models from medical data |
| TWI741359B (en) * | 2019-08-30 | 2021-10-01 | 國立中央大學 | Mixed reality system integrated with surgical navigation system |
| CN117462255A (en) * | 2023-10-09 | 2024-01-30 | 绍兴梅奥心磁医疗科技有限公司 | Surgical navigation planning method, system and storage medium based on surgical robot |
| CN117224817A (en) * | 2023-11-02 | 2023-12-15 | 上海微创医疗机器人（集团）股份有限公司 | Magnetic control guide wire control method and magnetic control interventional operation system |
| CN117530775B (en) * | 2024-01-09 | 2024-04-30 | 华中科技大学同济医学院附属协和医院 | Magnetic control intervention control method and system based on artificial intelligence and CT |
-
2024
- 2024-05-24 CN CN202410650722.5A patent/CN118217012B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2022188881A1 (en) * | 2021-03-12 | 2022-09-15 | 北京阅影科技有限公司 | Method for detecting guidewire-free ffr, guidewire-free imr, and guidewire-free cfr |
| CN116808408A (en) * | 2023-07-31 | 2023-09-29 | 深圳微美机器人有限公司 | Interventional guide wire control method and interventional navigation system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN118217012A (en) | 2024-06-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Sommer et al. | Efficient derivative computation for cumulative b-splines on lie groups | |
| Riley et al. | Methods for motion generation and interaction with a humanoid robot: Case studies of dancing and catching | |
| Tuceryan et al. | Single-point active alignment method (SPAAM) for optical see-through HMD calibration for augmented reality | |
| CN103429158B (en) | For the medical imaging device for the graphical representation for providing the positioning for supporting intervening equipment | |
| Gemeiner et al. | Simultaneous motion and structure estimation by fusion of inertial and vision data | |
| Ren et al. | A potential field model using generalized sigmoid functions | |
| CN107515606A (en) | Robot implementation method, control method and robot, electronic equipment | |
| CN109941274A (en) | Parking method and system, server and medium based on radar range finding identification gantry crane | |
| CN114041878A (en) | Three-dimensional reconstruction method and system for CT image of bone joint replacement surgical robot | |
| CN107688184A (en) | A kind of localization method and system | |
| CN115741732B (en) | Interactive path planning and motion control method for massage robot | |
| CN114034299B (en) | Navigation system based on active laser SLAM | |
| CN114027985A (en) | Robot for bone joint replacement surgery | |
| CN118217012B (en) | Magnetic control intervention control method, system and storage medium based on robot guide wire | |
| CN112990101A (en) | Facial organ positioning method based on machine vision and related equipment | |
| Krupa | A new duty-cycling approach for 3D needle steering allowing the use of the classical visual servoing framework for targeting tasks | |
| Kim et al. | Image-based force estimation of deformable tissue using depth map for single-port surgical robot | |
| JP2023068632A (en) | Device accelerating slam, and electronic device including the same | |
| CN113413214B (en) | Surgical robot manpower feedback method and device based on mixed reality guidance | |
| Yang et al. | Guidewire feeding method based on deep reinforcement learning for vascular intervention robot | |
| CN111916214A (en) | Catheter trajectory prediction method for vascular interventional surgeon training system | |
| WO2023274270A1 (en) | Robot preoperative navigation method and system, storage medium, and computer device | |
| US20220175485A1 (en) | Method for operating a visualization system in a surgical application, and visualization system for a surgical application | |
| CN109875659A (en) | Flexible needle plane based on brain emotion learning intelligent control algorithm punctures control device and method | |
| JP2024518533A (en) | Multi-scale Ultrasound Tracking and Display |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |