Autonomous Three-Dimensional Tissue Ablation Control System Based on Soft Robots

Abstract

The present invention provides an autonomous three-dimensional tissue ablation control system based on a soft robot, which relates to the field of soft robots. To achieve precise ablation of diseased tissues by a laser ablation system in a narrow in-vivo environment, the present invention pre-constructs a laser-tissue interaction model to describe the interaction relationship between blue laser and tissues, and thus develops a robust ablation depth control method; in addition, an adaptive laser ablation path tracking method combining the soft robot body with an OCT imaging system is innovatively proposed; finally, a surgical robot system with autonomous precise three-dimensional laser ablation capability is constructed, which can significantly improve the accuracy, safety and automation level of laser ablation surgery.

Description

Technical Field

[0001] The present invention relates to the field of soft robots, and particularly to an autonomous three-dimensional tissue ablation control system based on a soft robot. Background Art

[0002] Tissue ablation is an important treatment method in the field of tumor treatment. Traditional ablation methods mainly rely on manual operation by doctors, which have many drawbacks.

[0003] In related technologies, a soft robot is a new type of robot made of soft materials, which can adapt to various unstructured environments and is safer in interaction with humans. Through the flexibility and adaptability of the soft robot, a robot motion control method capable of freely navigating in narrow and complex natural cavities is developed to ensure that the laser ablation device can accurately reach the designated treatment area. For this purpose, researchers introduce the motion control technology of soft robots, and through the non-linear modeling, dynamic trajectory planning and adaptive control of flexible structures, achieve high-flexibility motion and precise operation for complex environments and non-uniform targets.

[0004] However, in the application of flexible actuators or soft robots, the inherent non-linear characteristics of soft materials (including material, geometric and drive non-linearity), dynamic uncertainties (such as material hysteresis and environmental coupling) make it impossible to establish an accurate model, and the existing technologies lack efficient motion control algorithms, resulting in large path tracking errors and insufficient motion flexibility.

[0005] In addition, in existing laser ablation surgeries, the control of tissue ablation depth mainly relies on preset experimental data and fixed model parameters to determine the laser output power and irradiation time. This open-loop control method cannot dynamically adjust the laser parameters according to the actual response of the tissue during the surgery, easily resulting in insufficient ablation or over-ablation, affecting the surgical effect and safety. Summary of the Invention

[0006] (I) Technical Problems to be Solved

[0007] In view of the deficiencies of the prior art, the present invention provides an autonomous three-dimensional tissue ablation control system based on a soft robot, which solves the technical problems of large path tracking errors and insufficient dynamic control of ablation depth.

[0008] (II) Technical Solutions

[0009] To achieve the above object, the present invention is realized through the following technical solutions:

[0010] An autonomous three-dimensional tissue ablation control system based on a soft robot, the soft robot includes a soft robot body, a laser ablation system, and an OCT imaging system; including:

[0011] A building block for pre - constructing a laser - tissue interaction model to describe the interaction relationship between blue laser and tissue and determine the target ablation depth of the area to be ablated;

[0012] A planning module for receiving multiple two - dimensional ablation points planned by a doctor on an endoscopic image and generating a laser ablation path;

[0013] A tracking module for controlling the soft robot body to track the laser ablation path; including:

[0014] An establishment unit for establishing a linearized model of the soft robot body by using the Koopman method;

[0015] A control unit for optimizing the robot control strategy based on the linearized model of the soft robot body and using the MPC method to ensure the tracking accuracy of the laser ablation path;

[0016] An ablation module for adjusting the energy output of the laser ablation system during the path - tracking process by combining the interaction relationship between the blue laser and tissue and the real - time feedback of the OCT imaging system, so that the tissue ablation depth of the area to be ablated is consistent with the target ablation depth.

[0017] Preferably, during the process of laser irradiating tissue, the state changes of the tissue are sequentially divided into three stages: light absorption, heat diffusion, and tissue damage. The corresponding method for pre - constructing the laser - tissue interaction model is as follows:

[0018] For the light - absorption stage, the laser - tissue interaction model is constructed by using the dynamic Monte Carlo method;

[0019] For the heat - diffusion stage, the laser - tissue interaction model is constructed by using the finite - difference element method;

[0020] For the tissue - damage stage, the laser - tissue interaction model is constructed by using the Arrhenius integral.

[0021] Preferably, the process of receiving multiple two - dimensional ablation points planned by a doctor on an endoscopic image and generating a laser ablation path includes:

[0022] Capturing the endoscopic image containing the area to be ablated through a camera and a light source;

[0023] Receiving multiple two - dimensional ablation points planned by a doctor on the endoscopic image and traversing all points on the area to be ablated;

[0024] Generating the laser ablation path by making the energy distribution of each two - dimensional ablation point uniform.

[0025] Preferably, establishing the linearized model of the soft robot body using the Koopman method includes:

[0026] Defining the state dynamics of the soft robot body using a nonlinear system:

[0027]

[0028] where x(t) is the system state of the soft robot body at time t, corresponding to indicating the position of the laser spot; is the state dynamics of the soft robot body at time t; u(t) is the control input, corresponding to the pneumatic control signal; f(⋅) is the nonlinear dynamic function of the system, used to describe the working principle of the pneumatic system;

[0029] For the nonlinear system , using the Koopman operator to map the observation function g(x) of the system from time t to time t + Δt:

[0030]

[0031] Approximating the state dynamics of the soft robot body of the system using a linear system:

[0032]

[0033] where A is the state transition matrix, representing the evolution of the observation function; B is the control input matrix, representing the influence of the pneumatic control signal on the observation function;

[0034] Using the least squares method to solve for A and B to transform the evolution of the system into the following linear state space equation:

[0035]

[0036] where g(x(t + 1)) represents the observation function of the system at time t + 1;

[0037] Taking the linear state space equation as the linearized model of the soft robot body.

[0038] Preferably, based on the linearized model of the soft robot body and using the MPC method to optimize the robot control strategy to ensure the tracking accuracy of the laser ablation path, includes:

[0039] Based on the linearized model of the soft robot body, the state x(t + k) of the system at the next k moments of the state x(t) at time t is calculated by the following prediction equation:

[0040]

[0041] Among them, g(x(t + k + 1)) and g(x(t + k)) are the observation functions of the system at times t + k + 1 and t + k respectively; u(t + k) is the control input at time t + k;

[0042] Combining the deviation of the system state and the energy consumption of the control input, minimizing the first cost function J within the next N time instants is taken as the objective function of the MPC problem:

[0043]

[0044] Among them, p(t + k) is the predicted position of the laser spot at time t + k; p d (t + k) is the desired position of the two-dimensional ablation point on the laser ablation path at time t + k; Q path , R path are both symmetric positive definite weight matrices in the path tracking process, controlling the cost of the path tracking error and the control input respectively, and the subscript path represents path tracking; represents the norm of the vector;

[0045] And set the constraint conditions of the MPC problem:

[0046]

[0047] Among them, u min , u max are the minimum and maximum values of the control input respectively;

[0048] Based on the objective function and constraint conditions of the MPC problem, it is transformed into solving the following optimization problem:

[0049] The objective function is:

[0050]

[0051] Among them, min is the minimization function; J(u) is the first cost function with respect to the variable u;

[0052] The constraint conditions are:

[0053]

[0054] By solving the said optimization problem, the optimal control input sequence u * (t), u * (t + 1), …, u * (t + N - 1) is obtained, and the first optimal control input u * (t) is selected to adjust the motion of the soft robot body in real time to ensure the tracking accuracy of the laser ablation path.

[0055] Preferably, during the path tracking, in combination with the interaction relationship between the blue laser and the tissue and the real-time feedback of the OCT imaging system, the energy output of the laser ablation system is adjusted so that the tissue ablation depth in the to-be-ablated area is consistent with the target ablation depth, including:

[0056] Define the function mapping relationship between the laser energy output per unit time and the tissue depth change:

[0057]

[0058] where, d t is the tissue ablation depth at time t, determined by the tissue depth change real-time feedback of the OCT image; the input variable of the system is composed of the tissue ablation depth d t at time t and the energy output per unit time u t , Δd t = d t+1 - d t is the ablation depth change amount at adjacent moments, which is the output variable of the system; f is the mapping function between the laser energy output per unit time and the tissue depth change;

[0059] Data is obtained through offline sampling, and the Gaussian process regression GPR is used to obtain the dynamic relationship of tissue depth change and establish a non-linear fitting:

[0060]

[0061] where, are respectively the training data sets of the input and output variables obtained from the real environment, and it is considered that the output variable of the system conforms to the Gaussian distribution with a mean of 0 ; K(X,X) is the covariance matrix; I is the identity matrix; σ 2 is the variance of the Gaussian noise;

[0062] Given the training data set and the input variable of the system , construct the following joint probability distribution and predict the depth change of the system under any given laser energy input sequence :

[0063]

[0064] where, K(X,X)+Iσ 2 is the covariance function between the input data set X and itself plus the noise term; is the covariance function between the input data set X and ; is the input variable The covariance function between the input data set X; is the input variable The covariance function between itself and itself;

[0065] Based on the dynamic relationship of the tissue depth change, a loss function is constructed with the target ablation depth, and the tissue ablation depth within the next K time instants is predicted through forward mapping:

[0066]

[0067] where J ' is the second cost function; is the average value of the random variable; d K is the tissue ablation depth within the next K time instants; d d is the target ablation depth; Q depth and S depth are both symmetric positive definite weight matrices in ablation depth control, controlling the cost of ablation depth control error and control input respectively. The subscript depth represents ablation depth control; l(d t ) is the process loss function, and h(d T ) is the terminal loss function; the superscript T represents transpose;

[0068] Minimize the second cost function J ' , determine the optimal tissue ablation depth at each of the next K time instants , and then obtain a set of optimal energy laser sequences. Combining with the depth change , set the laser power at the current moment and adjust the energy output of the laser ablation system in real time;

[0069] Utilize the tissue depth change feedback in real time by the OCT image to update the dynamic relationship of the tissue depth change, and repeat the above optimization process until the tissue ablation depth in the area to be ablated is consistent with the target ablation depth.

[0070] A storage medium stores a computer program for autonomous three-dimensional tissue ablation control based on a soft robot. Among them, the computer program enables a computer to control the autonomous three-dimensional tissue ablation control system described above to execute the autonomous three-dimensional tissue ablation method. The soft robot includes a soft robot main body, a laser ablation system, and an OCT imaging system. The autonomous three-dimensional tissue ablation method includes:

[0071] Pre-construct a laser-tissue interaction model to describe the interaction relationship between the blue laser and the tissue, and determine the target ablation depth of the area to be ablated;

[0072] Receive multiple two-dimensional ablation points planned by a doctor on an endoscopic image and generate a laser ablation path;

[0073] Control the soft robot body to track the laser ablation path, including:

[0074] Establish a linearized model of the soft robot body using the Koopman method;

[0075] Based on the linearized model of the soft robot body, and use the MPC method to optimize the robot control strategy to ensure the tracking accuracy of the laser ablation path;

[0076] During the path tracking process, combine the interaction relationship between the blue laser and the tissue and the real-time feedback of the OCT imaging system to adjust the energy output of the laser ablation system so that the tissue ablation depth in the area to be ablated is consistent with the target ablation depth;

[0077] During the path tracking process, combine the interaction relationship between the blue laser and the tissue and the real-time feedback of the OCT imaging system to adjust the energy output of the laser ablation system so that the tissue ablation depth in the area to be ablated is consistent with the target ablation depth, including:

[0078] Define the function mapping relationship between the laser energy output per unit time and the tissue depth change:

[0079]

[0080] where d t is the tissue ablation depth at time t, determined by the tissue depth change real-time feedback of the OCT image; the input variable of the system is composed of the tissue ablation depth d t at time t and the energy output per unit time u t , Δd t =d t+1 -d t is the ablation depth change amount between adjacent times, which is the output variable of the system; f is the mapping function between the laser energy output per unit time and the tissue depth change;

[0081] Obtain data through offline sampling, use Gaussian process regression GPR to obtain the dynamic relationship of tissue depth change, and establish a non-linear fitting:

[0082]

[0083] where, are the training data sets of the input and output variables obtained from the real environment respectively, and it is considered that the output variable of the system conforms to a Gaussian distribution with a mean of 0 ; K(X,X) is the covariance matrix; I is the identity matrix; σ 2 is the variance of Gaussian noise;

[0084] Given the training data set and the input variables of the system , construct the following joint probability distribution and predict the depth change of the system given any input sequence of laser energy :

[0085]

[0086] where \(K(X,X)+I\sigma\) 2 is the covariance function between the input data set \(X\) and itself plus the noise term; is the covariance function between the input data set \(X\) and ; is the covariance function between the input variable and the input data set \(X\); is the covariance function between the input variable and itself;

[0087] Based on the dynamic relationship of the tissue depth change, and construct a loss function with the target ablation depth, and predict the tissue ablation depth within the next \(K\) time instants through forward mapping:

[0088]

[0089] where \(J\) ' is the second cost function; is the mean value of the random variable; \(d\) K is the tissue ablation depth within the next \(K\) time instants; \(d\) d is the target ablation depth; \(Q\) depth , \(S\) depth are both symmetric positive definite weight matrices in ablation depth control, which control the ablation depth control error and the cost of the control input respectively. The subscript "depth" represents ablation depth control; \(l(d\) t ) is the process loss function, \(h(d\) T ) is the terminal loss function; the superscript \(T\) represents transpose;

[0090] Minimize the second cost function \(J\) ' , determine the optimal tissue ablation depth at each moment within the next \(K\) time instants , and then obtain a set of optimal energy laser sequences. Combine the depth change to set the laser power at the current moment and adjust the energy output of the laser ablation system in real time;

[0091] Utilize the tissue depth change feedback in real time by the OCT image to update the dynamic relationship of the tissue depth change, and repeat the above optimization process until the tissue ablation depth of the area to be ablated is consistent with the target ablation depth.

[0092] An electronic device, comprising:

[0093] One or more processors; a memory; and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, the programs including those for controlling the autonomous three-dimensional tissue ablation control system as described above to execute an autonomous three-dimensional tissue ablation method, the soft robot including a soft robot body, a laser ablation system, and an OCT imaging system, and the autonomous three-dimensional tissue ablation method including:

[0094] Pre-build a laser-tissue interaction model to describe the interaction relationship between the blue laser and the tissue and determine the target ablation depth of the area to be ablated;

[0095] Receive a plurality of two-dimensional ablation points planned by a doctor on an endoscopic image and generate a laser ablation path;

[0096] Control the soft robot body to track the laser ablation path; including:

[0097] Adopt the Koopman method to establish a linearized model of the soft robot body;

[0098] Based on the linearized model of the soft robot body and adopt the MPC method to optimize the robot control strategy to ensure the tracking accuracy of the laser ablation path;

[0099] During the path tracking process, combine the interaction relationship between the blue laser and the tissue and the real-time feedback of the OCT imaging system to adjust the energy output of the laser ablation system so that the tissue ablation depth of the area to be ablated is consistent with the target ablation depth;

[0100] During the path tracking process, combine the interaction relationship between the blue laser and the tissue and the real-time feedback of the OCT imaging system to adjust the energy output of the laser ablation system so that the tissue ablation depth of the area to be ablated is consistent with the target ablation depth, including:

[0101] Define the function mapping relationship between the laser energy output per unit time and the tissue depth change:

[0102]

[0103] wherein, d t is the tissue ablation depth at time t, determined by the tissue depth change real-time feedback of the OCT image; the input variable of the system is composed of the tissue ablation depth d t at time t and the energy output per unit time u t and Δdt = d t+1 - d t is the change in ablation depth at adjacent times and is the output variable of the system; f is the mapping function between the laser energy output per unit time and the tissue depth change;

[0104] Data is obtained through offline sampling, and the Gaussian process regression GPR is used to obtain the dynamic relationship of tissue depth change, and a non-linear fitting is established:

[0105]

[0106] where are the training data sets of the input and output variables obtained from the real environment respectively, and it is considered that the output variable of the system conforms to a Gaussian distribution with a mean of 0 ; K(X,X) is the covariance matrix; I is the identity matrix; σ 2 is the variance of the Gaussian noise;

[0107] Given the training data set and the input variable of the system , construct the following joint probability distribution and predict the depth change of the system under any given laser energy input sequence :

[0108]

[0109] where K(X,X)+Iσ 2 is the covariance function between the input data set X and itself plus the noise term; is the covariance function between the input data set X and ; is the covariance function between the input variable and the input data set X; is the covariance function between the input variable and itself;

[0110] Based on the dynamic relationship of the tissue depth change and with the target ablation depth, construct a loss function, and predict the tissue ablation depth within the next K time instants through forward mapping:

[0111]

[0112] where J ' is the second cost function; is the average value of the random variable; d K is the tissue ablation depth within the next K time instants; d d is the target ablation depth; Q depth 、S depthThey are all symmetric positive definite weight matrices in ablation depth control, respectively controlling the cost of ablation depth control error and control input. The subscript "depth" represents ablation depth control; l(d t ) is the process loss function, and h(d T ) is the terminal loss function; the superscript "T" represents transpose;

[0113] Minimize the second cost function J ' , determine the optimal tissue ablation depth at each moment within the next K moments , and then obtain a set of optimal energy laser sequences. Combine the described depth change to set the laser power at the current moment and adjust the energy output of the laser ablation system in real time;

[0114] Utilize the tissue depth change feedback in real time by the OCT image to update the dynamic relationship of the tissue depth change, and repeat the above optimization process until the tissue ablation depth in the area to be ablated is consistent with the target ablation depth.

[0115] (III) Beneficial Effects

[0116] The present invention provides an autonomous three-dimensional tissue ablation control system based on a soft robot. Compared with the prior art, it has the following beneficial effects:

[0117] To achieve precise ablation of diseased tissues by the laser ablation system in a narrow in-vivo environment, the present invention pre-constructs a laser-tissue interaction model to describe the interaction relationship between the blue laser and the tissue, and thus develops a robust ablation depth control method; in addition, an adaptive laser ablation path tracking method combining the soft robot body and the OCT imaging system is innovatively proposed; finally, a surgical robot system with autonomous precise three-dimensional laser ablation ability is constructed, which can significantly improve the accuracy, safety and automation level of laser ablation surgery. Description of the Drawings

[0118] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0119] Figure 1 It is a schematic structural diagram of a laser ablation system provided by an embodiment of the present invention;

[0120] Figure 2 It is an integrated schematic diagram of an OCT imaging optical fiber, a laser ablation optical fiber, and an indication laser optical fiber provided by an embodiment of the present invention;

[0121] Figure 3 This is a schematic structural diagram of an autonomous three-dimensional tissue ablation control system based on a soft robot provided by an embodiment of the present invention. Detailed implementation manners

[0122] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention are clearly and completely described. Apparently, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0123] The embodiment of the present application provides an autonomous three-dimensional tissue ablation control system based on a soft robot, and solves the technical problems of large path tracking error and insufficient dynamic control of ablation depth.

[0124] The technical solutions in the embodiments of the present application to solve the above technical problems are generally as follows:

[0125] First, supplement the term explanations:

[0126] 1) Optical coherence tomography (OCT for short) is a three-dimensional tomography technology. Based on the principle of low-coherence interference, it obtains the tomographic ability in the depth direction. Through scanning, two-dimensional or three-dimensional images of the internal structure of biological tissues or materials can be reconstructed. Its signal contrast originates from the spatial variation of the optical reflection (scattering) characteristics inside biological tissues or materials.

[0127] 2) Model Predictive Control (MPC for short) is a model-based closed-loop optimization control strategy, which is widely used in fields such as autonomous driving and path planning. MPC optimizes the control input by predicting the behavior of the system in the future for a period of time, making the system as close as possible to the target state while satisfying various constraint conditions.

[0128] 3) Gaussian Process Regression (GPR for short) is a non-parametric model that uses a Gaussian Process (GP for short) prior to perform regression analysis on data. Its model assumptions include noise (regression residuals) and a Gaussian process prior, and its solution is carried out according to Bayesian inference.

[0129] In view of the deficiencies of the existing laser ablation technology, the embodiments of the present invention construct a set of laser ablation surgical systems to achieve precise control of the ablation depth. The core points involved at least include:

[0130] (1) A robust laser ablation depth control algorithm based on real-time OCT feedback

[0131] By combining the interaction mechanism between blue laser and tissue and OCT image feedback, the ablation depth is monitored and adjusted in real time. And the interaction relationship between tissue and laser is learned online to update the system model in real time to predict the ablation depth, so as to dynamically adjust the laser parameters to achieve robust ablation depth control.

[0132] (2) A laser ablation path tracking algorithm based on soft robots

[0133] Aiming at the problem of insufficient combination of the flexible structure and precise control of soft robots, the method of Koopman plus MPC is adopted. The Koopman operator can convert a complex nonlinear system into a high-dimensional linear system representation, automatically capture complex dynamic characteristics such as material hysteresis in a pure data-driven manner, avoid the difficulties of traditional physical modeling, and support online model update to adapt to system parameters and environmental changes; at the control level, the predictive optimization characteristics of MPC can plan the optimal trajectory in advance, explicitly consider system constraints such as actuator limitations, and balance the tracking accuracy and control efficiency through multi-objective optimization.

[0134] It should be particularly emphasized that the autonomous three-dimensional tissue ablation control system provided by the embodiments of the present invention relies on soft robots. Specifically, the soft robot includes a soft robot body, a laser ablation system, and an OCT imaging system; where:

[0135] The soft robot body is accurately positioned by the beam emitted by the indicating laser fiber, performs flexible path tracking in space, and then moves the laser to perform laser ablation operations. Exemplarily, the soft robot body uses a pneumatic drive mechanism to make the movement of the robot flexible and adaptable, and can freely bend and stretch in a narrow and complex tissue environment.

[0136] The laser ablation system provides a high-energy laser beam through a laser to ablate the diseased tissue, that is, the area to be ablated. In theory, the laser ablation system can precisely control the power, frequency, and irradiation time of the laser to ensure precise control of the ablation depth. Exemplarily, the tissue ablation depth is adjusted by controlling the laser power here.

[0137] The OCT imaging system includes an OCT imaging fiber, which uses optical coherence tomography (OCT) technology to provide real-time tissue imaging feedback. The system can obtain high-resolution three-dimensional images of the diseased tissue and provide feedback information for real-time depth control.

[0138] Furthermore, Figure 1 a feasible laser ablation system is shown. In Figure 1 , the system consists of a passive bending part, an active bending part, and a laser deflector, and can be pre-placed in a designated ablation area (such as the laryngeal tumor in the figure). Then, by controlling the movement of the pneumatic soft robot, the laser ablation optical fiber of the laser ablation system is deflected to achieve tissue ablation.

[0139] Furthermore, Figure 2 a feasible integration method of an OCT imaging optical fiber, a laser ablation optical fiber, and an indicating laser optical fiber is shown. The laser beam irradiates the sample through the OCT imaging optical fiber. The tiny structures inside the sample reflect the beam back. The optical fiber receives the reflected light signal and transmits it to the imaging system, and finally forms a tomographic image of the sample. The laser ablation optical fiber conducts the laser beam to the target tissue, and through the thermal effect of the laser, raises the tissue temperature to a level sufficient to destroy the diseased tissue. The indicating laser optical fiber emits a visible red laser beam to clearly identify the position of the operation area or the diseased tissue, thereby ensuring the accuracy and safety during the treatment process.

[0140] In addition, in Figure 2 , it is also shown that the beams emitted by the ablation laser optical fiber and the indicating laser optical fiber intersect at a certain point. This intersection point is the position where the laser and the indicating laser act together, which ensures that the ablation surgery can accurately act on the area to be ablated.

[0141] To better understand the above technical solutions, the above technical solutions will be described in detail below in conjunction with the accompanying drawings of the specification and specific implementation manners.

[0142] Example 1:

[0143] As Figure 3 shown, an embodiment of the present invention provides an autonomous three-dimensional tissue ablation control system based on a soft robot. The soft robot includes a soft robot main body, a laser ablation system, and an OCT imaging system, and includes:

[0144] a construction module for pre-constructing a laser-tissue interaction model to describe the interaction relationship between the blue laser and the tissue, and determining the target ablation depth of the area to be ablated;

[0145] a planning module for receiving multiple two-dimensional ablation points planned by a doctor on an endoscopic image and generating a laser ablation path;

[0146] a tracking module for controlling the soft robot main body to track the laser ablation path, including:

[0147] a establishing unit for establishing a linearized model of the soft robot main body using the Koopman method;

[0148] A control unit, configured to optimize the robot control strategy based on the linearized model of the soft robot body and by using the MPC method to ensure the tracking accuracy of the laser ablation path;

[0149] An ablation module, configured to adjust the energy output of the laser ablation system during path tracking by combining the interaction relationship between the blue laser and the tissue and the real-time feedback of the OCT imaging system, so that the tissue ablation depth of the region to be ablated is consistent with the target ablation depth.

[0150] To achieve precise ablation of diseased tissues in a narrow in-vivo environment by the laser ablation system, the embodiments of the present invention pre-construct a laser-tissue interaction model to describe the interaction relationship between the blue laser and the tissue, thereby developing a robust ablation depth control method; in addition, an adaptive laser ablation path tracking method combining the soft robot body and the OCT imaging system is innovatively proposed; finally, a surgical robot system with autonomous precise three-dimensional laser ablation capabilities is constructed, which can significantly improve the accuracy, safety and automation level of laser ablation surgery.

[0151] Next, each component of the above system and its working principle will be introduced in detail:

[0152] For the construction module, it is configured to pre-construct a laser-tissue interaction model to describe the interaction relationship between the blue laser and the tissue and determine the target ablation depth of the region to be ablated.

[0153] The robust laser ablation depth control algorithm based on the real-time feedback of learning-based OCT needs to online learn the interaction relationship between the tissue and the laser, update the system model in real time to predict the ablation depth, and thus dynamically adjust the laser parameters to achieve robust ablation depth control.

[0154] Specifically, the related steps of the pre-constructed laser-tissue interaction model include:

[0155] During the process of laser irradiating the tissue, the state change of the tissue is sequentially divided into three stages: light absorption, heat diffusion, and tissue damage. The corresponding pre-construction method of the laser-tissue interaction model is:

[0156] (1) For the light absorption stage, the dynamic Monte Carlo method is used to construct the laser-tissue interaction model.

[0157] The simulation results of the absorption and scattering of a large number of photons in the tissue are statistically obtained to get the distribution S(r,z,t) of the laser energy in the tissue. The absorption and scattering processes of photons are equivalent to the photon energy being respectively in proportion μ a / (μ a +μ s ) at the collision point, μs / (μ a + μ s ) is absorbed and scattered by the node region where it is located. Among them, r and z are the radial and axial coordinates based on the cylindrical coordinate system, t is the time, and μ a , μ s are the tissue absorption and scattering parameters respectively, representing the absorption and scattering capabilities of the tissue for laser energy.

[0158] (2) For the heat diffusion stage, the finite difference element method is used to construct the laser tissue interaction model.

[0159] That is, according to the light absorption distribution function S(r, z, t), based on the constraint equation formula (1), the finite difference method is used to solve the heat diffusion equation to obtain the time and space distributions of the heat deposition H(r, z, t) and the tissue temperature T(r, z, t) in the tissue:

[0160]

[0161] Among them, represents the partial derivative of the heat deposition H(r, z, t) with respect to the time t, is the partial derivative symbol; β is the tissue thermal conductivity, representing the ability of the tissue to conduct heat; is called the Laplace operator, describes the second-order spatial derivative of the tissue temperature T(r, z, t) in the cylindrical coordinate system.

[0162] (3) For the tissue damage stage, the Arrhenius integral is used to construct the laser tissue interaction model.

[0163] That is, the Arrhenius integral Ω(r, z, t) is used to quantify the tissue damage parameter α(r, z, t):

[0164]

[0165] Among them, Ω(r, z, t) is used to describe the cumulative damage effect of the tissue under thermal action; C(r, z, t) is the remaining proportion of undamaged tissue, representing the proportion of undamaged tissue at the position (r, z) and time t; C(r, z, 0) is the proportion of undamaged tissue at the initial moment; A is the frequency factor, representing the reaction rate; E a is the activation energy, representing the energy required for tissue damage; R is the gas constant.

[0166] When α > 1, the tissue is considered to be solidified, and when H(r, z, t) reaches the threshold, the tissue is removed and becomes air, thereby updating the tissue model. Considering the dynamic changes of tissue characteristics during the laser transmission process, a dynamic change model of the tissue ablation process based on tissue absorption and scattering parameters is constructed to simulate the real changes of tissue characteristics. The relevant parameters are updated through in vitro experiments to approximate the real ablation process, so as to analyze the influence of different blue laser parameters.

[0167] For the planning module, it is used to receive multiple two-dimensional ablation points planned by the doctor on the endoscopic image and generate a laser ablation path.

[0168] The endoscopic image of the diseased tissue is captured by an external endoscopic system through a high-resolution camera and a light source. During the operation, the endoscopic probe can be brought close to the diseased tissue, and two-dimensional plane images can be obtained in video stream or static image mode. Moreover, the goal of the path planning in the embodiment of the present invention is to set the trajectory information on the two-dimensional plane, so as to facilitate subsequent path tracking by the soft robot body using MPC control.

[0169] Specifically, the related steps of generating the laser ablation path include:

[0170] First, capture the endoscopic image containing the area to be ablated through the camera and the light source.

[0171] Secondly, receive multiple two-dimensional ablation points planned by the doctor on the endoscopic image and traverse all points on the area to be ablated.

[0172] Finally, generate the laser ablation path by making the energy distribution of each two-dimensional ablation point uniform.

[0173] Exemplarily, making the energy distribution of each two-dimensional ablation point uniform can be achieved by using the following objective function:

[0174]

[0175] where min is the minimization function; p is the path parameter of the laser ablation path; n is the number of two-dimensional ablation points; z i is the target ablation depth of the i-th two-dimensional ablation point; E i is the energy distribution error of the i-th two-dimensional ablation point; α and β are the weight coefficients for adjusting the depth accuracy and the energy distribution uniformity respectively.

[0176] For the tracking module, it is used to control the soft robot body to track the laser ablation path; it includes:

[0177] A building unit, which is used to establish a linearized model of the soft robot body by using the Koopman method;

[0178] A control unit, configured to optimize the robot control strategy based on the linearized model of the soft robot body and using the MPC method to ensure the tracking accuracy of the laser ablation path.

[0179] Specifically:

[0180] In the establishment unit, perform linear modeling of the pneumatic soft robot body. Specifically, use the Koopman method to establish the linearized model of the soft robot body.

[0181] The pneumatic soft robot system has strong non - linear characteristics. On the premise of selecting the pneumatic soft robot, the control of the soft robot body refers to realizing the deformation of the robot by adjusting the air pressure, and then controlling the turning of the laser.

[0182] Specifically, the related steps of the linear modeling include:

[0183] First, define the state dynamics of the soft robot body using a non - linear system:

[0184]

[0185] Where x(t) is the system state of the soft robot body at time t, corresponding to indicating the position of the laser point; is the state dynamics of the soft robot body at time t; u(t) is the control input, corresponding to the air pressure control signal, for example, including the motor step signal corresponding to the pressure values [P1, P2, P3] of each air chamber; f(⋅) is the non - linear dynamic function of the system, used to describe the working principle of the pneumatic system.

[0186] Secondly, for the non - linear system , use the Koopman operator to map the observation function g(x) of the system from time t to time t + Δt:

[0187]

[0188] Thirdly, approximate the state dynamics of the soft robot body of the system using a linear system:

[0189]

[0190] Where A is the state transition matrix, representing the evolution of the observation function; B is the control input matrix, representing the influence of the air pressure control signal on the observation function.

[0191] Then, use the least - squares method to solve for A and B to transform the evolution of the system into the following linear state - space equation:

[0192]

[0193] Among them, g(x(t + 1)) represents the observation function of the system at time t + 1.

[0194] Specifically, the use of the least squares method to solve for A and B specifically means:

[0195] For the evolution matrix of the observation function, i.e., the observation function matrix G, it is desired to estimate A and B by minimizing the following objective function:

[0196]

[0197] where G t+1 represents the value of the observation function at the next time step, G t is the observation function matrix at the current time step, and U is the control input matrix. It is desired to find A and B to minimize this error. Least squares solution: Solve this optimization problem using the least squares method to obtain:

[0198]

[0199] As shown in formula (10), A and B are calculated through the pseudoinverse of the observation function matrix. Specifically, among them represents the transpose of matrix Gt, represents the inverse matrix of the matrix product . The entire expression constructs an optimal linear mapping relationship from the observation function matrix Gt at the current time step to the observation function matrix Gt+1 at the next time step, thereby minimizing the sum of the squares of the prediction errors, which is the core step of linearizing the dynamic model of the soft robot by the Koopman method.

[0200] Finally, take the linear state space equation as the linearized model of the soft robot body, which describes how the state g(x(t + 1)) of the system at the next time step depends on the current state g(x(t)) and the control input u(t).

[0201] In the control unit, path tracking control is performed. Specifically, the MPC (Model Predictive Control) method is used to optimize the robot control strategy to ensure the tracking accuracy of the laser ablation path.

[0202] To implement MPC, the Koopman model is used above to predict the future state x(t + k), thereby achieving optimal control of the system. The core idea of MPC is to optimize the control input u(t) within a future period of time so that the system state x(t) reaches the desired target and is adjusted within the control constraints.

[0203] Specifically, the relevant steps of the path tracking control include:

[0204] First, based on the linearized model of the soft robot body, the state x(t + k) of the system at the next k moments in the future of the state x(t) at time t is calculated by the following prediction equation:

[0205]

[0206] where g(x(t + k + 1)) and g(x(t + k)) are the observation functions of the system at times t + k + 1 and t + k respectively; u(t + k) is the control input at time t + k.

[0207] Secondly, by synthesizing the deviation of the system state and the energy consumption of the control input, minimizing the first cost function J within the next N moments is used as the objective function of the MPC problem:

[0208]

[0209] where p(t + k) is the predicted position of the indicating laser spot at time t + k; p d (t + k) is the desired position of the two-dimensional ablation point on the laser ablation path corresponding to time t + k; Q path , R path are both symmetric positive definite weight matrices in the path tracking process, controlling the path tracking error and the cost of the control input respectively, and the subscript path represents path tracking; represents the norm of the vector.

[0210] And set the constraint conditions of the MPC problem:

[0211]

[0212] where u min , u max are the minimum and maximum values of the control input respectively, representing that the control input is restricted by the maximum air cavity pressure, such as restricting the size of the step signal.

[0213] Thirdly, based on the objective function and constraint conditions of the MPC problem, it is transformed into solving the following optimization problem:

[0214] The objective function is:

[0215]

[0216] where min is the minimization function; J(u) is the first cost function with respect to the variable u.

[0217] The constraint conditions are:

[0218]

[0219] Finally, by solving the optimization problem, the optimal control input sequence u * (t), u * (t + 1), …, u * (t + N - 1) is obtained, and the first optimal control input u * (t) is used to adjust the motion of the soft robot body in real time to ensure the tracking accuracy of the laser ablation path.

[0220] For the ablation module, it is used to adjust the energy output of the laser ablation system during the path tracking process by combining the interaction relationship between the blue laser and the tissue and the real-time feedback of the OCT imaging system, so that the tissue ablation depth of the area to be ablated is consistent with the target ablation depth.

[0221] In this module, based on the study of the laser-tissue interaction model, the ablation depth can be predicted through real-time OCT data combined with Gaussian process regression, and the ablation depth control can be carried out while using MPC for trajectory tracking.

[0222] Specifically, the related steps of the ablation depth control include:

[0223] First, define the function mapping relationship between the laser energy output per unit time and the tissue depth change:

[0224]

[0225] where d t is the tissue ablation depth at time t, which is determined by the tissue depth change real-time feedback of the OCT image; the input variable of the system is composed of the tissue ablation depth d t at time t and the energy output per unit time u t , Δd t = d t+1 - d t is the ablation depth change amount at adjacent times, which is the output variable of the system; f is the mapping function between the laser energy output per unit time and the tissue depth change.

[0226] Secondly, data is obtained through offline sampling, and the Gaussian process regression GPR is used to obtain the dynamic relationship of the tissue depth change and establish a non-linear fitting:

[0227]

[0228] where are the training data sets of the input and output variables obtained from the real environment respectively, and it is considered that the output variable of the system conforms to the Gaussian distribution with a mean of 0 ; K(X, X) is the covariance matrix; I is the identity matrix; σ 2 is the variance of Gaussian noise.

[0229] Given the training data set and the input variables of the system , construct the following joint probability distribution and predict the depth change of the system under any given laser energy input sequence :

[0230]

[0231] where, K(X, X) + Iσ 2 is the covariance function between the input data set X and itself plus the noise term; is the covariance function between the input data set X and ; is the input variable and the covariance function between the input data set X; is the input variable and the covariance function between itself.

[0232] Again, based on the dynamic relationship of the tissue depth change, and construct a loss function with the target ablation depth, and predict the tissue ablation depth within the next K time instants through forward mapping:

[0233]

[0234] where, J ' is the second cost function; is the average value of the random variable; d K is the tissue ablation depth within the next K time instants; d d is the target ablation depth; Q depth , S depth are both symmetric positive definite weight matrices in ablation depth control, controlling the cost of ablation depth control error and control input respectively, the subscript depth represents ablation depth control; l(d t ) is the process loss function, h(d T ) is the terminal loss function; the superscript T represents transpose.

[0235] Then, minimize the second cost function J ' , determine the optimal tissue ablation depth at each of the next K time instants , and then obtain a set of optimal energy laser sequences, combined with the depth change , to set the laser power at the current moment and adjust the energy output of the laser ablation system in real time.

[0236] Finally, using the tissue depth change feedback in real time by the OCT image, update the dynamic relationship of the tissue depth change, and repeat the above optimization process until the tissue ablation depth in the ablation area to be ablated is consistent with the target ablation depth.

[0237] It should be specifically noted that in the embodiment of the present invention: the two-dimensional coordinates (x, y) in the endoscope coordinate system and the three-dimensional coordinates (x ' , y ' , d) in the (x ' , y ' ) correspond to the same physical point. Specifically, the three-dimensional coordinates of the OCT camera can be mapped to the two-dimensional coordinates of the endoscope camera through the coordinate conversion relationship, so as to ensure that the corresponding tissue depth control is realized at the two-dimensional ablation point of the laser ablation path.

[0238] So far, the embodiment of the present invention has introduced the complete content of the autonomous three-dimensional tissue ablation control system based on the soft robot.

[0239] Embodiment 2:

[0240] The embodiment of the present invention provides a storage medium, which stores a computer program for autonomous three-dimensional tissue ablation control based on a soft robot. Among them, the computer program enables a computer to control the autonomous three-dimensional tissue ablation control system described in Embodiment 1 to execute an autonomous three-dimensional tissue ablation method. The soft robot includes a soft robot body, a laser ablation system, and an OCT imaging system. The autonomous three-dimensional tissue ablation method includes:

[0241] Pre-construct a laser-tissue interaction model to describe the interaction relationship between the blue laser and the tissue, and determine the target ablation depth of the ablation area to be ablated;

[0242] Receive multiple two-dimensional ablation points planned by a doctor on the endoscope image, and generate a laser ablation path;

[0243] Control the soft robot body to track the laser ablation path; including:

[0244] Adopt the Koopman method to establish a linearized model of the soft robot body;

[0245] Based on the linearized model of the soft robot body, and adopt the MPC method to optimize the robot control strategy to ensure the tracking accuracy of the laser ablation path;

[0246] During the path tracking process, combine the interaction relationship between the blue laser and the tissue and the real-time feedback of the OCT imaging system to adjust the energy output of the laser ablation system, so that the tissue ablation depth in the ablation area to be ablated is consistent with the target ablation depth;

[0247] During the path tracking process, in combination with the interaction relationship between the blue laser and the tissue and the real-time feedback of the OCT imaging system, adjust the energy output of the laser ablation system so that the tissue ablation depth of the area to be ablated is consistent with the target ablation depth, including:

[0248] Define the function mapping relationship between the laser energy output per unit time and the tissue depth change:

[0249]

[0250] where d t is the tissue ablation depth at time t, determined by the tissue depth change real-time feedback of the OCT image; the input variable of the system is composed of the tissue ablation depth d t at time t and the energy output per unit time u t , Δd t =d t+1 -d t is the ablation depth change amount between adjacent moments and is the output variable of the system; f is the mapping function between the laser energy output per unit time and the tissue depth change;

[0251] Obtain data through offline sampling, use Gaussian process regression GPR to obtain the dynamic relationship of tissue depth change, and establish a non-linear fitting:

[0252]

[0253] where are the training data sets of the input and output variables obtained from the real environment respectively, and it is considered that the output variable of the system conforms to a Gaussian distribution with a mean of 0 ; K(X,X) is the covariance matrix; I is the identity matrix; σ 2 is the variance of the Gaussian noise;

[0254] Given the training data set and the input variable of the system , construct the following joint probability distribution and predict the depth change of the system under any given laser energy input sequence :

[0255]

[0256] where K(X,X)+Iσ 2 is the covariance function between the input data set X and itself plus the noise term; is the covariance function between the input data set X and ; is the input variable The covariance function between the input data set X; is the input variable The covariance function between itself;

[0257] Based on the dynamic relationship of the tissue depth change, and constructing a loss function with the target ablation depth, predicting the tissue ablation depth within the next K time instants through forward mapping:

[0258]

[0259] where J ' is the second cost function; is the mean value of the random variable; d K is the tissue ablation depth within the next K time instants; d d is the target ablation depth; Q depth and S depth are both symmetric positive definite weight matrices in the ablation depth control, controlling the cost of the ablation depth control error and the control input respectively. The subscript depth represents the ablation depth control; l(d t ) is the process loss function, and h(d T ) is the terminal loss function; the superscript T represents the transpose;

[0260] Minimize the second cost function J ' , determine the optimal tissue ablation depth at each of the next K time instants , and then obtain a set of optimal energy laser sequences. Combining the depth change , to set the laser power at the current time instant and adjust the energy output of the laser ablation system in real time;

[0261] Utilize the tissue depth change feedback in real time by the OCT image to update the dynamic relationship of the tissue depth change, and repeat the above optimization process until the tissue ablation depth of the region to be ablated is consistent with the target ablation depth.

[0262] Example 3:

[0263] The embodiment of the present invention provides an electronic device, including:

[0264] One or more processors; a memory; and one or more programs, where the one or more programs are stored in the memory and configured to be executed by the one or more processors. The program includes controlling the autonomous three-dimensional tissue ablation control system as described in Example 1 to execute the autonomous three-dimensional tissue ablation method. The soft robot includes a soft robot body, a laser ablation system, and an OCT imaging system. The autonomous three-dimensional tissue ablation method includes:

[0265] Pre - build a laser - tissue interaction model to describe the interaction relationship between blue laser and tissue, and determine the target ablation depth of the area to be ablated;

[0266] Receive multiple two - dimensional ablation points planned by the doctor on the endoscopic image and generate a laser ablation path;

[0267] Control the soft robot body to track the laser ablation path; including:

[0268] Use the Koopman method to establish a linearized model of the soft robot body;

[0269] Based on the linearized model of the soft robot body and use the MPC method to optimize the robot control strategy to ensure the tracking accuracy of the laser ablation path;

[0270] During the path - tracking process, combine the interaction relationship between the blue laser and tissue and the real - time feedback of the OCT imaging system to adjust the energy output of the laser ablation system so that the tissue ablation depth of the area to be ablated is consistent with the target ablation depth;

[0271] The process of combining the interaction relationship between the blue laser and tissue and the real - time feedback of the OCT imaging system during the path - tracking process to adjust the energy output of the laser ablation system so that the tissue ablation depth of the area to be ablated is consistent with the target ablation depth includes:

[0272] Define the function mapping relationship between the laser energy output per unit time and the tissue depth change:

[0273]

[0274] where, d t is the tissue ablation depth at time t, determined by the tissue depth change real - time feedback of the OCT image; the input variable of the system is composed of the tissue ablation depth d t at time t and the energy output per unit time u t , Δd t = d t+1 - d t is the ablation depth change amount between adjacent times and is the output variable of the system; f is the mapping function between the laser energy output per unit time and the tissue depth change;

[0275] Obtain data through offline sampling, use Gaussian process regression (GPR) to obtain the dynamic relationship of tissue depth change, and establish a non - linear fitting:

[0276]

[0277] where, They are respectively training data sets of input and output variables obtained from the real environment, and it is considered that the output variable of the system conforms to a Gaussian distribution with a mean of 0. ; K(X, X) is the covariance matrix; I is the identity matrix; σ 2 is the variance of the Gaussian noise;

[0278] Given the training data set and the input variable of the system , construct the following joint probability distribution and predict the depth change of the system under any given laser energy input sequence. :

[0279]

[0280] where K(X, X)+Iσ 2 is the covariance function between the input data set X and itself plus the noise term; is the covariance function between the input data set X and ; is the input variable and the covariance function between the input data set X; is the input variable and the covariance function between itself;

[0281] Based on the dynamic relationship of the tissue depth change, and construct a loss function with the target ablation depth, and predict the tissue ablation depth within the next K moments through forward mapping:

[0282]

[0283] where J ' is the second cost function; is the average value of the random variable; d K is the tissue ablation depth within the next K moments; d d is the target ablation depth; Q depth , S depth are both symmetric positive definite weight matrices in ablation depth control, which respectively control the cost of ablation depth control error and control input. The subscript depth represents ablation depth control; l(d t ) is the process loss function, h(d T ) is the terminal loss function; the superscript T represents transpose;

[0284] Minimize the second cost function J ' , determine the optimal tissue ablation depth at each moment within the next K moments , and then obtain a set of optimal energy laser sequences, combined with the depth change , to set the laser power at the current moment and adjust the energy output of the laser ablation system in real time;

[0285] Utilize the tissue depth change real-time feedback of the OCT image to update the dynamic relationship of the tissue depth change, and repeat the above optimization process until the tissue ablation depth of the to-be-ablated area is consistent with the target ablation depth.

[0286] In summary, compared with the prior art, the following beneficial effects are achieved:

[0287] 1. The embodiments of the present invention integrate high-resolution imaging technologies, including endoscopic imaging and OCT imaging. OCT provides high-resolution depth information, while endoscopic images can provide more intuitive surface images. It provides real-time visualization of the surgical area, and combines an online learning algorithm to monitor and feedback control the ablation process in real time, ensuring the controllability of the ablation depth and range.

[0288] 2. The embodiments of the present invention utilize the flexibility and adaptability of the soft robot to develop the hardware structure of the soft robot that can freely navigate in narrow and complex natural cavities, and establish the dynamic model of the soft robot and realize trajectory tracking control to ensure that the laser ablation system can accurately reach the designated to-be-ablated area.

[0289] It should be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variant thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements not only includes those elements, but also includes other elements not expressly listed, or also includes elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "including a..." does not exclude the existence of additional identical elements in the process, method, article or device including the element.

[0290] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An autonomous three-dimensional tissue ablation control system based on soft robots, characterized in that, The soft robot includes a soft robot body, a laser ablation system, and an OCT imaging system, including: A construction module for pre-constructing a laser tissue interaction model to describe the interaction relationship between the blue laser and the tissue, and determining the target ablation depth of the area to be ablated. A planning module for receiving multiple two-dimensional ablation points planned by a doctor on an endoscopic image and generating a laser ablation path. A tracking module for controlling the soft robot body to track the laser ablation path, including: An establishment unit for establishing a linearized model of the soft robot body using the Koopman method. A control unit for optimizing the robot control strategy based on the linearized model of the soft robot body and using the MPC method to ensure the tracking accuracy of the laser ablation path. An ablation module for adjusting the energy output of the laser ablation system during path tracking by combining the interaction relationship between the blue laser and the tissue and the real-time feedback of the OCT imaging system, so that the tissue ablation depth of the area to be ablated is consistent with the target ablation depth. During path tracking, combining the interaction relationship between the blue laser and the tissue and the real-time feedback of the OCT imaging system to adjust the energy output of the laser ablation system, so that the tissue ablation depth of the area to be ablated is consistent with the target ablation depth, including: Defining the function mapping relationship between the laser energy output per unit time and the tissue depth change. Among them, d t is the tissue ablation depth at time t, determined by the tissue depth change real-time feedback of the OCT image; the input variable of the system is composed of the tissue ablation depth d t at time t and the output energy u per unit time t , Δd t = d t+1 - d t is the ablation depth change amount at adjacent times and is the output variable of the system; f is the mapping function between the laser energy output per unit time and the tissue depth change; Obtaining the dynamic relationship of tissue depth change through offline sampling, using Gaussian process regression (GPR) to obtain the dynamic relationship of tissue depth change, and establishing a non-linear fitting. Among them, are respectively the training data sets of the input and output variables obtained from the real environment, and it is considered that the output variable of the system conforms to a Gaussian distribution with a mean of 0 ; K(X,X) is the covariance matrix; I is the identity matrix; σ 2 is the variance of the Gaussian noise; Given the training data set and the input variables of the system , construct the following joint probability distribution and predict the depth change of the system given any input sequence of laser energy : Among them, \(K(X,X)+I\sigma\) 2 is the covariance function between the input data set \(X\) and itself plus the noise term; is the covariance function between the input data set \(X\) and ; is the covariance function between the input variable and the input data set \(X\); is the covariance function between the input variable and itself; Based on the dynamic relationship of tissue depth change and constructing a loss function with the target ablation depth, predicting the tissue ablation depth within the next K time instants through forward mapping. Among them, J ' is the second cost function; is the average value of the random variable; d K is the tissue ablation depth within the next K time instants; d d is the target ablation depth; Q depth , S depth are both symmetric positive definite weight matrices in ablation depth control, controlling the cost of ablation depth control error and control input respectively. The subscript depth represents ablation depth control; l(d t ) is the process loss function, h(d T ) is the terminal loss function; the superscript T represents transpose; Minimize the second cost function J ' , and determine the optimal tissue ablation depth at each moment within the next K moments , and further obtain a set of optimal energy laser sequences, combined with the depth change , to set the laser power at the current moment and adjust the energy output of the laser ablation system in real time; Using the tissue depth change real-time feedback from the OCT image to update the dynamic relationship of tissue depth change, and repeating the above optimization process until the tissue ablation depth of the area to be ablated is consistent with the target ablation depth.

2. The autonomous three-dimensional tissue ablation control system according to claim 1, wherein During the process of laser irradiating tissue, the state changes of the tissue are sequentially divided into three stages: light absorption, heat diffusion, and tissue damage. The corresponding pre-construction method of the laser tissue interaction model is: For the light absorption stage, using the dynamic Monte Carlo method to construct the laser tissue interaction model. For the heat diffusion stage, using the finite difference element method to construct the laser tissue interaction model. For the tissue damage stage, using the Arrhenius integral to construct the laser tissue interaction model.

3. The autonomous three-dimensional tissue ablation control system according to claim 1, wherein Receiving multiple two-dimensional ablation points planned by a doctor on an endoscopic image and generating a laser ablation path, including: Capturing the endoscopic image containing the area to be ablated through a camera and a light source. Receiving multiple two-dimensional ablation points planned by the doctor on the endoscopic image and traversing all points on the area to be ablated. Generating the laser ablation path by making the energy distribution of each two-dimensional ablation point uniform.

4. The autonomous three-dimensional tissue ablation control system according to claim 1, wherein Using the Koopman method to establish a linearized model of the soft robot body, including: Defining the state dynamics of the soft robot body using a non-linear system. where \(x(t)\) is the system state of the soft robot body at time \(t\), corresponding to the position of the indicated laser point; is the state dynamics of the soft robot body at time \(t\); \(u(t)\) is the control input, corresponding to the pneumatic control signal; \(f(\cdot)\) is the nonlinear dynamic function of the system, used to describe the working principle of the pneumatic system; For a non-linear system , the Koopman operator is used to map the system's observation function g(x) from time t to time t + Δt: Approximate the state dynamics of the soft robot body of the system using a linear system: Among them, A is the state transition matrix, representing the evolution of the observation function; B is the control input matrix, representing the influence of the pneumatic control signal on the observation function; Solve for A and B using the least squares method to transform the evolution of the system into the following linear state space equation: Among them, g(x(t + 1)) represents the observation function of the system at time t + 1; Take the linear state space equation as the linearization model of the soft robot body.

5. The autonomous three-dimensional tissue ablation control system according to claim 4, characterized in that, Based on the linearization model of the soft robot body and using the MPC method to optimize the robot control strategy to ensure the tracking accuracy of the laser ablation path, including: Based on the linearization model of the soft robot body, the state x(t + k) of the system at the next k time instants of the system state x(t) at time t is calculated by the following prediction equation: Among them, g(x(t + k + 1)) and g(x(t + k)) are the observation functions of the system at times t + k + 1 and t + k respectively; u(t + k) is the control input at time t + k; Integrate the deviation of the system state and the energy consumption of the control input to minimize the first cost function J within the next N time instants as the objective function of the MPC problem: where p(t + k) is the predicted position of the laser point at time t + k; p d (t + k) is the desired position of the two-dimensional ablation point corresponding to the laser ablation path at time t + k; Q path , R path are both symmetric positive definite weight matrices during the path tracking process, controlling the path tracking error and the cost of the control input respectively. The subscript path represents path tracking; represents the norm of the vector; And set the constraint conditions of the MPC problem: where u min and u max are the minimum and maximum values of the control input, respectively; Based on the objective function and constraint conditions of the MPC problem, convert it into the following optimization problem to solve: The objective function is: Among them, min is the minimization function; J(u) is the first cost function with respect to the variable u; The constraint conditions are: By solving the optimization problem, an optimal control input sequence u * (t), u * (t + 1), …, u * (t + N - 1) is obtained, and the motion of the soft robot body adjusted in real time by selecting the first optimal control input u * (t) is used to ensure the tracking accuracy of the laser ablation path.

6. A storage medium, characterized in that, It stores a computer program for autonomous three-dimensional tissue ablation control based on a soft robot, wherein the computer program causes a computer to control the autonomous three-dimensional tissue ablation control system according to claim 1 to execute an autonomous three-dimensional tissue ablation method. The soft robot includes a soft robot body, a laser ablation system, and an OCT imaging system. The autonomous three-dimensional tissue ablation method includes: Pre-construct a laser-tissue interaction model to describe the interaction relationship between the blue laser and the tissue, and determine the target ablation depth of the area to be ablated; Receive multiple two-dimensional ablation points planned by a doctor on an endoscopic image and generate a laser ablation path; Control the soft robot body to track the laser ablation path; including: Establish a linearization model of the soft robot body using the Koopman method; Based on the linearization model of the soft robot body and using the MPC method to optimize the robot control strategy to ensure the tracking accuracy of the laser ablation path; During the path tracking process, combine the interaction relationship between the blue laser and the tissue and the real-time feedback of the OCT imaging system to adjust the energy output of the laser ablation system so that the tissue ablation depth of the area to be ablated is consistent with the target ablation depth; During the path tracking process, combine the interaction relationship between the blue laser and the tissue and the real-time feedback of the OCT imaging system to adjust the energy output of the laser ablation system so that the tissue ablation depth of the area to be ablated is consistent with the target ablation depth, including: Define the function mapping relationship between the laser energy output per unit time and the tissue depth change: Among them, d t is the tissue ablation depth at time t, which is determined by the tissue depth change real-time feedback of the OCT image; the input variable of the system is composed of the tissue ablation depth d t at time t and the output energy u per unit time t . Δd t = d t+1 - d t is the ablation depth change amount at adjacent times and is the output variable of the system; f is the mapping function between the laser energy output per unit time and the tissue depth change; Obtain data through offline sampling, use Gaussian process regression (GPR) to obtain the dynamic relationship of tissue depth change, and establish a non-linear fitting: wherein, are respectively the training data sets of the input and output variables obtained from the real environment, and it is considered that the output variable of the system conforms to a Gaussian distribution with a mean of 0 ; K(X,X) is the covariance matrix; I is the identity matrix; σ 2 is the variance of the Gaussian noise; Given the training data set and the input variables of the system , construct the following joint probability distribution and predict the depth change of the system given any input sequence of laser energy : Among them, \(K(X,X)+I\sigma\) 2 is the covariance function between the input data set \(X\) and itself plus the noise term; is the covariance function between the input data set \(X\) and ; is the covariance function between the input variable and the input data set \(X\); is the covariance function between the input variable and itself; Based on the dynamic relationship of the tissue depth change, and construct a loss function with the target ablation depth, predict the tissue ablation depth within the next K time instants through forward mapping: Among them, J ' is the second cost function; is the average value of the random variable; d K is the tissue ablation depth within the next K time instants; d d is the target ablation depth; Q depth , S depth are both symmetric positive definite weight matrices in ablation depth control, controlling the cost of ablation depth control error and control input respectively. The subscript depth represents ablation depth control; l(d t ) is the process loss function, h(d T ) is the terminal loss function; the superscript T represents transpose; Minimize the second cost function J ' , and determine the optimal tissue ablation depth at each moment within the next K moments , thereby obtaining a set of optimal energy laser sequences, combined with the depth change , to set the laser power at the current moment and adjust the energy output of the laser ablation system in real time; Utilize the tissue depth change real-time feedback of the OCT image to update the dynamic relationship of the tissue depth change, and repeat the above optimization process until the tissue ablation depth of the area to be ablated is consistent with the target ablation depth.

7. An electronic device, characterized in that, Include: One or more processors; A memory; And one or more programs, wherein the one or more programs are stored in the memory and are configured to be executed by the one or more processors. The programs include those for controlling the autonomous three-dimensional tissue ablation control system as described in claim 1 to execute the autonomous three-dimensional tissue ablation method. The soft robot includes a soft robot body, a laser ablation system, and an OCT imaging system. The autonomous three-dimensional tissue ablation method includes: Pre-construct a laser-tissue interaction model to describe the interaction relationship between the blue laser and the tissue, and determine the target ablation depth of the area to be ablated; Receive multiple two-dimensional ablation points planned by a doctor on an endoscopic image and generate a laser ablation path; Control the soft robot body to track the laser ablation path; including: Establish a linearized model of the soft robot body using the Koopman method; Based on the linearized model of the soft robot body, and use the MPC method to optimize the robot control strategy to ensure the tracking accuracy of the laser ablation path; During the path tracking process, combine the interaction relationship between the blue laser and the tissue and the real-time feedback of the OCT imaging system to adjust the energy output of the laser ablation system so that the tissue ablation depth of the area to be ablated is consistent with the target ablation depth; The step of, during the path tracking process, combining the interaction relationship between the blue laser and the tissue and the real-time feedback of the OCT imaging system to adjust the energy output of the laser ablation system so that the tissue ablation depth of the area to be ablated is consistent with the target ablation depth, includes: Define the function mapping relationship between the laser energy output per unit time and the tissue depth change: where d t is the tissue ablation depth at time t, determined by the tissue depth change real-time feedback of the OCT image; the input variable of the system is composed of the tissue ablation depth d t at time t and the output energy u per unit time t , Δd t = d t+1 - d t is the ablation depth change amount at adjacent times and is the output variable of the system; f is the mapping function between the laser energy output per unit time and the tissue depth change; Obtain data through offline sampling, use Gaussian process regression (GPR) to obtain the dynamic relationship of tissue depth change, and establish a non-linear fitting: Among them, are respectively the training data sets of the input and output variables obtained from the real environment, and it is considered that the output variable of the system conforms to a Gaussian distribution with a mean of 0 ; K(X, X) is the covariance matrix; I is the identity matrix; σ 2 is the variance of the Gaussian noise; Given the training data set and the input variables of the system , construct the following joint probability distribution and predict the depth change of the system given any input sequence of laser energy : Among them, \(K(X, X)+I\sigma\) 2 is the covariance function between the input data set \(X\) and itself plus the noise term; is the covariance function between the input data set \(X\) and ; is the covariance function between the input variable and the input data set \(X\); is the covariance function between the input variable and itself; Based on the dynamic relationship of the tissue depth change, and construct a loss function with the target ablation depth, predict the tissue ablation depth within the next K time instants through forward mapping: Among them, J ' is the second cost function; is the average value of the random variable; d K is the tissue ablation depth within the next K time instants; d d is the target ablation depth; Q depth , S depth are both symmetric positive definite weight matrices in ablation depth control, controlling the cost of ablation depth control error and control input respectively, and the subscript depth represents ablation depth control; l(d t ) is the process loss function, h(d T ) is the terminal loss function; the superscript T represents transpose; Minimize the second cost function J ' , and determine the optimal tissue ablation depth at each moment within the next K moments , and then obtain a set of optimal energy laser sequences, combined with the depth change , to set the laser power at the current moment and adjust the energy output of the laser ablation system in real time; Utilize the tissue depth change real-time feedback of the OCT image to update the dynamic relationship of the tissue depth change, and repeat the above optimization process until the tissue ablation depth of the area to be ablated is consistent with the target ablation depth.

Images